Granola
无需机器人入会的 AI 会议记事本,正在转向企业级上下文层;最新估值 $1.5B,但没有公开财务披露
Granola 的无 bot 会议纪要产品有强劲且相互印证的融资动能和客户证明,但它的 $1.5B Series C 估值无法用任何已披露收入、ARR 或毛利数据核验。
封面要素
公司概况
Granola 是一家总部位于伦敦、在英国注册的 AI 会议笔记公司,由 Chris Pedregal 和 Sam Stephenson 于 2023 年 3 月创立。它的无机器人记事本把用户输入的笔记、实时 AI 转写(Deepgram、AssemblyAI)以及 OpenAI/Anthropic 模型融合成结构化会议回顾,并通过 Spaces、个人和企业 API、MCP 与 Briefs 扩展为全公司的会议上下文层。Granola 从 2024 年 5 月公开发布,到 2024 年 10 月完成 $20M Series A、2025 年 5 月以 $250M 估值完成 $43M Series B,再到 2026 年 3 月由 Index Ventures 领投、估值 $1.5B 的 $125M Series C;累计披露融资 $192M。公开客户佐证覆盖 Vanta、Brex、Zapier、Vercel、Asana、Gusto 等技术行业客户,但公司没有披露收入、ARR、毛利率、员工数或席位数;英国小公司审计豁免意味着近期法定文件也不会披露这些指标。
- 成立时间
- 2023-03-01
- 创始人
- Chris Pedregal, Sam Stephenson
- 创立地点
- London, United Kingdom
- 总部
- London, United Kingdom
- 产品
- 无需机器人入会的 AI 会议记事本,在本地采集麦克风 / 系统音频,经第三方 ASR(Deepgram、AssemblyAI)转写,再用 OpenAI/Anthropic 模型把转写稿与用户自己的笔记融合为结构化、可分享的会议记忆;Spaces、个人 / 企业 API、MCP、Recipes 和 Briefs 继续扩展这一层。
- 客户
- 知识工作者,以及科技 / SaaS / AI 原生团队;通常先由个人采用,再进入 IT 主导的全公司部署(例如 Vanta、Brex、Zapier、Vercel、Asana、Gusto、Thumbtack、Cursor)。
- 商业模式
- 免费增值 SaaS 订阅:免费层,加 $14/用户/月 Business 计划和 $35/用户/月 Enterprise 计划;个人和企业 API 打包进付费层;从个人用户切入,再扩到全公司许可证。
- 阶段
- late-stage private (Series C, $1.5B valuation, March 2026)
- 融资情况
- $125M Series C 由 Index Ventures 于 2026-03-25 领投,估值 $1.5B;Kleiner Perkins 以及老股东 Lightspeed、Spark Capital、NFDG 参投;Series C 后累计披露融资达到 $192M(TechCrunch、The SaaS News)。
执行摘要
主要优势
- 融资历史和可比公司事实,至少由两个独立交叉核验来源一致印证,包括公司官方声明和高声誉新闻报道。
- 无 bot、设备级架构真实存在,并被安全研究员拆解和逆向工程开源客户端独立印证;这支撑了大约三个月完成的 SOC 2 Type II 认证。
- 独立 B2B 交易数据(YipitData Signals)显示 2026 年采用率出现真实拐点:中型市场客户环比增长 25%、六个月支出增长 3x、实测流失接近零,并且用户从竞品切换过来的方向不对称。
- 四个公司发布的旗舰案例(Zapier、Vanta、Brex、Vercel)都显示,一个基层用户带动全组织 52-100% 采用,证明这种 land-and-expand 模式能在不同买方画像中重复。
- AI 会议纪要品类有持久且文档充分的结构性需求驱动;我们审阅的每一种市场规模口径都指向 2030 年代初前的双位数增长。
主要风险
- Granola、投资人或任何第三方数据商都没有披露收入、ARR、毛利率、现金余额或烧钱速度,因此 $1.5B 投后估值无法对照公司自身财务基底核验。
- 针对 Otter.ai 合并联邦集体诉讼中的窃听和生物声纹同意法律先例正在推进,并非假设;法律评论也将其视为可能波及 Granola 的品类级风险。
- 捆绑式巨头(Microsoft Teams/Copilot、Google Workspace/Gemini、Zoom AI Companion)把可比纪要功能放进买方已经付费的订阅里免费送,压住 Granola 独立定价权。
- 一年内披露的三起安全事后复盘,以及 March 2026 本地数据库加密变更打断第三方 agent 工作流,削弱了 Granola 自己「默认私密」的营销说法;独立报道也与之相矛盾。
- 资本充足性披露在结构上不透明:Granola Labs Ltd 享有英国小公司审计豁免,意味着未来普通法定文件不会披露营业额、毛利或现金烧钱。
未决问题
- Granola 的绝对收入或 ARR 数字,以及按 Business/Enterprise/API 分层的收入拆分。
- Granola 的账上现金、月度烧钱速度和以月计的现金跑道。
- 按客户细分拆分的净收入留存、总收入留存、logo 流失率或合同价值数据。
- 确认三篇已发布安全事后复盘是否覆盖全部重大事件,并提供 HIPAA 商业伙伴协议或 EU 托管路线图。
- 确认 Granola 是否任命 CFO、General Counsel 或 Head of Compliance,以及公司是否涉及任何诉讼或监管行动。
目录
01公司概况
1.1 身份、产品与阶段
Granola 是一家 AI 会议笔记创业公司,围绕一句简单定位搭建:没有会议机器人,也能留下笔记、行动项和记忆。官网、产品页和投资人文章都把它描述成有人参与的记事本:产品结合用户输入、AI 转写和结构化处理,而不是让机器人加入通话并自主总结一切。这个设计选择很关键,因为它同时界定了产品体验和市场类别:Granola 想更接近面向对话的 Apple Notes 或 Notion,而不是一个显眼的 Zoom copilot。 截至 2026-07-01 运行日,公司显然已经进入后期私营创业公司阶段。官方和独立的 2026 年 3 月报道都显示,总部在伦敦,支持 macOS、Windows 和 iPhone,并把业务描述为从个人专业用户记笔记工具扩展到企业会话基础设施。到 2026 年 3 月,Granola 估值已达 $1.5 billion,并把 Spaces、API 和连接器界面定位为公司会议上下文的操作系统,而不只是独立转写工具。[CO001, CO002, CO003, CO008, CO009, CO010]
| 指标 | 数值 / 状态 | 日期 / 时点 | 信心 | 尽调缺口 |
|---|---|---|---|---|
| 成立时间 | 2023 年 3 月 | 2023-03 | 高 | |
| 总部 | 英国伦敦 | 2026-03 | 高 | |
| 最新轮次 | Index Ventures 领投 $125M Series C | 2026-03-25 | 高 | |
| 最新公开估值 | 投后估值 $1.5B | 2026-03-25 | 高 | |
| 累计融资 | Series C 后公开报道 $192M | 2026-03-25 | 高 | 与已签署融资文件和 cap table 核对 |
| 平台 | macOS、Windows、iPhone | 2026-07-01 | 高 | |
| 公开牵引点 | 2024 年 10 月约 5,000 周活用户;2025 年 5 月每日转写数百万分钟 | 2024-10 至 2025-05 | 中 | 需要当前用户数、席位数和使用量数据 |
| 收入 / ARR | 2026-07-01 | 低 | 向管理层索取 ARR、收入 run-rate、毛利率、烧钱速度和现金续航 | |
| 员工数 | 2026-07-01 | 低 | 索取准确员工数、招聘计划和组织架构图 | |
| 客户数 / 席位数 | 2026-07-01 | 低 | 索取付费账户数、席位数和头部客户集中度 |
null 字段表示该指标未在已审阅的 2024-2026 年官方和独立来源中公开披露;公开牵引点是历史快照,而不是当前 run-rate 披露。
[CO003, CO010, CO011, CO012, CO015, CO019]Granola 的产品逻辑从无 bot 采集出发,流向结构化笔记、共享 Spaces,再到下游 AI / 工作流激活。
[CO008, CO009, CO011, CO024, CO025, CO027]公开证据显示采用动能和参考客户 ROI 都强,但财务披露深度还达不到 $1.5B 估值应有水平。
[CO015, CO016, CO019, CO021, CO023, CO031]1.2 创始人、领导层与治理
近期新闻中的公开创始人记录保持一致:Granola 由 Chris Pedregal 和 Sam Stephenson 创立。Pedregal 在公开材料中更可见,也有更清晰的创始人-市场匹配履历,来自此前 AI 与产品工作的积累。BusinessCloud 报道称,他曾创办 Socratic,2018 年出售给 Google,之后在 Google 的 Area 120 孵化器内部打造 Stack,直到该产品并入 Google Drive。Stephenson 在报道和公司材料中作为联合创始人与产品 / 设计搭档出现,包括在 2024 年 10 月 Series A 报道中以联合创始人身份被引用。 更不透明的是更宽的高管和治理班底。抓取到的公司、投资人和媒体页面反复聚焦两位创始人,但没有公布完整董事会名单、独立董事阵容,或清晰具名的 CFO/COO 架构。从尽调角度看,Granola 仍显得以创始人为中心:有号召力、产品驱动、可能行动很快,但相对其 Series C 后估值,关键人集中和公开治理披露不足是真问题。[CO004, CO005, CO006, CO007, CO044]
| 人物 / 职务 | 公开可支持背景 | 当前公开角色 | 创始人 - 市场匹配 / 依赖 | 披露提示 |
|---|---|---|---|---|
| Chris Pedregal / 联合创始人兼 CEO | 创办 Socratic(出售给 Google),之后在 Google Area 120 内部搭建 Stack | 融资和产品发布中的主要公开高管声音 | AI / 产品领域的创始人 - 市场匹配很强;外部叙事和产品战略上存在清晰的关键人物依赖 | 已获取来源中未发现公开 CFO/COO 对位人 |
| Sam Stephenson / 联合创始人 | 融资报道中公开识别为联合创始人及设计 / 产品搭档 | 产品 / 设计联合创始人;在官方叙事中不如 Pedregal 显眼 | 对 Pedregal 的运营画像形成重要平衡,尤其体现在 UX 和产品工艺上 | 详细履历和治理角色公开披露较少 |
| 董事会 / 治理梯队 | 已审阅表面未发现独立董事、董事会委员会和具名财务运营高管披露 | 已获取官方 / 新闻材料中未披露 | 相对其估值阶段,显示创始人中心化治理 | 实时尽调中索取董事会材料、董事名单和高管组织架构图 |
这只是公开可见领导梯队的部分列举;缺少公开董事会和财务负责人披露,本身就是一个尽调信号。
[CO004, CO005, CO006, CO007, CO044]1.3 融资历史、牵引力与利益相关方
Granola 的融资弧线压缩得异乎寻常。2024 年 10 月之前,TechCrunch 报道其完成由 Lightspeed 领投的 $4.25M 种子轮。2024-10-23,公司宣布完成 Spark Capital 领投的 $20M Series A。2025-05-14,公司发布 Granola 2.0,并同步完成 NFDG 领投、估值 $250M 的 $43M Series B。不到一年后,2026-03-25,Granola 宣布完成 Index Ventures 领投、估值 $1.5B 的 $125M Series C,Kleiner Perkins 和老股东也参与。TechCrunch 与 The SaaS News 均称,Series C 后累计融资达到 $192M。 公开牵引力信号方向上很强,尽管并不完整。TechCrunch 报道称,到 2024 年 10 月,Granola 约有 5,000 名周活用户,首周复用率 70%,后来周用户基数增长 10%。BusinessWire 和招聘页称,Granola 每天转写数百万分钟对话。对于一家年轻软件公司,客户佐证异常具体:Vanta、Brex、Zapier、Vercel 都有公开案例,披露了量化节省时间或清晰的流程改善。股权结构表质量也高——Index、Kleiner、Spark、Lightspeed、NFDG,以及来自 Vercel、Replit、Linear、Shopify 等公司的运营型天使,既代表资本支持,也带来进入技术团队的强分发邻近性。[CO013, CO014, CO015, CO016, CO017, CO018]
| 利益相关方 | 角色 / 关系 | 为什么重要 | 公开信号 | 尽调问题 |
|---|---|---|---|---|
| Lightspeed | 种子轮领投方和连续支持者 | 锚定第一轮机构融资,并延续支持至后续轮次 | 释放早期产品信念和持续支持信号 | 确认持股比例、董事会权利及任何 pro-rata 承诺 |
| Spark Capital | Series A 领投方和连续支持者 | 在 2024 年 10 月设定首个大规模发布后估值基准 | 在 Granola 尚未广泛企业成熟前提供支持 | 确认从种子轮到 A 轮再到 C 轮的 step-up 经济条款 |
| NFDG(Nat Friedman 与 Daniel Gross) | Series B 领投方 | 支持 2025 年 5 月企业 / workspace 转向,估值 $250M | 带来 AI builder 可信度和运营者网络 | 审查治理权利和后续融资义务 |
| Index Ventures | Series C 领投方 | 2026 年 3 月将公司定价至独角兽状态 | 为后期扩张预期给出领投信号 | 审查 C 轮 preference stack 和增长计划 |
| Kleiner Perkins | Series C 参投方 | 增加后期品牌验证和企业软件模式匹配 | 发布明确论点,将 Granola 视为公司情境层 | 明确持股比例和 go-to-market 支持预期 |
| 运营者天使群体 | 天使投资人包括 Vercel、Replit、Shopify、Linear、Intercom、Ramp 和 Warp 的领导者 | 有助于产品分发、招聘和技术可信度 | 显示异常强的从业者拉力 | 识别任何客户引荐或产品顾问协议 |
| 企业客户背书 | Vanta、Brex、Zapier 和 Vercel 案例研究提供公开 ROI 证明 | 验证其部署超出 VC 小圈子 | 具名背书降低玩具 AI 工具风险 | 要求实时客户背调电话和合同条款 |
| AI 平台连接器生态 | Claude、ChatGPT、Figma Make、Replit 和其他连接器界面扩展效用 | 可能成为进入 agent 工作流的分发楔子 | 同时抬高机会和依赖问题 | 梳理合作伙伴集中度和 API 使用经济性 |
这张图混合了财务利益相关方、参考客户和生态合作伙伴,因为 Granola 的规模化路径同时依赖三者,而不只依赖资本。
[CO013, CO014, CO017, CO021, CO022, CO030]Granola 的公开叙事把从种子轮到独角兽的完整弧线压缩到约三年,产品边界扩张与每一步融资同步推进。
[CO003, CO014, CO017, CO021, CO026, CO027]1.4 里程碑、控制与负面信号
里程碑记录显示,公司从发布到平台扩展走得很快。Granola 于 2024 年 5 月公开发布,2024 年 10 月完成 Series A,2025 年 5 月推出协作式 Granola 2.0 工作区,2025 年 9 月发布 Granola Chat 的 Recipes,2026 年 2 月通过 Granola MCP 暴露会议上下文,随后借 2026 年 3 月 Series C 继续推进 Spaces、文件夹和 API。2026 年 5 月,公司又加入 Briefs,让用户开会前就能拿到已整理的参会者和历史对话上下文。 信任与控制图景有好有坏。正面看,Granola 的隐私政策称 OpenAI、Anthropic 等第三方不能用客户个人数据训练模型,DPA 也明确覆盖 GDPR、UK GDPR、瑞士数据保护和美国州隐私法。负面看,同一隐私政策允许在用户未退出时使用去标识化数据训练;2026 年 3 月,Granola 决定加密本地数据库,打断早期本地智能体工作流,引发用户反弹。这个事件不否定产品牵引力,但确实说明,平台控制决策会在早期帮助品牌发酵的技术重度用户中制造采用摩擦。[CO020, CO025, CO026, CO027, CO028, CO035]
| 日期 | 事件 | 类型 | 金额 / 状态 | 参与方 | 含义 |
|---|---|---|---|---|---|
| 2023-03 | Granola 成立 | 创立 | Chris Pedregal;Sam Stephenson | 开启公司的公开运营时钟 | |
| 2024-05 | 封闭 beta 后公开发布产品 | 产品 | 发布 | 早期重度用户 / VC / 创始人 | 形成首个可见采用浪潮和周活用户增长基础 |
| 2024-10-23 | Series A 公告 | 融资 | $20M | Spark Capital 及 AI Grant、Lightspeed、Betaworks、Firstminute 等 | 资助发布后扩张,并强化 product-market fit |
| 2025-05-14 | Granola 2.0 与 Series B | 产品 | Workspace 发布 + $43M、估值 $250M | NFDG 及回投投资人 | 标志着从个人记事本转向团队 workspace |
| 2025-09-30 | 新 Granola Chat 推出 Recipes | 产品 | 功能发布 | Granola 产品团队 | 在会议情境之上增加可复用的 prompt / 工作流层 |
| 2026-02-04 | Granola MCP 发布 | 合作 | 连接器发布 | Granola 及 Claude/ChatGPT 相邻工具生态 | 将会议情境扩展到外部 AI agent 和应用 |
| 2026-03-17 | 本地数据库加密反弹公开化 | 负面 | 用户信任事件 | 重度用户,包括 a16z 合伙人 Guido Appenzeller | 显示早期技术采用者中的开放性 / 互操作性张力 |
| 2026-03-25 | Series C 与 Spaces/API 推进 | 融资 | $125M、估值 $1.5B | 投资方:Index Ventures、Kleiner Perkins、Lightspeed、Spark、NFDG | 确立独角兽估值和企业情境叙事 |
| 2026-05-20 | Briefs 在 Mac 和 Windows 上发布 | 产品 | 功能发布 | Granola 产品团队 | 将产品进一步前移到会议准备,而不只是会后回顾 |
这是后续章节使用的标准日期顺序;产品事件和融资事件交错排列,因为市场叙事同时依赖二者。
[CO003, CO014, CO017, CO018, CO021, CO025]1.5 展示材料
02市场分析
2.1 市场边界、邻近类别与现状替代品
Granola 自己的产品和博客页面把公司描述为 AI 记事本:一个桌面和 iPhone app,在不向通话加入可见机器人参与者的情况下采集会议音频和用户输入的笔记,并可配合任何视频会议平台(Zoom、Google Meet、Microsoft Teams)使用。这个框定对划分市场边界很重要,因为它把 Granola 与两个容易混淆的相邻类别分开。第一类是既有协作套件内置的记笔记:Microsoft 365 Copilot 和 Teams Recap、由 Gemini 驱动的 Google Meet “Take notes for me”、Zoom AI Companion 的 Meeting Summary 都已经包含在现有协作套件订阅中,不另收费;这会限制任何独立厂商仅凭基础转写和总结能收取的价格。第二类是 Gong 这类专用销售会话智能软件,面向大型 SDR 组织,定价约 $1,300-$1,600 每用户每年,另加 $5,000-$50,000 的强制年度平台费;它的使用场景和价格点都与轻量记事本本质不同。 在独立记笔记类别内部,Granola 与 Fathom、Fireflies(声称覆盖超过 1 million 家公司)和 Otter.ai(主打基于日历的自动入会采集)等可选机器人或基于机器人的产品竞争。因此,尽调用的实际纳入市场,是独立按席位计费的 AI 会议笔记软件,加上 Granola 在 2026 年 3 月 Series C 轮推出的新兴“上下文层”:用于共享团队上下文的 Spaces、个人和企业 API,以及连接 Claude、ChatGPT、Replit 等第三方智能体工具的新版 MCP。排除在边界之外的是底层视频会议平台许可证本身、面向大型销售组织定价的专用会话智能平台,以及没有会议衍生上下文的通用公司 wiki 或文档管理套件。一个值得注意的现状替代品是,一些对隐私敏感的用户在 Granola 加密本地数据库后,选择回到普通本地 markdown 文件(配合 Obsidian 等工具),而不是信任任何厂商托管数据库;这提醒我们,对于最重视可携带性的买方,“根本不用专门工具”仍是活选项。[CM001, CM002, CM003, CM004, CM005, CM006]
| 细分 / 类别 | 纳入支出 | 排除支出 | 买方 / 付款方 | 相关性 |
|---|---|---|---|---|
| 独立 AI 会议笔记 / 记事软件 | Granola、Fathom、Fireflies、Otter 等专用记事应用的按席位 SaaS 订阅 | 底层视频会议平台许可证费用本身 | 团队负责人、IT,或自助个人用户 | Granola 产品的核心直接市场 |
| 在位厂商打包的会议笔记功能 | 零增量价格;已包含在 Teams、Google Workspace 和 Zoom 订阅中 | 不作为独立项目单独货币化 | 现有协作套件管理员 | 相邻替代品,压住独立供应商定价权 |
| 企业 conversation-intelligence 平台(Gong 类型) | 每用户年费约 $1,300-$1,600,另加 $5,000-$50,000 年平台费 | 小团队基础通话记事 | VP Sales / Sales Enablement | Granola 不直接竞争的相邻高 ACV 类别 |
| 团队知识 / 情境层(Spaces、API、MCP) | Business / Enterprise 计划附加功能,支持共享情境、API 访问和第三方 agent 连接器 | 不含会议衍生情境的通用公司 wiki 或文档管理套件 | IT / 产品领导层、agent 工具团队 | Granola 可服务支出的较新扩展层 |
| 对保密敏感的选择性捕获(VC、高管搜索、M&A) | 同样的按席位记事本订阅,但价值专门来自无机器人带来的低可见度 | 任何需要可见录制参与者的工具 | 管理合伙人 / 公司运营负责人 | 无机器人架构本身就是购买标准的差异化利基 |
纳入 / 排除支出边界来自 Granola 自身产品和博客页面,以及竞争对手和分析师页面;Gong 的美元数字来自第三方买方估计,因为 Gong 不公布价格。
[CM001, CM002, CM003, CM004, CM005, CM006]2.2 规模测算:用一组估算取代一个宽泛数字
没有单个 TAM 数字能公平代表 Granola 的机会,因为发布方对类别边界本身就有分歧。宽口径“AI meeting assistant”视角把转写、日程安排和分析打包在一起,将 2025 年全球市场估为约 $3.14 billion(The Business Research Company)、$3.47 billion(Grand View Research;由于原始报告被机器人挑战拦住,此处来自 GII Research 摘要)和 $3.50 billion(Statifacts);各发布方分别预计市场到 2030 年达 $9.33 billion、到 2032 年达 $17.55 billion、到 2033 年达 $21.48 billion,CAGR 为 24.3% 到 25.9%。更窄、专门隔离独立“AI note-taking”软件的视角小五到七倍:Congruence Market Insights 将该类别规模估为 2024 年 $450.7 million,到 2032 年增至 $1.80 billion(18.9% CAGR);Precedence Research 估为 2025 年 $623.5 million,到 2035 年增至约 $3.48 billion(18.75% CAGR)。Startup Fortune 引用的第三个独立来源估算称,PitchBook 将“AI transcription and meeting intelligence”2024 年估值定为 $3.5 billion,到 2030 年增至 $11.8 billion;这个数字更接近宽口径而非窄口径。 审阅到的每个视角都指向持久的两位数增长,但没有一个隔离出 Granola 自己可服务的市场。本研究集中没有公开来源披露 Granola 的收入、ARR 或付费席位数,因此无法推导公司特定的 SAM 或 SOM;这里将其保留为开放尽调问题,而不是猜一个数字。承销时的实际结论是:把宽口径视为类别顺风的外边界,把窄口径视为独立记笔记软件需求的更近似(但仍通用)代理,再按既有厂商捆绑功能已经免费满足的未知需求份额,对两者都打折。[CM012, CM013, CM014, CM015, CM016, CM017]
| 发布方 | 年份 | 地域 | 价值($B) | CAGR | 方法论 | 信心 | 局限 |
|---|---|---|---|---|---|---|---|
| Grand View Research(经 GII Research 摘要) | 2025 | 全球 | 3.47(→ 21.48,2033 年) | 25.8% | 自上而下的 AI 会议助手市场模型 | 中 | 主报告被机器人挑战拦截;本章依赖辛迪加摘要 |
| Statifacts | 2025-2035 | 全球 | 3.50 → 35.02 | 25.9% | 宽口径 AI 会议助手预测 | 中 | 独立研究供应商;底层方法论未被独立审计 |
| The Business Research Company | 2025-2030 | 全球 | 3.14 → 9.33 | 24.3% | 宽口径 AI 驱动会议助手预测 | 中 | 2030 年期限短于多数同业,因此不能在同一终点年份直接比较 |
| Precedence Research | 2025-2035 | 全球 | 0.6235 → 3.47674 | 18.75% | 更窄的「AI 记笔记市场」模型 | 中 | 类别定义较窄,可能排除更广的会议分析收入 |
| Congruence Market Insights(市场研究机构) | 2024-2032 | 全球(美国领先) | 0.4507 → 1.8003 | 18.9% | 更窄的「AI 记笔记解决方案」模型 | 中 | 2024 年为基准年,比多数同业估计早一年 |
| PitchBook(经 Startup Fortune 引用) | 2024-2030 | 全球 | 3.5 → 11.8 | 约 22% 隐含值 | 投资人数据机构市场分析,经二手引用 | 低 | 未独立获取;仅来自第三方新闻文章 |
| Grand View Research(主报告) | 2026 | 全球 | 主报告 | 低 | 直接获取返回机器人挑战响应(403);无许可访问则无法使用 |
宽口径「AI 会议助手」视角和窄口径「AI 记笔记软件」视角衡量的是不同类别边界,不应平均成一个 TAM;null 值标记本章无法直接验证的数据。
[CM012, CM013, CM014, CM015, CM016, CM017]广义 AI 会议助手估算构成远大于狭义 AI 记笔记软件估算的外层包络,而套件化既有厂商功能和 Granola 自身 SAM / SOM 仍未量化。
本图是视角堆叠,不是严格的 TAM-SAM-SOM 级联,因为各发布方使用的品类边界、基准年份和预测周期不同。
[CM012, CM015, CM018, CM019, CM021, CM044]发布的估算区间会随品类定义而明显变化:广义是会议助手,狭义则专指记笔记软件。
中点只是算术中心或居中的发布方估算,用来让比较更清楚;每个发布方都有自己的基准年份和预测期,所以图中呈现估算区间,不是一个三角校准后的数字。
[CM012, CM013, CM014, CM015, CM016, CM018]2.3 买方、用户、付款方分层与采用路径
Granola 的公开客户案例显示,多个不同买方细分里都出现了相似的自下而上采用模式,而不是一套统一的企业销售。在 Brex,IT 负责人是在发现销售、工程和支持团队已由员工自发采用之后,才把全公司访问正式化;批准推广前,还要求 SOC 2 Type 2 合规、SSO 和 SCIM。在 Vanta,买方是客户成功副总裁,需要替代每天 1:1 和跨职能同步中人工对齐现场笔记与原始转写稿的工作。在 Zapier,一名 Principal Designer 采用 Granola,为设计团队搭建共享、可搜索的知识库;在 Vercel,一名销售工程师需要更快、可搜索的通话回顾,以跟上潜在客户跟进节奏。Granola 自己的用例页面进一步把 Product Managers、Designers 和 Product Leaders 放在一类工作流里(客户研究库反哺路线图决策),把 Account Executives 放在另一类工作流里(通话回顾同步到 HubSpot、Affinity、Attio 或 Zapier 等 CRM)。 证据中还出现了一个高度重视私密性的细分:Granola 自己关于 second-brain 工作流的博客强调,风投合伙人用捕捉到的会议上下文加速投资委员会备忘录写作;混合团队指南则引用高管猎头公司 Daversa Partners,该公司在 150 名员工中有 136 人采用无机器人采集,原因正是在保密 CEO 搜索中,可见机器人参与者会显得唐突。公司自己的指引也在销售场景中按规模划线:15 名销售代表以下的创始人与 SMB 团队,被引向 $14 每用户每月的 Business 计划中的轻量 AI 记事本;30 名以上销售代表且有专职销售赋能经理的团队,则被引向 Gong 式会话智能。横跨这些细分,采用路径相当一致:个人先试用无机器人采集模式,使用在团队内非正式扩散,IT 或安全审查引入合规要求,之后才形成正式、有预算的企业部署;部分账户随后进一步扩展到 API、Spaces 和智能体集成。[CM023, CM024, CM025, CM026, CM027, CM028]
| 细分 | 买方 | 用户 | 付款方 | 工作流 | 预算负责人 | 采用触发器 |
|---|---|---|---|---|---|---|
| SMB / 创始人主导的销售与运营 | 创始人或团队负责人 | 个人贡献者 | 自助团队卡 | 非正式通话笔记流入 CRM | 创始人 | 启动无需合规负担 |
| 企业 IT / 安全治理推广 | IT / 安全负责人 | 全公司员工 | 中央 IT 预算 | 全组织可搜索会议知识库 | IT / 中央采购 | 自下而上自然使用达到阈值后,触发正式 SSO/SCIM 推出 |
| 客户成功 / 客户管理 | 客户成功副总裁 | CS/AM 负责人 | CS 部门预算 | 1:1 与客户通话上下文采集 | CS/营收负责人 | 碎片化手工笔记核对无法规模化 |
| 产品 / 设计 / 研究 | 设计或产品负责人 | PM、设计师、研究员 | 产品组织预算 | 用于路线图验证的客户访谈库 | 产品 / 设计负责人 | 缺少可搜索的研究事实源 |
| 销售 / GTM 与客户经理 | 销售运营 / AE 经理 | AE、销售工程师 | 销售组织预算或自助购买 | 同步 CRM 的通话摘要与跟进事项 | 销售负责人 | 现有工具产出可执行摘要太慢 |
| 对保密敏感的机构(VC、高管猎头) | 管理合伙人 / 机构运营 | 合伙人、招聘顾问 | 机构运营预算 | IC 备忘录与候选人上下文采集 | 机构合伙人 / 运营负责人 | 敏感对话需要无 bot 的私密性 |
各行来自 Granola 自有客户案例、用例页面和博客文章;如果未给出正式头衔,预算负责人标签按买方披露角色推断。
[CM023, CM024, CM025, CM026, CM027, CM028]企业级 IT 治理部署和高度重视保密的机构,对合规和低调性要求最高;SMB 自助购买者最低。
[CM023, CM024, CM025, CM027, CM030, CM031]公开案例研究指向一条可重复路径:个人无 bot 试用,扩散到团队,进入 IT / 安全审查,最后扩展到 APIs 和智能体集成。
[CM023, CM026, CM034, CM039, CM041, CM042]2.4 增长驱动、采用约束与尽调缺口
AI 会议笔记软件的结构性需求驱动看起来持久。Microsoft Work Trend Index 遥测显示,平均 Microsoft 365 用户在核心工作时段每两分钟就被会议、邮件或聊天打断一次(每天 275 次),57-60% 的会议是临时或 ad hoc 会议,近三分之一会议现在跨越多个时区,比 2021 年上升 35%。Granola 自己的混合团队指南引用 Gallup 调查数据称,美国可远程办公员工中,52% 现在采用混合安排,26% 完全远程,持续支撑对可靠捕捉现场和远程音频工具的需求。Granola 引用的一项来自 IDC 的估算认为,Fortune 500 公司每年因机构知识流失付出 $31.5 billion 成本;这强化了 Granola 2026 年 3 月从个人记事本扩展到 Spaces、API 和 MCP 连接器的商业逻辑:目标是捕捉并复用组织上下文,而不只是单次会议。 约束也同样具体。Microsoft、Google 和 Zoom 都把免费的 AI 会议笔记功能直接捆绑进许多买方已经付费的订阅中,这压低了独立厂商仅凭记笔记能收取的价格,并推动 Granola 这类公司转向差异化上下文和智能体工作流,以证明单独预算项合理。公开记录中也能看到信任和监管摩擦:Startup Fortune(引用此前 The Verge 报道)发现,Granola 默认向任何持有分享链接的人暴露笔记,并在用户未退出时用转写稿训练;BuildBetter 的独立评测给 Granola 企业就绪度打出 3/10,理由是只用美国 AWS 托管,没有欧盟区域或本地部署选项;2026 年 3 月,Granola 加密本地笔记数据库的决定在一个广泛传播的社交帖子中遭到 a16z 合伙人 Guido Appenzeller 公开批评,促使一些用户考虑基于 markdown 的替代方案。单看任何一项约束都不致命,但合在一起说明,Granola 的增长取决于同时补上信任、数据驻留和锁定缺口;与此同时,捆绑式既有厂商正在把基础记笔记功能商品化。两个规模问题仍真正开放,因此保留而不估算:宽口径分析师估算市场中,有多少已被既有厂商捆绑功能而非独立厂商捕获;以及 Granola 自己的 SAM/SOM 会更接近宽口径还是窄口径分析师视角。[CM035, CM036, CM037, CM038, CM039, CM040]
| 驱动因素 / 约束因素 | 方向 | 时点 | 影响 | 尽调问题 |
|---|---|---|---|---|
| 会议 / 消息过载(Microsoft 遥测数据) | 驱动 | 当前 | 支撑自动采集与摘要工具的基线需求 | 跟踪过载指标在 2026 年后继续恶化还是趋于平台期 |
| 混合 / 远程办公延续(Gallup) | 驱动 | 当前 | 让异步、跨格式会议采集需求保持韧性 | 监测返办公室强制要求是否压缩混合办公人群 |
| 组织知识流失成本(IDC 来源估计) | 驱动 | 中期 | 强化从记笔记扩展到“第二大脑”/ 上下文层的理由 | 直接核验 IDC 估计的原始方法和发布时间 |
| 产品扩展到 Spaces/API/MCP/智能体集成 | 驱动 | 当前至中期 | 把 Granola 单账户可触达支出从单席记事本扩大出去 | 跟踪 API 与智能体使用带来的附加率和增量 ACV |
| 既有套件捆绑功能(Teams、Meet、Zoom) | 约束 | 当前 | 压住基础记笔记的独立定价权,推动厂商转向差异化上下文 / 智能体价值 | 单独量化面对既有捆绑功能时的赢单 / 输单率 |
| 默认隐私 / 信任设置反弹 | 约束 | 当前 | 监管与声誉风险可能拖慢企业信任建设 | 确认当前默认分享与训练退出设置 |
| 数据驻留 / 监管缺口(无 EU 区域) | 约束 | 当前 | 限制 EU 总部或对数据驻留敏感的企业买家采用 | 确认当前托管区域及是否有 EU 数据驻留路线图 |
| 2026 年 3 月本地数据库锁定反弹 | 约束 | 当前 | 一位知名投资人公开凸显了更高切换成本和信任风险 | 跟踪变更后的流失与公开舆情 |
| 更高 ACV 的会话智能板块(Gong 风格) | 约束 | 当前 | 限制 Granola 在最大、最复杂销售组织中的可触达支出 | 评估 Granola 计划向上竞争,还是让出该板块 |
方向反映当前可观察到的净效应:它对 Granola 可触达市场和定价权的影响;不是对公司特定结果的预测。
[CM035, CM036, CM037, CM038, CM039, CM040]2.5 展示材料
03竞争格局
3.1 格局:直接同行、捆绑式既有厂商、相邻会话智能、硬件进入者与现状替代品
Granola 自己的 “AI Meeting Assistant - Without the Bot” 页面围绕一个轴线定义整个竞争类别:竞品工具是否以可见机器人身份加入通话,还是在本地无机器人采集音频(CP001);该页面还声称,在列出的选项中,Granola 是唯一默认保持笔记私密且无需可见机器人的产品(CP002)。这个框定在 Granola 发布时是准确的,但到 2026 年已经不再是清晰差异点。tl;dv 现在向超过 2 million 用户宣传 “NO BOT REQUIRED”(无需机器人)选项(CP003),Fathom 自己的 2026 年定价页也在个人层级中提供“无机器人(beta)或机器人采集二选一”(CP004)。按声称用户数看,Otter.ai 是最大的独立竞品,仍默认使用连接日历的机器人,自动加入 Zoom、Microsoft Teams 或 Google Meet(CP005);因此,无机器人 / 基于机器人的分界越来越是在区分产品选择,而不是区分类别。 除了直接软件同行,Granola 至少还面对另外四类替代。捆绑式既有厂商——Microsoft Teams/Copilot、Google Workspace/Gemini 和 Zoom AI Companion——在大多数买方已经付费的订阅里提供可比的记笔记功能。Gong 相邻但不同,是为大型销售组织打造的高得多 ACV 会话智能平台,而不是轻量记笔记(CP021)。TechCrunch 2026 年 3 月的硬件综述直接把 Granola 与 Read AI、Fireflies、Fathom 并列,称一批增长中的可穿戴和口袋式 AI 记笔记设备可以替代这些数字竞品,用于线下采集(CP006、CP028)。最后,一些对隐私敏感的用户面对任何厂商托管的笔记数据库——包括 Granola 自己的数据库——都会回到本地 markdown 文件或通用笔记 app,让“没有专用工具”持续成为现状替代品(CP029)。合在一起,本章把买方的真实选择集视为五个结构不同的类别:直接无机器人 / 可选机器人同行、捆绑式既有厂商、相邻高 ACV 类别、可能的硬件进入者,以及内部自建 / 维持现状选项(CP048)。[CP001, CP002, CP003, CP004, CP005, CP006]
Granola 的无 bot / 隐私优先定位很强,但分发触达弱;捆绑型既有厂商正好相反:无 bot 卖点几乎为零,分发杠杆最高。
坐标轴分数是证据支撑的序数判断,来自本章审阅的 2026 年产品、定价、营销页面和独立报道;不是标准化调研或分析师评分方法。
[CP001, CP003, CP004, CP005, CP009, CP011]3.2 按竞品类别拆解:直接同行、捆绑式既有厂商,以及相邻会话智能层
在直接独立同行中,Fathom 的 2026 年定价页列出免费层,以及 Team($19/$15)和 Business($34/$25)按用户月付 / 年付层级,并包含 MCP 与 Claude/ChatGPT 集成(CP007、CP008)。Fireflies.ai 声称覆盖超过 1 million 家公司,并在首页直接展示 SOC 2 Type II、GDPR 和 HIPAA 徽章,同时给出“零数据留存”AI 训练政策(CP009);其 Pro、Business、Enterprise 层级大致覆盖 $10-$39 每席位每月(CP010)。Otter.ai 是审阅同行中的规模领导者:2026 年 4 月重启发布称拥有超过 35 million 用户、处理超过 1 billion 场会议,并被 Fortune 500 的 86% 使用(CP011);同时推出新的 “Conversational Knowledge Engine”,带 MCP server/client,并声称可触达机会约 $100 billion(CP012)——这一策略几乎逐项映射 Granola 自己的 Spaces/API/MCP 扩展。tl;dv 报告超过 2 million 用户和 4.7/5 评分(CP016),Read AI 则把名为 Ada 的 “digital twin” 助手与官方 Zoom “Essential App” 合作伙伴身份绑定(CP019)。 捆绑式既有厂商和相邻类别又是另一种形态。Microsoft 自己的支持页面确认,智能 Teams recap 需要 Teams Premium 附加许可证或 Microsoft 365 Copilot 许可证(CP023),而 Microsoft 将 Teams Premium 定价为 $10.00 每用户每月——低于本章审阅的所有独立厂商标价(CP024)。Google 的 “Take notes for me” 与 Zoom 的 AI Companion 都包含在现有付费订阅中,不另列单独价格(CP025、CP026);Zoom 报告称,AI Companion 月活用户在 FY26 Q4 同比增长超过三倍(CP026)。相比之下,Gong 声称拥有超过 5,000 家客户;独立研究估计其 Foundations 计划约 $1,600 每用户每年,另加强制 $5,000-$50,000 年平台费(CP021、CP022)——比任何独立记事本高一个数量级,并且只面向大型 SDR 组织;这恰好也是 Granola 自己销售指南所说不需要 AI 记事本的买方(CP034)。[CP007, CP008, CP009, CP010, CP011, CP012]
| 竞争对手 | 类别 | 规模 / 融资 | 目标客群 | 差异化 | 局限 |
|---|---|---|---|---|---|
| Granola | 直接同业 / 无 bot 品类开创者 | $125M Series C(2026 年 3 月),估值 $1.5B;累计披露融资 $192M | 连续开会的个人与团队;正扩展到 IT 治理的企业铺开 | 无 bot 本地采集、可跨数月笔记推理的智能体式 Granola Chat、持续扩展的 Spaces/API/MCP 上下文层 | 其“默认私密”营销主张,被独立报道的默认链接分享和模型训练情况相矛盾 |
| Fathom | 直接同业(bot 可选) | 定价 / 规模只由自有产品页面披露;据一份独立 2026 年报告,G2 评分为独立品类最高 | 个人到中端市场销售 / CS 团队 | 免费无限层级、无 bot 测试版采集选项、公开 API/MCP 以及 Claude/ChatGPT 集成 | 已审阅公开来源未披露公司层面融资或客户数 |
| Fireflies.ai | 直接同业(基于 bot、工作流重) | 据自有营销材料,覆盖 1M+ 家公司;已审阅来源未披露融资 | 希望获得 CRM/ATS 自动化的 SMB 到企业团队 | 200+ 项 AI Skills,广泛 CRM/PM/ATS/Slack 集成,明确零数据留存的 AI 训练政策 | 据独立报道,因同意相关录音做法面临与 Otter.ai 类似诉讼 |
| Otter.ai | 直接同业(基于 bot、规模领先者) | 截至 2026 年 4 月重新发布,声称拥有 35M+ 用户、1B+ 场会议,并瞄准 >$100B 可触达市场 | 从个人到大型企业;将自身定位为对标 CRM/ERP 规模的平台 | MCP 客户端 / 服务器、AI Chat Connectors,在已审阅同业中声称用户基数最大 | 因同意与生物识别数据主张面临合并后的联邦集体诉讼 |
| tl;dv | 直接同业(bot 可选) | 据自有首页,拥有 2M+ 用户,450 条评论综合评分 4.7/5 | 需要 CRM 自动化和通用会议采集的销售与 CS 团队 | “无需 bot”采集选项、多语言自定义摘要,声称接入 5,000+ 个 Zapier 集成 | 其 Granola 对比内容由自己撰写,本研究会话无法访问其专用定价页 |
| Read AI | 直接同业(参与度分析) | 定价层级最高 $39.75/用户/月;已审阅来源未披露融资 | 希望获得会议参与度指标和个人“数字孪生”助手的团队 | Ada 数字孪生助手、官方 Zoom “Essential App” 合作伙伴身份 | 独立报道点名其因同意担忧面临大学层面使用禁令 |
| Microsoft Teams / Copilot | 捆绑型既有厂商 | Teams Premium 附加包 $10/用户/月;完整 Copilot 摘要需要单独 Microsoft 365 Copilot 许可证 | 现有 Microsoft 365/Teams 企业订阅客户 | 在已购买协作套件内几乎零摩擦采用 | 据独立 2026 年市场评论,功能深度与跨平台覆盖落后于独立厂商 |
| Google Workspace / Gemini | 捆绑型既有厂商 | 包含在 Google Workspace 订阅内;已审阅来源未披露单独价格 | 现有 Google Workspace 订阅客户 | 零增量成本,自动集成 Google Docs/Drive/Calendar | 仅限 Google Meet;不采集 Zoom 或 Teams 会议 |
| Zoom AI Companion | 捆绑型既有厂商 | 包含在付费 Zoom Workplace 方案内,或单独提供;Q4 FY26 月活跃用户同比增长超过三倍 | 现有 Zoom Workplace Pro/Business/Enterprise 许可证客户 | 声称不使用客户数据训练模型,采用增长迅速 | 绑定 Zoom 会议平台;不采集 Teams/Meet 通话 |
| Gong | 邻近品类(更高 ACV 的会话智能) | 声称 5,000+ 客户;Foundations 方案估计约 $1,600/用户/年,另加 $5,000-$50,000 强制年度平台费 | 拥有专门赋能资源的大型 SDR / 销售组织 | 深度辅导、预测与营收分析工具,超出基础记笔记 | 成本结构只适合大型团队;Gong 不直接公布定价 |
| 硬件可穿戴设备(Plaud、Mobvoi、Omi、Viaim、Anker、Pocket) | 潜在进入者 / 替代品 | 设备价格大约 $89-$200,可选 $15-$20/月订阅;已审阅来源中多数厂商未披露融资 | 纯软件工具无法采集的线下会议和对话 | 用专用硬件形态采集离线 / 线下内容 | 除任何软件订阅之外,或作为软件订阅替代,还需购买实体设备 |
| 内部自建 / 现状(无专用工具) | 替代品 / 现状 | 无厂商关系;成本是内部工程与流程时间 | 最重视控制或可移植性的团队,包括部分对本地数据库反弹作出反应的团队 | 完全数据控制且零厂商锁定 | 没有重大内部工程投入,就没有 AI 摘要、搜索或跨会议记忆 |
各行覆盖本章公开证据中,2026 年 Granola 买家可见的最重要直接、既有、邻近、硬件进入者和现状替代方案;规模 / 融资数字按相应主张标注,混合了公司声称与独立估计。
[CP001, CP002, CP004, CP007, CP008, CP009]竞争对手自称的 2026 年规模远高于 Granola 任何可比公开数据;打包型在位者的采用在加速,Gong 的强制费用也凸显相邻品类定价逻辑差异很大。
[CP009, CP011, CP016, CP022, CP024, CP026]3.3 能力、定价、GTM / 分发,以及信任 / 隐私姿态
从原始能力看,单场会议转写这一基础层已经商品化。Otter.ai、Fathom 和 Granola 都在 2026 年各自推出跨会议 AI 聊天或智能体层——Otter 的 AI Chat Connectors、Fathom 的全账户 “Ask Fathom”、Granola 自己重建的智能体式 Chat(CP030);Granola 2026 年 5 月设计文章称,它选择让 Chat 高层读取最多 100 场会议,而不是完整读取 10 场会议的转写细节,并把这个架构取舍视为战略核心(CP031)。在信任上,Granola 自己的安全文档要求企业买方先申请 Trust Center 访问,才能看到 SOC 2 报告;Fireflies 和 Otter 则更直接地在公开营销和定价页展示合规声明(SOC 2/GDPR/HIPAA 徽章、HIPAA add-on)(CP032)。 在定价上,独立厂商标价集中在相似的 $0-$40 每用户每月区间——Granola 自己的销售指南称 Business 计划为 $14 每用户每月(CP050),对比 Fathom 的 $15-$34、Fireflies 的 $10-$39、Otter 的 $17-$30、Read AI 的 $15-$40(CP033)——而 Microsoft Teams Premium 以 $10 每用户每月压低所有厂商价格,Gong 按年计则高出一个数量级(CP033)。Otter 和 Fireflies 自己的定价页都显示相对标价有大幅促销折扣(CP036),说明价格竞争活跃,可能长期挤压 Granola 平直的 $14 标价。分发上,Microsoft、Google 和 Zoom 的功能都搭载在买方已经付费的协作套件订阅里;没有任何独立厂商,包括 Granola,可以在没有自身分发动作的情况下复制这种渠道优势(CP040)。[CP030, CP031, CP032, CP033, CP034, CP036]
| 购买标准 | Granola | Fathom | Fireflies | Otter.ai | 捆绑型既有厂商(Teams/Meet/Zoom) | Gong |
|---|---|---|---|---|---|---|
| 可用无 bot 采集选项 | 是 | 是(测试版) | unknown | 否(基于 bot) | n/a(原生平台功能) | 否 |
| 跨会议 AI 聊天 / 智能体记忆 | 是 | 是(“Ask Fathom”) | 部分(“AskFred”助手) | 是(AI Chat Connectors) | 部分(会议上下文内的 Copilot 聊天) | 是(Gong Assistant) |
| 公开 API/MCP 支持 | 是 | 是 | 是(Free 方案起开放 API 访问) | 是(MCP 服务器) | 部分(Graph/Copilot Studio API,不是会议专用 MCP) | unknown |
| 免费自助层级 | unknown | 是 | 是 | 是 | n/a(捆绑进付费套件) | 否 |
| 合规徽章直接发布在营销 / 定价页 | 否(需通过 Trust Center 申请) | 部分(Enterprise 提供 SSO/SCIM;未看到 SOC2 徽章) | 是(SOC2 Type II/GDPR/HIPAA 徽章) | 部分(列出 HIPAA 附加项;未看到醒目 SOC2 徽章) | 是(企业合规是核心卖点) | unknown |
| 发布自助标价(相对仅定制报价) | 是 | 是 | 是 | 是 | 部分(Teams Premium 是;Copilot/Gemini/Zoom 捆绑条款不同) | 否(仅定制报价) |
| 专用销售对话辅导 / 分析 | 否 | 部分(Business 层级辅导指标) | 部分(Business 层级团队分析) | 部分(“Assist”辅导附加项) | 否 | 是(核心产品) |
单元格按等级概括本章已审阅的 2026 年产品、定价与营销页面;来源组无法支持更明确判断的地方标为未知,而不是猜测。
[CP001, CP004, CP005, CP007, CP008, CP009]| 厂商 | 公开套餐 | 价格 / 单位 / 合同模式 | 包含能力 | 折扣或未知项 | 含义 |
|---|---|---|---|---|---|
| Granola | Business 方案(据 Granola 自有销售买方指南) | $14/用户/月 | 无限会议历史、CRM 集成、跨团队对话的文件夹级查询 | 本章来源组未核验 Free/Individual 层级定价和 Enterprise 合同条款 | 相较 Fathom/Fireflies/Otter/Read AI 付费层级处于中游,且明显高于 Microsoft $10 的 Teams Premium 附加包 |
| Fathom | Team / Business 层级 | $19/$15(Team)至 $34/$25(Business)/用户/月,月付 / 年付 | 无限 AI 摘要、CRM 同步、辅导指标、MCP 以及 Claude/ChatGPT(Business) | 直接宣传 25%+ 年付折扣 | 免费层级加透明标价,让 Fathom 在已审阅同业中拥有最宽的自助漏斗 |
| Fireflies.ai | Pro / Business / Enterprise 套餐 | $18/$10(Pro)至 $39(Enterprise,仅年付)/席位/月 | 200+ 项 AI Skills、CRM/ATS 集成、Enterprise 层级零数据留存政策 | Pro/Business 层级年付最高约 44% 折扣 | Pro 层级标价低于 Otter,Business 的单席区间大致匹配 Granola |
| Otter.ai | Pro / Business / Enterprise 套餐 | $16.99(Pro)至 $30(Business)/用户/月,观察到 $8.49-$24 促销价 | MCP 服务器、AI Chat Connectors、Enterprise 上 100 用户起的 SSO 要求 | Pro 和 Business 层级促销价约为标价一半 | 在直接同业中声称用户基数最大,但 Enterprise 层级增加了 Granola 不要求的席位下限 |
| tl;dv | 免费加付费层级(本会话未从 tl;dv 自有网站确认准确标价) | 独立估计约 $29-$98/用户/月 | 永久免费方案、GDPR/SOC2 徽章、多语言自定义摘要 | 本研究会话中,tl;dv 自有定价页返回 404 错误;数字依赖第三方估计 | 如果独立估计准确,tl;dv 即使有免费入口层,价格仍高于多数已审阅同业 |
| Read AI | Pro / Team / Enterprise 套餐 | $15/$19.75(Pro)至 $39.75(Enterprise,5+ 许可证)/用户/月 | Ada 数字孪生助手、高级集成、Enterprise 提供 HIPAA/SAML/SCIM | 标价组合内含约 24% 年付折扣 | 已审阅同业中最细的定价阶梯,但被硬性的 5 席 Enterprise 下限限制 |
| Gong | Foundations(另有 Engage/Forecast 附加项) | 独立估计为 $1,000-$1,600/用户/年,另加 $5,000-$50,000 强制年度平台费 | 会话智能、辅导、预测、CRM 集成 | Gong 不直接公布定价;所有数字均来自第三方或买方估计 | 按用户年费计,比任何独立记事本贵一个数量级;只有大型 SDR 组织才划算 |
| Microsoft Teams Premium / Copilot | Teams Premium 附加包或 Microsoft 365 Copilot 许可证 | $10.00/用户/月,按年计费(Teams Premium);Copilot 许可证单独定价,已审阅来源未完整披露 | 智能摘要、会中问答、水印、敏感度标签 | 本章来源组未直接检索到 Microsoft 365 Copilot 精确单席价格 | 标价低于所有独立厂商,但要求已有 Microsoft 365/Teams 订阅 |
| Zoom AI Companion | 包含在付费 Zoom Workplace 方案内,或单独提供 | 已审阅来源未找到单独项目价格;符合条件的付费方案无需额外成本 | 会议摘要、转录、应用内 Companion 聊天 | 已审阅来源未披露单独产品价格 | 现有 Zoom 客户的边际价格为零,竞争厂商单靠价格无法低于这个地板 |
| Google Workspace / Gemini | 包含在 Google Workspace 订阅内 | 已审阅来源未找到单独价格;捆绑进现有 Workspace 层级 | “帮我记笔记”、自动集成 Google Docs/Drive/Calendar | 本章来源组未核验 Gemini for Workspace 附加包定价历史 | 任何已经为 Google Workspace 付费的买家,都面临与 Zoom 相同的零边际价格动态 |
除标注为独立估计或公司估计外,数字均为 2026 年发布的标价 / 自助价格;实际企业合同价格、折扣以及 Granola 自身付费席位经济性均未公开,因此保留为缺口而不估算。
[CP007, CP008, CP009, CP010, CP012, CP013]直接软件同业中,Fireflies 和 tl;dv 直接发布最多合规与集成信号;Granola 则把自身合规文档放在门控后,还面临独立报道对其「默认私密」营销说法的反向证据。
该图单独从信任、合规、锁定视角切入,区别于 TP002 中更宽的功能与定价能力矩阵;若已审阅来源不足以支撑更明确判断,单元格标为未知。
[CP009, CP013, CP014, CP015, CP016, CP017]3.4 切换成本、锁定、多栖使用、分发权力与护城河耐久性
切换成本和锁定对 Granola 的影响不止一个方向。2026 年 3 月,Granola 决定加密本地笔记数据库,遭到一位知名投资人公开批评;一个独立博客也认为,这一变化让想要本地优先可携带性的用户面对一个“zero value”(零价值)app(CP037)——这种锁定风险与 Granola 自己“private by default”(默认私密)的营销主张并不协调(CP044)。独立 2026 年报道直接反驳了这一主张,发现 Granola 会把笔记分享给任何持有链接的人,并默认用转写稿训练,除非用户退出(CP045)。同一报道把批评延伸到更宽的类别,指出 Otter.ai、Fireflies.ai 和 Fellow 也有类似的宽松默认设置模式(CP046);另一个独立企业就绪度评测则因托管 / 数据驻留给 Granola 打出 3/10,Fireflies 和 Fathom 通过专用存储与自定义留存控制部分补上的缺口,Granola 已审阅页面尚未匹配(CP047)。多栖使用方面,Fathom、Fireflies 和 Otter 都发布了接入 CRM、项目管理和自动化工具的原生集成,降低了把结构化会议数据迁出单一厂商的成本(CP038);Otter 自己的 MCP client/server 设计也明确映射 Granola 的 MCP 策略,削弱了“智能体连接性”作为单独护城河的力度(CP039)。 耐久性方面,一份由竞争性 AI 会议助手厂商发布的 2026 年市场报告称,干净音频上的转写准确率已在全行业收敛到 95% 以上,并把市场分层为平台捆绑、独立和垂直专家三档;Zoom AI Companion 和 Microsoft Teams Premium 已按原始用户数领先,Fathom 则在独立同行中拥有最高披露 G2 评分(CP041、CP042)。TechCrunch 也另行观察到,竞争价值正从转写本身转向把笔记变成行动;其称 Read AI、Fireflies 和 Quill 正与 Granola 一起追逐这一方向(CP043)。这些都不致命,但意味着 Granola 自己的护城河不靠任何单一持久技术优势,而是同时取决于智能体功能、信任姿态和分发上的持续执行速度(CP048)。[CP037, CP038, CP039, CP040, CP041, CP042]
| 护城河主张 | 威胁 | 严重性 | 缓释措施 / 尽调问题 |
|---|---|---|---|
| 无 bot 本地采集是 Granola 核心差异化 | Fathom 和 tl;dv 现在都已推出无 bot 或“无需 bot”采集选项,Otter 基于 bot 的工作流反映的是选择,而非技术限制 | 高 | 要求 Granola 提供证据,证明无 bot 采集相较 bot 可选竞争对手,具体带来留存或赢单率优势 |
| Granola 的智能体式 Chat 和 Spaces/API/MCP 扩展,构成耐久的上下文层护城河 | Otter 2026 年 4 月重新发布已推出等价 MCP 客户端 / 服务器和 AI Chat Connectors,Fathom 也已提供账户范围“Ask Fathom”和 MCP 支持 | 高 | 索取 Granola 自身 Chat/Spaces 采用率与留存数据,并与对手智能体功能对比,检验护城河来自架构还是执行速度 |
| 既有厂商捆绑功能压住独立定价权 | Microsoft、Google、Zoom 都在已购买订阅中包含会议笔记 AI,且 Zoom 报告 AI Companion MAU 同比增长超过三倍 | 高 | 专门量化与捆绑式在位功能竞争时的胜负率,而不是只对比其他独立厂商 |
| Granola 的信任基础能撑住企业销售 | 一份独立评测将 Granola 在托管 / 数据驻留等企业就绪度上评为 3/10;报道显示,它「默认私密」的营销说法与实际默认设置之间存在落差 | 高 | 索取更新且注明日期的 SOC 2 报告,并要求其直接书面确认当前托管区域以及默认分享 / 训练设置 |
| Granola 的本地优先、加密笔记存储抬高了切换成本 | 一位知名投资人公开批评 2026 年 3 月本地数据库加密调整降低了数据可携性;独立博客也认为,这让这款应用对重视可携性的用户「价值归零」 | 中 | 确认 Granola 是否提供完整保真的导出路径,并跟踪加密调整后的流失信号 |
| 独立 AI 笔记是一个耐久、可单独收费的品类 | 一份厂商发布的 2026 年市场报告称,转写在全行业已以 95%+ 准确率商品化;TechCrunch 报道,多家竞争对手的竞争重心正在同时转向行动导向工作流 | 中 | 跟踪 Granola 自身的行动导向功能(Briefs、Recipes、CRM 推送)采用速度是否可衡量地快于 Otter、Fireflies 或 Read AI 的同类工作流功能 |
| Granola 能守住由 SMB / 创始人主导的客群,不被 Gong 式上市场竞争者挤压 | Gong 的成本结构只适用于大型 SDR 组织,但 Fathom、Fireflies、Otter 都已推出销售辅导 / 分析功能,从低端收窄这道差距 | 中 | 评估 Granola 是否计划自行加入销售辅导功能,还是有意把这一客群让给 Fathom/Fireflies/Otter |
本登记表聚焦最可能改变承销假设的耐久性问题,而不是逐项罗列所有战术层面的功能对战。
[CP021, CP022, CP023, CP025, CP026, CP037]3.5 展示材料
04财务情况
4.1 收入模式、定价与收入结构
Granola 通过分层按席位 SaaS 订阅变现。其博客称,Business 计划价格为 $14 每用户每月,包含无限会议历史、CRM 集成,以及跨团队对话的文件夹级查询(CI001)。SiliconANGLE 对 2026 年 3 月 Series C 的报道还补充了免费层和 $35 每月的 Enterprise 层,后者带有更扩展的网络安全控制;但报道从未说明 Enterprise 数字是按席位还是组织整体价格(CI002)——Granola 自己的 Enterprise 页面也印证了这种模糊性:它列出企业专属功能,却没有发布任何合同价格(CI003)。Series C 公告又增加了两个变现入口——打包进 Business/Enterprise 计划的个人 API,以及只打包进 Enterprise 的企业管理员 API——目前都未作为独立项目定价(CI004)。没有审阅来源披露收入中有多少来自 Business、Enterprise 或 API 使用(CI005)。 对比竞品组合定价,Granola 的 $14 标价处在 Fireflies、Otter 和 Fathom 共同覆盖的 $10-$39 每席位区间内(CI010、CI011、CI012),这些公司也都把自己的 Enterprise 层放在销售沟通之后。Granola 自己的销售买方指南明确把 ACV 高得多的会话智能类别让给 Gong,并引用第三方估算:Gong 每用户每年 $1,300-$1,600,另加 $5,000-$50,000 平台费(CI008);独立 CloudTalk 分析也佐证了这一数字(CI009)。Bloomberg 经 SiliconANGLE 报道称,Granola 在 Series C 前一个季度收入增长 250%,但两家媒体都没有披露这一增速对应的绝对基数(CI006);因此,收入质量方向上积极,却无法财务核验。[CI001, CI002, CI003, CI004, CI005, CI010]
| 收入流 | 机制 | 单位 | 当前数值 / 状态 | 质量 | 尽调问题 |
|---|---|---|---|---|---|
| Business 订阅 | 按席位 SaaS 订阅,自助购买或销售协助 | $/user/month | $14/user/month 标价 | 官方标价,不是实际收入 | 索取付费席位数,以及扣除折扣后的实际净价格 |
| Enterprise 订阅 | 按席位或固定费用的 SaaS 订阅,定制合同 | $/user/month(有争议) | 据 SiliconANGLE 约 $35/month(按席位还是固定费用未确认) | 第三方估算,单位含糊 | 确认 Enterprise 的准确计价单位,以及是否收取平台 / 管理费 |
| 免费层 | 免费增值获客漏斗 | n/a(未变现) | 已审阅来源未披露限制 | 不直接变现 | 确认免费层功能 / 用量限制,以及免费转付费转化率 |
| 个人 API 访问 | Business / Enterprise 方案捆绑附加功能 | 包含在内,不单独定价 | 未披露单独价格 | 变现方式未知 | 询问 API 访问未来是否会变成按量 / 付费附加项 |
| Enterprise API 与管理员分析 | 仅随 Enterprise 捆绑的附加功能 | 包含在 Enterprise 方案中 | 未披露增量价格 | 变现方式未知 | 确认是否计划推出按用量计费的 API 定价 |
| 专业服务 / 上线导入 | 已审阅来源均无证据 | n/a | 没有服务收入线证据 | 证据缺口 | 直接询问公司是否存在实施或服务收入 |
Business 方案价格来自公司一手来源;Enterprise 价格是单一第三方估算,单位不清;免费层、API 和服务行都是证据缺口,不是已确认的空值。
[CI001, CI002, CI003, CI004, CI005]| 厂商 | 价格 / 单位 / 合同 | 标价与实收 | 折扣 / 未知项 | 来源 |
|---|---|---|---|---|
| Granola Business | 每用户每月 $14 | 仅标价 | 未披露实收 / 折后价格 | Granola 博客(官方) |
| Granola Enterprise | 约 $35/月(单位含糊) | 第三方估算,非标价 | 按席位还是固定费未确认;无折扣数据 | SiliconANGLE(第三方) |
| Fireflies.ai | Pro $10、Business $19、Enterprise $39/席位/月(年付) | 标价 | 未披露月付价格和企业折扣 | Fireflies.ai 定价页 |
| Otter.ai | Pro $16.99(促销 $8.49)、Business $30(促销 $24)/用户/月 | 标价 + 促销价 | 未披露促销价持续时间 | Otter.ai 定价页 |
| Fathom | Team $19/$15、Business $34/$25/用户/月;Enterprise 定制 | 标价;Enterprise 需咨询 | Enterprise 价格完全未披露 | Fathom 定价页 |
| Gong(相邻品类,非直接竞品) | 约 $1,300-$1,600/用户/年 + $5,000-$50,000 强制平台费 | 第三方买家估算 | Gong 不公开定价;估算随谈判而变 | Granola 买家指南 + CloudTalk(独立) |
所有数字都是标价或第三方估算;都不代表实收、折后或谈判后的价格。Gong 只作为相邻品类价格锚点纳入,不是 Granola 事实。
[CI001, CI002, CI010, CI011, CI012, CI008]一个 Granola 用户如何转化为已确认订阅收入,并在扣除第三方转录和 LLM 推理成本后,最终形成未披露的毛利润。
节点中的美元数字来自标价或单一第三方估计,不是实测收入;Granola 未披露收入成本和毛利润数据,因此相关节点为定性判断。
[CI001, CI002, CI022, CI023, CI024]4.2 GTM 动作与销售效率代理指标
Granola 的市场进入动作是自下而上、产品驱动,而不是销售驱动。Brex 的公开案例最能说明:实施先从部分团队试用 Business 计划开始,随着使用自发扩散,数周内转为完整 Granola Enterprise 部署(CI013);Vercel 案例则描述了整个 GTM 团队采用产品,而不是只有一名内部倡导者(CI014)。Index Ventures 自己的投资文章把这写成其支持的增长机制:个人用户变成传播者,采用以草根方式从个人扩散到团队,再扩到整个企业(CI017);Granola 招聘页也呼应这一叙事,称产品“从领导者快速扩散到整个组织”,同时每天转写数百万分钟对话(CI018)。 这些信号没有一个以美元计量。无论公司发布还是第三方来源,都没有披露 Granola 的 CAC、LTV 或 CAC 回本周期(CI019)。最接近的标尺来自外部:一份 2026 年 SaaS 基准报告称,PLG / 自下而上 SaaS 公司的 CAC 回本周期中位数为 12-15 个月,头部四分位约 9-12 个月,超过 24 个月则被标为风险信号(CI020);同一报告还指出,风投支持 SaaS 的销售与营销开支占收入比例中位数,已从 2021 年超过 50% 降至 2024-2026 年约 35-40%(CI021)。这些只是竞品组合基准;它们并不能说明 Granola 自身获客支出或效率在该区间之上、之内还是之下。客户案例量化的是节省时间(260-830+ 小时),不是节省成本或支付金额(CI015、CI016)。[CI013, CI014, CI017, CI018, CI019, CI020]
基于客户案例研究推断的获客到扩张路径;Granola 没有实测 CAC、转化率或 LTV 输入。
所有节点均为定性,并按公开客户案例研究推断的 GTM 逻辑排序,不是实测 CAC、转化率或 LTV 输入;本图中的任何数字都不是经过验证的 Granola 指标。
[CI017, CI013, CI014, CI019]4.3 成本结构与毛利驱动
Granola 自己的安全页面是本章最重要的成本结构披露:它点名 Deepgram 和 AssemblyAI 为第三方转写提供商,OpenAI 和 Anthropic 为用于转写和总结会议的第三方 AI 提供商(CI022)。由于这两类输入是授权采购,而不是建立在自有基础设施上,Granola 的收入成本很可能随使用量联动——跟着会议量和笔记长度增长——而不是像更简单 SaaS 产品那样主要是固定托管成本(CI023)。除此之外,Granola 要求企业买方采用 SSO/SAML、SCIM 账号配置、管理员仪表盘和加密,并称自己持有 SOC 2 Type II 认证(CI026);这些都代表持续的安全工程投入,叠加在核心转写 / 推理成本基础上。2026 年 3 月,Granola 加密本地笔记数据库的决定因打断第三方智能体工作流而遭 a16z 合伙人公开批评,又增加了尚未量化的工程和信任修复成本(CI031)。 没有来源披露 Granola 自己的毛利率、收入成本或运营费用拆分;TheNextWeb 的 Series C 报道直接称,Granola 没有公开披露收入数字、用户数或留存指标(CI027)。外部基准显示,SaaS 订阅毛利率中位数约 76%(头部四分位 84%+),但也明确提示,计算或 AI 推理成本重的产品常低于 70%(CI024)——这正切中 AI 转写产品的结构。Zoom FY2026 10-K 报告毛利率从 75.8% 提升到 77.0%(CI025),只能作为成熟上市协作软件对比,而不是 Granola 数据点。结构上,Companies House 确认 Granola Labs Ltd 已为 FY2024 和 FY2025 提交 “total exemption full accounts”,这是英国小公司豁免,取消了近期发布营业额或利润的任何法定义务(CI028);因此,这种不透明应被理解为结构性申报状态,而不是公司刻意回避。[CI022, CI023, CI026, CI031, CI027, CI024]
| 指标 | 数值 / 空值 | 置信度 | 为什么重要 | 尽调问题 |
|---|---|---|---|---|
| 收入 / ARR 年化规模 | 低(未披露) | 锚定所有其他单位经济比率;没有它,增长率说法无法定量。 | 直接索取过去 12 个月收入和 ARR。 | |
| 毛利率 | 低(未披露;可比基准 70%-84%) | 决定业务在承受 LLM 推理成本敞口时,能否盈利扩张。 | 索取按供应商拆分的收入成本(转写 / LLM 推理 / 托管)。 | |
| CAC | 低(未披露) | 评估自下而上 / 有机增长是否真的像表面那样便宜,必须有这个指标。 | 索取按季度拆分的付费营销支出和新增客户数。 | |
| LTV / CAC 回收期 | 低(未披露;PLG SaaS 可比基准为 12-15 个月) | 决定 GTM 动作的资本效率。 | 索取随时间变化的队列级单账户收入。 | |
| 净收入留存率 | 低(未披露) | 判断 ROI 案例是否能转化为耐久的扩张收入。 | 索取过去四个季度的队列留存和扩张数据。 | |
| 员工数(现金消耗代理) | 116 | 中(Tracxn 估算,2026 年 5 月) | 在未披露现金消耗率时,这是运营成本规模的最佳可得代理。 | 索取公司确认的员工数快照。 |
| 烧钱倍数 | 低(未披露;$1M-$10M ARR 队列的可比基准中位数约 1.5x) | 反映增长支出相对于新增净收入的资本效率。 | 索取过去四个季度的净现金消耗和新增净 ARR 数字。 | |
| CAC 回收期(仅可比基准) | 9-24 个月 | 低(外部 SaaS 基准,非 Granola 专属) | 在没有 Granola 专属数字时,提供外部效率标尺。 | CAC/LTV 数据可得后,用实测 Granola 数字替换。 |
除员工数外,上述每个指标对 Granola 本身都是空值;可比基准数字(CAC 回收期范围、烧钱倍数)仅作为外部 SaaS 背景引用,来源为 2026 年行业基准报告,不应解读为 Granola 事实。
[CI019, CI024, CI020, CI021, CI034, CI037]4.4 公开牵引力与私人指标缺口
Granola 的公开牵引力代理指标偏定性、基于参考客户,而不是量化。招聘页称产品每天转写数百万分钟对话(CI018);Series C 文章和 Kleiner Perkins 文章共同点名 Asana、Vanta、Gusto、Thumbtack、Cursor、Lovable、Decagon 和 Mistral AI 为客户,并有 Brex、Vanta、Zapier、Vercel 的专门 ROI 案例(CI036)。这些案例都没有披露席位数、合同金额或续约 / 扩张率。历史上,一份 2026 年报告称,Granola 在 2024 年 10 月 Series A 时约有 5,000 名周活用户,但截至 2026 年中没有披露当前用户数(CI032);同一报告还指出,Fireflies.ai 已独立超过 16 million 用户和 $1 billion 估值——这是一个 Granola 自身披露既无法匹配也无法反驳的规模对比(CI033)。 即使第三方数据供应商也补不上缺口:Tracxn 公开资料估计,截至 2026 年 5 月公司有 116 名员工(CI034),但把 Granola 的收入和逐轮估值倍数字段挡在付费数据墙后,意味着专门研究平台也没有免费露出 Granola 的收入或估值倍数(CI035)。TheNextWeb 直接称 Granola 没有公开披露留存指标;本章审阅来源中,也没有任何来源披露净收入留存、logo 留存或流失率(CI037)。结果是,牵引力图景完全建立在 logo 质量和节省时间的轶事上,而不是任何可测量的使用、收入或留存趋势。[CI018, CI036, CI032, CI033, CI034, CI035]
在 Granola 相关财务输入未被实测时,区间展示外部 SaaS 可比组基准或不同出版方报道的跨度,而非 Granola 特定数字。
前两行是外部 SaaS 行业可比基准,仅供背景参考,不是 Granola 的实测数据;第三行反映独立出版方对 Granola 自身累计融资总额确实存在分歧。
[CI020, CI024, CI025, CI040]4.5 资本充足性与融资依赖
公司概况章节记录了 Granola 的完整融资时间线;本章只重述评估未来资本充足性所需的事实。公司于 2026-03-25 宣布完成 $125 million Series C,由 Index Ventures 领投,Kleiner Perkins 与老股东 Lightspeed、Spark、NFDG 参投,投后估值 $1.5 billion(CI038)。TechCrunch 和 The SaaS News 都称融资后累计融资 $192 million,而 BusinessCloud 报道的则约为 $188 million——这是发布方之间围绕 Granola 自身累计总额的小但真实差异(CI040)。资金用途只被宽泛说明:StartupRise 报道只给出“扩展产品并支撑持续增长”这类通用表述(CI041);Granola 自己的 Series C 文章则围绕发布 Spaces、API、更新版 MCP server 和管理员使用可见性来描述融资,而不是给出数字化预算拆分(CI042);Bloomberg 另报道称 Granola 计划继续招聘,这与招聘页开放职位一致(CI043)。 没有来源披露现金余额、月度烧钱速度或现金跑道(CI044)。作为一家提交豁免小公司账目的英国私人公司,Granola 近期没有法律义务公开这些指标。审阅到的 Companies House 申报历史列出确认声明、账目、PSC 和高管申报,但没有登记的有担保债务押记;这与迄今为止全股权资本结构一致,尽管这反映的是申报历史摘要,而非专门调取押记登记册(CI045)。Granola 及其投资人都没有公开下一轮触发条件;Index Ventures 和 Kleiner Perkins 的文章转而把本轮描述为多年上下文层建设(CI046)。这种表述并没有阻止公开质疑:TheNextWeb 指出,在没有披露收入、用户或留存的情况下,估值十个月内跳升六倍(CI047);一家声誉较低的媒体 BritishWire 又放大了互联网泡沫类比,质疑 Granola 最终收入能否支撑十位数估值(CI048)。[CI038, CI040, CI041, CI042, CI043, CI044]
| 账面现金 | 月度现金消耗 | 现金续航(月) | 计划资金用途 | 下一轮触发因素 | 债务 / 项目融资义务 |
|---|---|---|---|---|---|
| 未披露(Series C 总融资 $125M 于 2026-03-25 完成;净留存现金未知) | 未披露 | 未披露 | Spaces、个人 / 企业 API、更新版 MCP、管理员用量仪表盘,以及持续招聘(据公司 Series C 文章和 Bloomberg 来源的招聘报道);未发布数字化预算拆分 | 未说明;投资人评论把本轮融资定位为多年期企业转型建设,而非近期追加融资 | 已审阅 Companies House 备案历史中未发现债务或担保抵押登记;披露的四轮(种子轮、A、B、C)均为股权融资 |
| 未披露(截至 2025 年 3 月 31 日的财年账目按小公司总豁免提交) | 未披露 | 未披露 | n/a(法定备案不覆盖资金用途) | n/a | 截至已审阅的最近一次确认声明(2026 年 3 月 2 日),Granola Labs Ltd 无抵押登记 |
第 1 行反映公司 / 媒体来源给出的 Series C 后运营图景;第 2 行反映 Companies House 的英国法定备案图景。两行的现金 / 现金消耗 / 现金续航都为空,因为两类来源均未披露;资金用途和债务单元格是最佳可得代理,不是已确认数字。
[CI038, CI042, CI041, CI043, CI044, CI028]| 缺失的私有指标 | 影响 | 精确尽调路径 |
|---|---|---|
| 绝对收入 / ARR | 无法量化报道中的 250% 季度增长率,也无法对照 $1.5B 估值做基准。 | 在签署 NDA 的数据室中索取过去 12 个月收入和月度 ARR 桥接表。 |
| 毛利率 / 收入成本拆分 | 无法判断 LLM 推理和转写供应商成本是否会把利润率压到 SaaS 中位数以下。 | 索取过去四个季度按供应商拆分的收入成本明细(Deepgram/AssemblyAI、OpenAI/Anthropic 支出)。 |
| CAC / LTV / 回收期 | 无法验证自下而上 GTM 是资本高效,还是只是增长快。 | 索取付费营销支出、新增客户数和队列收入曲线。 |
| 净收入留存 / 流失 | 无法确认 ROI 案例是否会转化为耐久的扩张收入。 | 索取过去四个季度的队列级留存和扩张数据。 |
| 按层级拆分的付费席位数 | 无法把标价换算成实际收入估算。 | 索取当前付费席位数,并按 Business 和 Enterprise 层级拆分。 |
| 现金余额和现金续航 | 无法判断 Series C 资金预计能撑多久,或是否急需追加融资。 | 索取最新管理账和董事会批准的现金续航模型。 |
| 月度现金消耗率 | 无法独立核对由招聘推动的成本增长是否匹配收入增长。 | 索取过去各月实际运营支出。 |
| Enterprise 合同经济性(ACV、折扣深度) | 无法确认 SiliconANGLE 报道的 $35/month Enterprise 数字是否现实,还是底价 / 上限估算。 | 索取 3-5 份经脱敏的 Enterprise 订单,显示席位数和年度合同价值。 |
所有行都描述本章已审阅的所有来源中缺失的指标,包括 Granola 自有网站、投资人文章、媒体报道和 Companies House 备案。
[CI005, CI027, CI019, CI037, CI036, CI044]Granola 的资本结构只有股权,在传统资本强度上偏低(未披露债务或重资本开支),但资本依赖度高,因为现金余额、烧钱速度和现金续航期均未披露。
节点 k2 和 k5 属于结构性未披露,并非估算值;该图展示依赖逻辑,不代表实测现金流金额。
[CI049, CI045, CI022, CI034, CI043, CI044]4.6 财务结论
合在一起,Granola 的财务图景是:商业牵引力真实且快速增长,但几乎没有可验证的财务披露可供承销。收入质量方向上看起来强——Bloomberg 称季度增长 250%,Brex 也有具体的 Business 到 Enterprise 升级——但绝对金额无法验证,因为公司、投资人或任何审阅到的第三方数据供应商,包括付费墙后的 Tracxn 资料,都没有披露以美元计的收入、ARR 或席位数(CI050)。毛利率路径是真正的开放问题,而不是理所当然的 SaaS 结果:Granola 依赖第三方 LLM 推理和转写供应商,这与竞品组合基准所指出的低于 70% 毛利率成本模式一致;作为对照,订阅 SaaS 中位数到头部四分位区间为 76%-84%,Zoom FY2026 可比值为 77.0%(CI051)。 从传统意义看,资本强度低——没有披露债务,没有重资本开支基础设施——但资本依赖高:不到三年完成四轮股权融资,刚刚完成 $125 million 轮融资,却没有披露现金余额、烧钱速度或现金跑道来说明这笔资本预计能支撑多久(CI052)。核心尽调阻碍是结构性的,而不是公司沉默寡言:Granola Labs Ltd 的英国小公司审计豁免意味着,普通未来 Companies House 文件不会披露营业额、毛利或利润;因此,承销这家公司需要谈判进入资料室,而不是等待公开文件(CI053)。在获得这些访问之前,任何贴在 Granola 身上的收入、毛利或估值倍数数字,都应被视为估算或第三方主张,而不是事实。[CI050, CI051, CI052, CI053]
4.7 展示材料
05产品与技术
5.1 产品概览与核心模块 / SKU 图谱
Granola 是一款会议记忆应用,差异化集中在一个机制:它在用户自己的设备上本地采集麦克风和系统音频,而不是以可见机器人身份加入通话;会议开始前,用户必须手动启动应用(CE001, CE002)。这层采集能力如今支撑的是更宽的产品线,而不再只是单一记笔记工具。AI Notepad 本体覆盖 macOS、Windows 和 iOS;其上叠加了可在用户完整会议历史中答问的 agentic Granola Chat、自动生成会前准备的 Briefs、Recipes(可复用 AI 提示词模板)、用于共享组织记忆的 Team Spaces/Folders、Personal API 和独立的 Enterprise Admin API,以及 Granola MCP——一个把会议上下文暴露给外部 AI 工具的集成层(CE003)。Granola for Enterprise 在核心产品之上增加的是治理层:SSO/SAML、SCIM 配置、管理员仪表盘、域名管理、灵活数据留存窗口和 SOC 2 Type II 文档,而不是另一套产品架构(CE004)。 商业化上,Granola 以免费层加付费 Business、Enterprise 计划分发,主要靠会议历史深度、CRM 集成、管理员控制等功能权限区分(CE006);Personal API 随 Business 和 Enterprise 捆绑,Enterprise Admin API 仅 Enterprise 可用,二者都未作为独立项目定价(CE007)。模块图谱里有两个明显缺口。第一,Granola 自己的招聘页面列出了 Android Engineer,但截至本次报告运行日,Granola 的站点地图和所有公开应用商店都看不到 Android 应用——这是清晰路线图信号,却没有承诺发布日期(CE005)。第二,评审时 Granola 自己的 /integrations 营销 URL 返回 HTTP 404,因此没有一个官方、最新页面能完整列出公司的集成表面(CE008)。[CE001, CE002, CE003, CE004, CE006, CE007]
| 模块 / 资产 | 主要用户 | 状态 / 成熟度 | 差异化 | 尽调缺口 |
|---|---|---|---|---|
| AI Notepad 应用(macOS、Windows、iOS) | 个人会议参与者 | 成熟——自 2024 年 7 月(桌面)、2025 年 5 月(iOS)、2025 年 7 月(Windows)起为核心产品 | 无机器人本地音频捕获;通话中没有可见参会者 | 没有独立准确率基准(CE048) |
| Granola Chat(智能体式) | 个人 / 团队,跨会议问答 | 2026 年 4 月重构为智能体式;仍在快速演进 | 设计上明确权衡完整转写与浓缩笔记的 token 成本(CE024) | 没有公开答案准确率或幻觉率基准 |
| Briefs | 个人,会前准备 | 2026 年 5 月上线;较新 | 基于日历 + 过往笔记 + 可选 Gmail 上下文自动生成 | 未披露使用采用率或准确率数据 |
| Recipes | 个人 / 团队,可复用提示词 | 2025 年 9 月上线 | 专家编写的模板市场(Lenny Rachitsky、Matt Mochary 等) | 未披露 Recipes 专属使用或留存数据 |
| Team Spaces / Folders | 团队 / 组织 | 已上线,并在 2026 年 3 月 Series C 前后扩展 | 把组织记忆从个人笔记扩展到共享层 | 未披露组织级采用指标 |
| Personal API + Enterprise Admin API 接口 | 开发者 / IT 管理员 | 2026 年 3 月宣布;未单独定价 | 将笔记访问捆绑给智能体 / 自动化场景 | 未发布用量、速率限制或 SLA |
| Granola MCP | 外部 AI 工具(Claude、ChatGPT、Cursor) | 2026 年 2 月推出;Enterprise 访问为早期访问测试版,默认关闭 | 采用开放 Model Context Protocol 标准,而非只走专有 API | 高级访问控制和用量监控明确尚未上线(CE033) |
| Granola for Enterprise(管理层) | IT / 安全管理员 | 成熟——自 2025 年 7 月起具备 SOC 2 Type II | SSO/SAML、SCIM、管理员仪表盘、留存控制、水印 / 通知透明度 | 无 HIPAA BAA,也无 EU / 本地部署托管选项(CE041) |
| Android 应用 | Android 用户 | 未上线——团队名单有 Android 工程师职位,但没有产品或日期 | n/a | 没有承诺发布日期(CE005、CE032) |
状态 / 成熟度和差异化来自 Granola 自有产品页和更新日志;尽调缺口标注该模块没有独立指标或公司披露指标的位置。
[CE003, CE004, CE006, CE007, CE025, CE036]5.2 工作流与用例证明
Granola 自己发布的客户案例最清楚地展示产品在组织内的真实用法,覆盖四类不同买方 / 用户场景。Zapier 的 Principal Designer 在设计 / 产品 / 工程组织中采用 Granola,原因是其他 AI 转写工具显得侵入感强;公司发布的案例称每月节省 830+ 小时,组织级采用率 52%(CE009)。Vanta 的 VP of Customer Success 将 Granola 用于内部 1:1、站会和跨职能同步,明确因为 SOC 2 Type II 认证、加密和通知同意式录音控制对安全与合规公司很重要;案例称组织级采用率 69%,会议可见性提升 61%(CE010)。Brex 的 Head of IT 在员工已自发采用后把 Granola 推向全公司,强调受监管 fintech 场景里的 SSO/SCIM 交付和不用于公开模型训练的承诺;案例称六个月内知识员工采用率超过 50%,每周节省 427 小时(CE011)。Vercel 的销售工程团队用 Granola 替换了更慢的转写工具,现在用其 “chat with the transcript” 功能检索潜在客户提出的具体技术异议;案例称 GTM 团队采用率 100%,每位用户每周节省 11 小时,GTM 生产力提升 25%(CE012)。 除这四个旗舰客户外,Granola 自己的报道点名高管猎头公司 Daversa Partners 在 150 名员工中的 136 人部署 Granola,原因是无机器人、设备级采集比可见机器人工具更适合保密候选人与客户对话,侵入感更低(CE013);独立 Product Hunt 社区帖还提到面对面、多语言对话采集——例如把日语对话转成英文笔记——是移动端应用一个突出的真实用例(CE014)。不过,四个旗舰案例都没有披露席位数、合同金额、续约率或流失数据;所有量化收益都停留在采用率和节省小时数估算,而不是与收入挂钩的证明(CE015)。这是本章工作流证据的核心限制。[CE009, CE010, CE011, CE012, CE013, CE014]
| 用户任务 | 既有工作流 | Granola 方案 | 可衡量收益(公司发布) | 限制 |
|---|---|---|---|---|
| 设计 / 产品 / 工程站会和 1:1(Zapier) | 每周 30+ 场会议靠手工记录;没有共享事实源 | 无机器人桌面捕获 + 共享团队文件夹 | 每月节省 830+ 小时;全组织采用率 52% | 未披露席位数、合同金额或留存数据 |
| 客户成功 1:1、站会、同步会(Vanta) | 实时手记还要人工对照原始转写 | 无机器人捕获 + SOC 2 / 加密 / 同意控制 | 全组织采用率 69%;会议可见度提升 61% | 未披露流失或续约数据 |
| 全企业 IT 推广(Brex) | 手工与部分 AI 笔记混用;外部会议未覆盖 | 全公司 Enterprise 部署,配 SSO/SCIM | 6 个月内采用率超过 50%;节省 25k+ 小时;每周 427 小时 | 未披露每席位成本或以美元计的 ROI |
| 销售工程发现电话(Vercel) | 原转写工具更慢,UX 笨重,处理延迟 | 用「Chat with the transcript」回忆技术异议 | GTM 采用率 100%;每用户每周节省 11 小时;生产率提升 25% | 未量化胜率或交易推进速度影响 |
| 保密高管猎聘电话(Daversa Partners) | 候选人 / 客户通话中,手工或机器人工具都显得打扰 | 无机器人设备级捕获 | 150 名员工中有 136 人使用 Granola | 没有与采用率绑定的结果或成单率指标 |
| 线下多语对话(个人用户) | 没有结构化方式捕获并翻译线下非英语对话 | 移动应用捕获外语音频,并输出英文笔记 | 社区轶事性好评(Product Hunt);没有实测准确率 | 单条社区帖子,不是公司验证的基准 |
所有可衡量收益数字均来自 Granola 自己发布的客户案例,或其自有社区 / 营销渠道;均未经独立审计。
[CE009, CE010, CE011, CE012, CE013, CE014]5.3 技术与运营架构
Granola 桌面客户端是为 macOS 和 Windows 打包的 Electron 应用——这一点既有 Granola 自家工程博客谈到定制 dock / taskbar 图标的难度,也有安全研究公司 Tenable 拆包 macOS .asar 档案查看客户端 JavaScript 的独立确认(CE016)。核心 AI 管线完全依赖供应商:转写由第三方 ASR 提供商 Deepgram 和 AssemblyAI 执行,笔记增强、聊天和摘要则调用第三方 OpenAI 和 Anthropic LLM API;这是公司公开说明的架构,也得到 Tenable 在客户端中发现 AssemblyAI 引用和实时 API key 的佐证(CE017)。同一份 Tenable 研究发现了具体漏洞:一个未经认证的 api.granola.ai 端点(“get-feature-flags”)返回实时 AssemblyAI key,可被用来直接从 AssemblyAI 自己的 API 拉取会议转写 ID 和内容(CE018);Granola 后来的事后复盘确认,2024 年 11 月至 2025 年 3 月期间约 300-333 名 iOS TestFlight beta 测试者受影响,除研究者测试外没有被利用的证据(CE019)。 在设备端,Granola 将会议元数据、转写文本和 AI 摘要存入本地应用缓存文件(cache-v3.json),内部是嵌套的 “double-JSON” 结构;一个独立开源 Python 项目完全离线逆向了该结构,没有调用 Granola 自家服务器(CE020)。同一项目还记录了另一条非公开路径:从本地 supabase.json 配置文件提取 session token,直接调用 Granola 云端 API;维护者出于隐私和可靠性考虑没有实现(CE021)。2026 年 3 月,Granola 加密了本地数据库,一位知名 a16z 合伙人公开批评(该 tweet 约 327,000 次浏览)此举切断了此前可直接读取明文缓存文件的第三方 agent(CE022);Granola 联合创始人指向 Granola MCP 作为替代路径,但批评者反驳称 MCP 需要支持 MCP 的特定客户端,而不是开放文件或 API 访问,底层争议仍未解决(CE023)。Granola 自家工程博客用 2026 年 4 月 agentic Chat 重构背后的取舍说明了同类问题:读取完整转写比读取压缩笔记大约多消耗 10x token,跨会议 agent 查询必须在深度和成本之间明确取舍(CE024)。Granola MCP 于 2026 年 2 月 4 日发布,实现开放的 Model Context Protocol,让 Claude、ChatGPT、Cursor、Claude Code 等工具能把用户笔记作为上下文查询;但 Enterprise MCP 访问仍是默认关闭的 early-access beta(CE025)。已存储数据——转写、笔记、AI 摘要——位于 AWS 托管、美国区域的 Virtual Private Cloud,静态和传输中加密,每日备份;原始音频完全不存储(CE026)。认证通过绑定用户日历的 Google Workspace 或 Microsoft OAuth 完成(CE027)。[CE016, CE017, CE018, CE019, CE020, CE021]
| 层 / 组件 | 角色 | 依赖 | 风险 |
|---|---|---|---|
| 桌面客户端(Electron、macOS/Windows) | 本地音频捕获、UI、笔记编辑 | Electron 框架;由 .asar 拆解确认 | 客户端代码可解包检查,Tenable 研究已证明这一点 |
| iOS 客户端(原生) | 移动 / 线下 + 电话捕获 | Apple App Store 分发 | 分发完全受 Apple 审核 / 政策卡口约束(CE052) |
| 转写流水线 | 实时语音转文字 | Deepgram、AssemblyAI(第三方 ASR 供应商) | 2024-2025 年,供应商 API 密钥曾通过未认证端点暴露(CE018) |
| AI 推理 / 摘要 | 笔记增强、Chat、Briefs、Recipes | OpenAI、Anthropic(第三方 LLM API) | 核心产品质量和成本绑定外部模型供应商,而非自有 IP |
| 存储层 | 转写 / 笔记持久化 | AWS 托管的美国区 VPC,静态 / 传输中加密,每日备份 | 仅美国托管;未发现 EU / 本地部署选项(CE012) |
| 本地设备端缓存(cache-v3.json) | 离线访问会议元数据 / 转写稿 / 摘要 | 专有嵌套 JSON 格式;已被独立逆向解析 | 自 Mar 2026 起加密,切断了此前第三方 agent 的读取权限(CE022) |
| 认证层(Google Workspace / Microsoft OAuth) | 登录 + 日历同步 | Google / Microsoft 身份平台 | 三份已披露事后复盘中有两份追溯到 Google Workspace 账户状态边界场景(CE044、CE045) |
| Granola MCP 服务器 | 把笔记作为上下文暴露给外部 AI 工具 | Model Context Protocol 开放标准 | Enterprise MCP 仍处早期访问 beta;且发生过一次有日期记录的中断(CE030) |
本表汇总 Granola 自身安全页面和博客、Tenable 的独立安全研究文章,以及一个逆向解析本地缓存格式的第三方开源客户端。
[CE016, CE017, CE018, CE026, CE020, CE022]Granola 的技术栈从本地设备采集出发,经第三方 ASR/LLM 推理,到加密云存储和代理集成层,外面包着企业管理 / 控制层。
各层综合自 Granola 自有安全 / 企业页面、工程博客、独立安全研究拆解和第三方开源客户端,并非来自单一官方架构图。
[CE016, CE017, CE026, CE020, CE025, CE003]Granola 的产品依赖第三方 ASR/LLM 供应商、仅限美国的云托管、应用商店分发和 Google/Microsoft 身份平台——其中两个已产生已披露的安全复盘。
依赖边来自 Granola 自有安全 / 企业页面、MCP 和更新日志文章,以及独立的 Tenable 安全研究;地图只覆盖外部可见依赖。
[CE017, CE018, CE026, CE027, CE029, CE025]5.4 部署、集成、可靠性、支持与路线图
Granola 的公开发布历史显示,自 2024 年 7 月桌面端上线后,功能交付几乎按月推进:伴随原始 notepad 的 $20M 融资(2024 年 11 月)、原生 iOS 可用(2025 年 5 月)、Windows 可用并完成 SOC 2 Type II(2025 年 7 月)、Recipes(2025 年 9 月)、Granola MCP(2026 年 2 月)、$125M Series C(2026 年 3 月)、agentic Chat 重构(2026 年 4 月)以及 Briefs(2026 年 5 月)(CE028)。集成表面同步扩张:Slack 发帖(2024 年 10 月)、Notion 分享(2025 年 3 月),以及 Zapier 将 Granola 连接到 8,000+ 个应用并接入 Attio 和 HubSpot(2025 年中后期);这些都记录在 Granola 自己的 changelog 中(CE029)。 但可靠性可见度很薄,且主要来自第三方。一个独立 uptime 监控聚合器记录了运行日前 90 天内 Granola 的 4 起事故,中位解决时间约 2 小时 34 分钟,包括一次 iOS 转写中断(2026 年 6 月 25 日)、一次因 “broken deployment” 导致的 Granola Chat 中断(2026 年 3 月 10 日),以及——值得注意的是——一次 Claude 用户无法连接 Granola MCP 的独立事故(2026 年 3 月 9 日)。这是 MCP 集成层存在自身运营故障模式的第一条直接证据(CE030)。审阅过的来源没有披露 Granola 发布的 uptime SLA 或公司维护的公开状态页;目前唯一事故可见性来自这个非关联第三方监控器(CE031)。部署侧还存在两个路线图缺口:尽管团队名册中有 Android Engineer,Android 应用仍未发布,也没有日期(CE032);Granola 自己的 MCP 公告明确把“高级访问控制、使用监控与分析、更高速率限制、灵活定价选项”标为待推出,而不是已交付的 Enterprise MCP 能力(CE033)。平台推出本身也是顺序推进而非同步——iOS 于 2025 年 5 月(CE050)、Windows 于 2025 年 7 月(CE051)——当下 iOS 分发完全依赖 Apple App Store 审核和政策环境,这是识别出的唯一公开 iOS 渠道(CE052)。[CE028, CE029, CE030, CE031, CE032, CE033]
| 日期 | 功能 / 里程碑 | 状态 | 含义 |
|---|---|---|---|
| Jul 2024 | Granola 桌面端发布(面向会议的 AI 笔记本) | 已上线 | 确立「无 bot 采集」作为创始产品命题 |
| Nov 2024 | $20M 融资 + 原版 AI 笔记本同步宣布 | 已上线 | 把早期资本直接绑到核心产品叙事上 |
| May 2025 | 原生 iOS 应用发布 | 已上线 | 把采集从桌面扩展到线下 / 电话会议 |
| Jul 2025 | Windows 可用 + SOC 2 Type II 认证 | 已上线 | 两个企业就绪里程碑同月落地 |
| Sep 2025 | Recipes(可复用、由专家撰写的 AI 提示模板) | 已上线 | 增加竞品普遍缺少的内容 / 社区层 |
| Feb 2026 | Granola MCP 发布(Model Context Protocol 集成) | 已上线;Enterprise MCP 处于早期访问 beta | 把 Granola 上下文开放给外部 AI 工具,但 Enterprise 只开放了一部分 |
| Mar 2026 | 本地笔记数据库加密 + $125M Series C | 已上线 | 安全加固与外界批评第三方 agent 访问丢失同时发生(CE022) |
| Apr 2026 | Granola Chat 重构为 agentic 形态 | 已上线 | 转写稿与笔记之间的 token 成本取舍被明确写进产品设计 |
| May 2026 | Briefs(自动生成会前准备) | 已上线 | 产品从被动笔记延伸到主动式 agentic 会前准备 |
| 未注明日期(截至运行日) | Android 应用 | 仅在路线图中——已有 Android Engineer 岗位,但产品尚未上线 | 时间点和范围仍未确认(CE005、CE032) |
| 未注明日期(截至运行日) | 高级 Enterprise MCP 能力(访问控制、使用监控、更高限速) | 公司公告称仅在路线图中 | Enterprise MCP 成熟度仍落后于基础产品(CE033) |
日期和里程碑来自 Granola 的公开更新日志、博客和 MCP 公告;「仅在路线图中」的行没有公司承诺日期。
[CE028, CE029, CE050, CE051, CE025, CE033]Granola 用户如何从一场已排期会议,走到采集、会中辅助、会后综合,再到组织级记忆。
流程根据 Granola 自有产品、博客和更新日志页面对连续功能发布的描述推断,并非来自单一官方工作流图。
[CE002, CE027, CE049, CE025, CE036, CE028]5.5 差异化
Granola 对核心差异化的表述不是功能,而是架构:无机器人、本地采集、服务器端零音频留存,转写后立即删除音频;公司称这正是其约三个月完成 SOC 2 Type II 认证、而不是通常 12-18 个月的直接原因(CE034)。在这套架构之上,Granola 用 “CODE” 框架(Capture、Organize、Distill、Express)定义 Chat 产品,认为只对原始转写做关键词搜索,无法沉淀可复用的组织记忆(CE035)。Recipes 则用来自 Lenny Rachitsky、Matt Mochary 等具名外部专家的可分享模板,而不只提供公司自写的通用提示词,借此区分提示词库路线(CE036)。 独立审视下,这些主张只有部分耐久性。竞争对手发布的评测(tl;dv)估计 Granola 转写准确率约 94-95%,低于 tl;dv 自称的 97%+;又称 Granola 支持约 15 种语言,而 tl;dv 支持 30+。这把 Granola 的真实差异化界定为工作流和采集设计,而不是原始转写性能,不过该来源本身带有明显竞争偏见(CE037)。更实质的是,Granola “bot-free” 和 “private by default” 的定位已在两个战线遭到公开、独立质疑:一名独立评测者将 Granola 企业就绪度只评为 3/10,理由是默认分享设置弱、没有 EU 或本地部署选项;另一家媒体称,默认情况下,任何拿到链接的人都可访问笔记,除非用户手动更改设置(CE038)。合在一起看,Granola 的差异化在架构层面真实存在,但既非独占,也未在营销主张层面得到完全佐证;这一区分直接影响该架构究竟能支撑多少定价权和转换成本韧性。[CE034, CE035, CE036, CE037, CE038]
Granola 的无机器人核心采集和企业控制能力证据充分、较成熟;代理 / MCP 集成,以及数据驻留 / 质量保证能力已经上线,但存在争议或仍有缺口。
成熟度位置是基于公司披露、独立安全研究和独立评测综合而来的定性判断,不是打分式供应商评估框架。
[CE001, CE016, CE033, CE022, CE030, CE005]5.6 信任、安全、隐私、合规与质量控制
Granola 称其已获得 SOC 2 Type II 认证(首次于 2025 年 7 月 7 日宣布)并符合 GDPR,企业采购团队可通过 Trust Center 获取独立审计文档(CE039)。其 Data Processing Addendum 将 Granola 定义为客户提交个人数据的数据处理者,并要求在 Personal Data Breach 后“不得无故拖延”通知;同时,Granola 对自身业务联系人和使用数据另行作为独立控制者(CE040)。一个具体合规缺口仍未关闭:Granola 目前没有 HIPAA Business Associate Agreement。即便其帮助中心另有专门的 Healthcare & HIPAA FAQ,公司仍在自家博客直接披露这一缺口(CE041)。模型训练方面,Granola Enterprise 计划默认关闭训练,并在合同上禁止 OpenAI 和 Anthropic 用客户数据训练;免费和标准消费者账户则默认加入匿名化训练,需手动 opt-out(CE042)。 Granola 维护公开 Vulnerability Disclosure Policy,自 2025 年中以来发布了三份详细安全事后复盘(CE043)。第一起由独立安全公司 Tenable 发现,涉及未经认证的 API 端点泄露实时 AssemblyAI 转写 key(CE018, CE019)。第二起披露称,在 2025 年 5 月 23 日修复前,workspace auto-join 只检查用户 Google 邮箱域名是否匹配 workspace 域名,可能让未托管的个人账户自动加入企业 workspace;不过 Granola 自查数据库后称没有 workspace 实际受影响(CE044)。第三起披露称,187 名从关联 Google Workspace 中移除、暂停或删除的用户仍保留活跃 Granola session,可继续访问历史笔记和转写,直到 2025 年 5 月 28 日至 6 月 2 日部署 session-validity 检查(CE045)。正面看,Granola 的安全页面和 local-first 架构博客都称会议音频本身从不存储在任何服务器上——只保留生成的文本——这把未来泄露的爆炸半径收窄到文本,而不是已存音频或生物识别语音数据(CE046)。一个独立开源项目也确认,本地缓存可完全离线读取,无需联系 Granola 服务器;但 2026 年 3 月本地数据库加密上线后,这条路径已部分被封闭(CE047)。审阅过的来源没有发布 Granola 的独立转写准确率基准;唯一可用数字(约 94-95%)来自单一竞争对手撰写的评测,只能作为方向性参考(CE048)。Enterprise 透明度控制——可选的屏幕录制水印,以及转写开始时自动发送的会中聊天通知——补齐了管理员可配置的信任层(CE049)。[CE039, CE040, CE041, CE042, CE043, CE044]
| 控制项 / 认证 / 指标 | 状态 | 范围 | 缺口 |
|---|---|---|---|
| SOC 2 Type II | 已认证(Jul 7, 2025 首次宣布) | 公司级安全控制 | 公司自行声明;本章未独立审阅审计报告本身 |
| GDPR 合规 | 公司称已合规 | 欧盟个人数据处理 | 未发现 ISO 27001 或欧盟托管选项,难以覆盖重视欧盟数据主权的买家 |
| HIPAA 商业伙伴协议(BAA) | 目前不提供 | 医疗 / 医疗相邻数据 | BAA 可用前,受监管医疗买家基本被挡在门外(CE041) |
| 漏洞披露政策 + 事后复盘 | 自 2025 年中以来公开 3 份事后复盘 | API key 暴露;2 次 Google Workspace 访问控制缺口 | 尚不清楚 3 起已披露事件是否就是完整事故记录(evidenceGap) |
| AI 模型训练退出选项 | Enterprise 默认关闭;免费 / 标准版默认开启(可退出) | 客户会议数据与匿名化产品改进数据 | 用户不操作时,消费者层默认加入、可退出,而非主动加入(CE042) |
| 静态 / 传输中加密 | AWS 美国 VPC 上采用 AES 系列加密;本地缓存自 Mar 2026 起加密 | 转写稿、笔记、本地缓存 | 音频本身从不存储;反过来也没有原始音频审计轨迹 |
| 独立转写准确率基准 | Granola 或独立第三方均未发布 | 核心产品质量 | 唯一数据点来自一个竞品撰写的估算(约 94-95%)(CE048) |
| 笔记隐私 / 分享默认设置 | 公司称「默认私密」 | 个人笔记 | 独立报道称,链接分享默认开放,除非用户另行修改(CE038) |
| Enterprise 透明度控制(水印、聊天通知) | 可用,可由管理员配置 | 会中向参会者披露 | 这些可选控制的启用率 / 使用率未披露 |
认证和政策表述除另有说明外,均为公司自行声明;事后复盘细节直接来自 Granola 已发布的事故后报告,并在可能时用独立安全研究交叉印证。
[CE039, CE040, CE041, CE043, CE042, CE046]5.7 展示材料
06客户
6.1 客户基础与分层
Granola 通过三个自助 / 企业计划销售,它们同时构成买方分层阶梯:免费 Basic 层;Business 计划,按 Granola 自家销售指南为每用户每月 $14;Enterprise 计划(定制定价,独立估计约每用户每月 $35),增加 SSO/SCIM、管理员控制和模型训练 opt-out(CU001)。公司 Enterprise 页面称其“受到技术、金融、媒体等团队信任”(CU002),但本章详细且经独立佐证的证据几乎全部集中在技术和技术相邻公司。G2 评测者覆盖小型企业、中型市场和企业公司规模,行业包括信息技术与服务、金融服务,以及临床 / 医疗相邻角色(CU003)。独立评测者也反复点名招聘人员、律师、顾问、VC、销售工程师和 IT / 客户成功负责人是 Granola 核心用户画像——这些专业人士会议频繁、风险高,且常常对保密敏感(CU004)。 Granola 四个公司发布的详细案例研究可清晰映射到不同买方 / 用户职能:设计 / 产品 / 工程(Zapier)、客户成功领导层(Vanta)、IT(Brex)和销售工程 / GTM(Vercel)(CU005)。四家之外,高管猎头公司 Daversa Partners 总裁在 150 名员工中的 136 人部署 Granola,明确因为无机器人、设备级采集比可见机器人替代品更适合保密候选人与客户对话,侵入感更低——这是一个保密需求尖锐、但在详细证据中尚未被代表的垂直领域(高管猎头)(CU006)。媒体和投资人报道还点名 Asana、Gusto、Thumbtack、Cursor、Lovable、Decagon、Mistral AI、Intercom、Linear、Replit 为客户,横跨工作管理、人力 / 薪酬、市场平台、开发者工具和 AI-native 初创公司(CU007)——但这十个客户标识本身也全部是技术或技术相邻公司,进一步强化了详细案例中可见的行业集中。审阅过的来源没有披露 Granola 的单客户收入、席位数、合同金额,也没有披露其客户基础在英国 / 美国本土市场之外的官方地理拆分;收入区间和地理区域仍是开放尽调问题,而非已回答的分层维度(CU008)。[CU001, CU002, CU003, CU004, CU005, CU006]
| 细分市场 | 买方 / 用户 / 付款方 | 用例 | 规模 / 方案 | 战略价值 | 尽调缺口 |
|---|---|---|---|---|---|
| 个人专业用户(招聘顾问、律师、咨询顾问、VC) | 买方和用户是同一个人,自助购买 Free / Business | 1:1、候选人 / 客户电话、尽调对话 | Free Basic 或 $14/mo Business;未披露席位数 | 高频入口,播下病毒式传播和先落地后扩张的采用种子 | 未披露个人用户总数 |
| 设计 / 产品 / 工程团队(如 Zapier) | 先由个人使用,自然扩散后团队采购 | 跨职能站会、1:1、共享知识库 | Enterprise 方案,组织范围(52% 采用率) | 旗舰标杆案例 | 未披露席位数或合同金额 |
| 客户成功 / 合规敏感组织(如 Vanta) | VP 级买方,高管和团队用户 | 内部 1:1、站会、跨职能同步;由安全需求推动采用 | Enterprise 方案,组织范围(69% 采用率) | 旗舰标杆;直接支撑 Granola 的 SOC 2 定位销售 | 未披露续约或 NRR |
| IT 主导的企业部署(如 Brex,受监管金融科技) | IT 负责人采购,各部门使用 | 面试、供应商电话、站会、事后复盘、销售电话 | Enterprise 方案,组织范围(6 个月内 >50% 采用率) | 证明即便在受监管环境中,IT 仍会批准 | 未披露流失或续约数据 |
| 销售 / GTM 团队(如 Vercel) | 销售工程买方,GTM 团队用户 | 发现电话、演示、POC 启动会 | Enterprise 方案,团队范围(100% GTM 采用率) | 证明「与转写稿聊天」在收入职能中的价值 | 单团队案例,而非全公司 |
| 保密敏感垂直领域(如 Daversa Partners,高管猎头) | 公司总裁 / 领导层采购 | 候选人和客户访谈采集 | 150 名员工中 136 人 | 验证无 bot 定位能打动信任敏感细分市场 | 除采用人数外,未披露量化生产力或结果指标 |
| 更广泛的中端市场 / 企业科技客户 logo(Asana、Gusto、Thumbtack、Cursor、Lovable、Decagon、Mistral AI、 Intercom、Linear、Replit) | 媒体提及显示为公司级买方 | 日常内部工作流仅被描述为「第二大脑」 | 企业 / 中端市场,规模未说明 | 印证客户集中在科技 / SaaS / AI-native 买方群体 | 十个 logo 均无使用深度、合同或续约数据 |
规模 / 方案数据来自 Granola 自身案例研究、销售指南和第三方 Enterprise 价格估算;战略价值为分析师判断,不是公司披露的收入区间,因为 Granola 在已审阅材料中从未披露单客户收入。
[CU001, CU002, CU003, CU004, CU005, CU006]基于四篇已发布案例研究和 Granola 自有产品 / 扩张公告,展示单个 Granola 用户的采用通常如何演进为全公司、接入代理的企业部署。
阶段综合自四篇已发布案例研究、Granola 自有 Enterprise 与 Series B/C 公告及 MCP 文档;没有单一官方图表描绘这一旅程,也并非每个账户都会走到每个阶段(例如 Vercel 部署停留在团队范围,而非全公司)。
[CU038, CU039, CU040, CU018, CU019, CU020]6.2 采用轨迹
Granola 自己披露的采用指标稀少且已经过时:公司在 2024 年 10 月 Series A 时报告每周 5,000 名用户(CU009);独立分析明确指出,Granola 自那之后“未披露当前用户数”(CU010)。公司招聘页面只称其每天“转写数百万分钟对话”,没有精确数字,也没有历史趋势(CU011);独立媒体报道引用公司称截至 2026 年 3 月 Series C 时每周用户增长率为 10%,同样没有披露适用该百分比的基数(CU012)。 最严谨的采用证据并非来自公司,而是 YipitData 的 Signals 面板;该面板追踪 900 多家中型市场公司的真实 B2B 软件交易。面板显示,Granola 被观察到的中型市场客户数在 2026 年 2 月环比增长 25%(CU013),新增总客户数从 2026 年 1 月到 2 月增长 9x,2 月净新增客户达到此前六个月均值的 4x——这是阶跃式拐点,而非平稳增长(CU014)。货币化正在跟随采用:2026 年 2 月观察到的中型市场支出约 $6,000,较前六个月增长 3x(CU015)。即便如此,YipitData 提示,三家具名竞争对手——Fireflies.ai、Otter.ai 和 Fathom——中型市场总客户基础仍大于 Granola(约大 20-40%),只是其增长相对放缓,而 Granola 在加速(CU016)。同一六个月窗口内,员工人数增长 49%,与公司扩张运营以匹配需求相一致(CU017)。合在一起,本章最强、最新的采用信号是独立交易数据,而非公司披露;尽调应谨慎加权这一区别。[CU009, CU010, CU011, CU012, CU013, CU014]
| 指标 | 值 | 日期 | 来源可信度 | 含义 | 缺失分母 |
|---|---|---|---|---|---|
| 每周活跃用户 | 5,000 | Oct 2024 (Series A) | 中(第三方报道) | 首次机构融资时的基线规模 | 此后未披露更新后的总用户数 |
| 每日转写的对话分钟数 | 「数百万」(未量化) | 截至运行日仍持续 | 低(公司说法笼统) | 指向较大的绝对使用量 | 未给出历史趋势或精确数字 |
| 每周用户增长率 | 10%(报道值) | 截至 Mar 2026 Series C | 中(媒体报道,非公司发布指标) | 若能持续,意味着复合增长 | 该百分比对应的用户基数未披露 |
| 中端市场客户数增长(YipitData Signals) | 环比 +25% | Feb 2026 | 高(独立、基于交易的样本面板) | 现有最强的独立采用动能信号 | 面板只覆盖中端市场,不覆盖 SMB、个人或企业层 |
| 净新增客户激增(YipitData) | 为 6 个月均值的 4x | Feb 2026 | 高(独立) | 这是阶跃式拐点,不是平稳递增 | 绝对客户数未披露 |
| 中端市场支出增长(YipitData) | 6 个月内 3x,观测支出达 ~$6,000 | Feb 2026 | 高(独立) | 变现开始追上采用 | 来源未说明 $6,000 的单位(单账户月度或年度) |
| 员工数增长 | +49% | 截至 Apr 2026 的 6 个月 | 高(独立,YipitData) | 可作为组织扩张追赶需求的代理指标 | 不是直接的客户或收入指标 |
| 当前付费客户 / 用户总数 | 未披露 | 截至运行日 2026-07-01 | n/a | 评估客户基数规模的核心尽调缺口 | 自 Oct 2024 以来,公司从未发布更新后的汇总数字 |
第 1-3 行来自公司或媒体报道;第 4-7 行来自 YipitData 的独立 B2B 支出面板(Signals),该面板观察真实交易,而非自报使用量。第 8 行记录采用轨迹的核心缺口。
[CU009, CU010, CU011, CU012, CU013, CU014]6.3 具名客户证明
Granola 最强客户证据是四份公司发布的详细案例研究。Zapier 的 Principal Designer 从个人试用开始,在设计 / 产品 / 工程组织内病毒式扩散,案例称每月节省 830+ 小时,组织级采用率 52%(CU018)。Vanta 的 VP of Customer Success 将 Granola 用于内部 1:1、站会和跨职能同步,明确因为 SOC 2 Type II 认证和基于同意的录音对安全与合规公司很重要;案例称每位用户每年节省 260+ 小时,组织级采用率 69%,会议可见性提升 61%(CU019)。Brex 的 Head of IT 在员工已自发采用 Granola 后,把它正式推向全公司;案例称六个月内知识员工采用率超过 50%,每周节省 427 小时(CU020)。Vercel 的销售工程团队用 Granola 替换了更慢的转写工具,并在 discovery calls 中使用其 “chat with the transcript” 功能;案例称 GTM 团队采用率 100%,每位用户每周节省 11 小时,GTM 生产力提升 25%(CU021)。关键是,这四个案例都没有披露席位数、合同金额、续约率或流失数据——所有量化收益都是采用率或节省小时数估算,不是与收入挂钩的指标(CU022)。 独立、非公司撰写的证据佐证了真实但衡量不均的满意度。Granola 的 G2 档案在 2025 年末快照时有 19 条评价、综合评分 4.9/5,评测者覆盖小型企业、中型市场和企业;其中几人的身份通过 business email、Google One Tap 或 LinkedIn 验证,并标记为 organic、unincentivized,另至少两人披露为评价收到了象征性礼品卡(CU023, CU024)。同一批 G2 评测者反复提到具体限制:没有音频 / 视频回放,大型会议里说话人识别不稳定,没有专用 Web 应用,设置问题上的客户支持较慢(CU025)。汇总后的 Product Hunt 评论也呼应了类似缺口——没有 Android 应用、iPad / iOS 采集限制、账户设置摩擦,以及希望更多集成(CU026)。三份独立专家评测给 Granola 的分数明显低于 G2 的自选评分,且彼此更不一致:thebusinessdive.com 给 3.6/5,理由是协作和分析功能薄;nubiapage.com 给 4.5/5,提到缺少 Android 应用、免费层限制不清;bestreviews.net 指出语言支持和说话人识别缺口,但没有给数字评分(CU027)。另一个置信度较低的数据点来自独立 Product Hunt 社区帖,称面对面、多语言对话采集——例如把日语对话转成英文笔记——是移动端应用一个突出的真实用例(CU028)。四个旗舰账户之外,媒体和投资人报道还点名十个客户标识(Asana、Gusto、Thumbtack、Cursor、Lovable、Decagon、Mistral AI、Intercom、Linear、Replit),但没有任何一个附带经独立验证的使用深度、合同或结果细节(CU029)。[CU018, CU019, CU020, CU021, CU022, CU023]
| 客户 | 细分市场 | 部署 / 用例 | 生产使用 / 试点 | 结果 | 局限 |
|---|---|---|---|---|---|
| Zapier | 设计 / 产品 / 工程(SaaS 自动化平台) | 全组织会议笔记和知识库 | 生产使用,企业范围(52% 采用率) | 每月节省 830+ 小时;会议总数下降 | 未披露席位数、合同金额、续约或流失 |
| Vanta | 安全 / 合规软件 | 高管 1:1、站会、跨职能同步 | 生产使用,企业范围(69% 采用率) | 每位用户每年节省 260+ 小时;会议可见性提升 61% | 证据限制同上;采用率由公司报告,未经独立审计 |
| Brex | 金融科技(受监管) | 用按角色定制的模板化回顾,覆盖全公司会议笔记 | 生产使用,企业范围(6 个月内 >50% 采用率) | 每周节省 427 小时;累计节省 25,000+ 小时 | 证据限制同上;单公司案例 |
| Vercel | 销售工程 / GTM(开发者平台) | 销售电话采集和「与转写稿聊天」 | 生产使用,团队范围(100% GTM 采用率) | 每位用户每周节省 11 小时;GTM 生产力提升 25% | 仅限 GTM 团队,不是全公司 |
| Daversa Partners | 高管猎头(保密敏感) | 候选人和客户访谈采集 | 生产使用(150 名员工中 136 人) | 专门因敏感对话需要无 bot、低调采集而采用 | 除采用人数外,未披露量化生产力或结果指标 |
| 其他公开具名 logo(Asana、Gusto、Thumbtack、Cursor、Lovable、Decagon、Mistral AI、Intercom、Linear、 Replit) | 混合:工作管理、HR / 薪酬、市场平台、开发者工具、AI-native 初创公司 | 仅被描述为团队「每天」依赖 Granola,把它当作「第二大脑」 | 不清楚——没有任何案例研究、采用率或发布日期 | 印证客户集中在科技 / SaaS / AI-native 买方群体 | 十个 logo 均没有独立的结果、使用或留存证据 |
四个详细案例研究中的结果数字由公司发布,未经独立审计。汇总行存在,是因为没有独立来源能为十个额外具名 logo 中任何一个提供案例研究级细节。
[CU005, CU006, CU007, CU018, CU019, CU020]本章来源中提到的每一个 Granola 客户里,有多少拥有已发布的量化结果、全公司采用,或与收入相关的指标。
15 = 4 个详细案例研究账户(Zapier、Vanta、Brex、Vercel)+ Daversa Partners + 10 个仅标识提及(Asana、Gusto、Thumbtack、Cursor、Lovable、Decagon、Mistral AI、Intercom、Linear、Replit)。Vercel 案例研究聚焦 GTM 团队而非全公司,因此被排除在第三阶段之外,但仍留在第二阶段。
[CU018, CU019, CU020, CU021, CU022, CU029]截至 2026-07-01 运行日,每个具名账户相对独立中端市场面板的证据质量、生产成熟度、结果具体性和留存可见度。
评级是分析师基于来源层级和佐证深度在高 / 中 / 低尺度上的评估,不是打分式供应商框架。注意相反模式:具名账户的结果具体性高但留存可见度低,而独立面板正好相反。
[CU005, CU006, CU013, CU018, CU019, CU020]6.4 留存与韧性
审阅过的任何来源都没有披露 Granola 的 Net Revenue Retention、Gross Revenue Retention、logo-churn rate 或平均合同续约率(CU030)。缺少公司披露时,最佳可用留存信号仍是独立交易数据:YipitData 的 B2B 支出面板分析发现,在约 2025 年 8 月至 2026 年 2 月间,Granola 被观察到的中型市场客户几乎零流失(CU031)。同时,切换模式不对称:采用 Granola 的客户常常放弃 Fathom、Fireflies.ai、Otter.ai 等竞争工具,反向则很少出现在观察数据中;YipitData 将此解读为主动替代的证据,而不只是试用(CU032)。不过,“近零流失”是单一独立面板给出的定性描述,不是 Granola 或 YipitData 发布的精确百分比,而且只覆盖 YipitData 跟踪账户中的中型市场细分。 满意度证据比案例研究或流失数据单独呈现的故事更分裂。Granola 19 条评价得出的 G2 评分为 4.9/5,明显高于两位独立专家评测者按标准化框架测试产品后给出的 3.6/5 和 4.5/5,说明自选式评价平台评分与上手专家评估,对 Granola 客户到底有多满意并不一致(CU033)。留存风险也有一个有记录的爆点:2026 年 3 月,Granola 加密本地笔记数据库,引发一位知名 a16z 合伙人公开批评;该 tweet 约 327,000 次浏览、630 次点赞、577 次收藏、177 条回复,发帖人明确向社区征求替代产品推荐(CU034)。独立评论认为,这次反弹暴露了“一个已让 Granola 用户暗中沮丧数月的问题”,问题源于失去本地文件和 API 访问,而这些访问此前支撑了构建在 Granola 之上的个人自动化和 AI-agent 工作流(CU035)。TechCrunch 报道确认,反弹不止来自这位 a16z 合伙人,也包括其他用户;Granola 联合创始人的回应是承诺推出新的 personal 和 enterprise API,恢复程序化数据访问(CU036)。没有来源量化有多少客户(如有)因此实际离开 Granola——案例研究中的采用率(52%、69%、超过 50%、100%)只衡量某一时点的组织级使用深度,不衡量续约或多周期留存,也无法回答这个问题(CU037)。[CU030, CU031, CU032, CU033, CU034, CU035]
| 指标 | 值 / 空值 | 细分市场 | 可信度 | 尽调索取项 |
|---|---|---|---|---|
| 净收入留存(NRR) | 空值——未披露 | 公司范围 | n/a | 直接向 Granola 或 Series C data room 材料索取 NRR |
| 总收入留存(GRR)/ logo 流失 | 空值(公司未披露);独立资料定性称「接近零」(非精确百分比) | 中端市场(YipitData 面板) | 中(独立但定性) | 向 YipitData 或 Granola 索取精确流失 / 留存百分比及底层队列定义 |
| 合同续约率 | 空值——未披露 | n/a | n/a | 向 Granola 企业销售或 data room 索取续约率数据 |
| 切换 / 替代模式 | 非对称:采用者会放弃竞品,反向很少发生 | 中端市场(YipitData 面板) | 高(独立、基于交易) | 验证该模式是否也适用于 SMB 和企业层,而不只是中端市场 |
| G2 评论满意度 | 4.9/5(19 条评论;自选择样本,部分有激励) | 小企业 / 中端市场 / 企业混合 | 低-中(自选择、小样本 n、部分礼品卡激励) | 先用独立专家评测交叉校准,再视为代表性信号 |
| 独立专家评测满意度 | 三家测试方给出 3.6/5 至 4.5/5 | 通用 / 个人用户 | 中(标准化测试方法,n=3 家媒体) | 委托更大样本、具统计代表性的满意度调查 |
| 组织范围采用深度(案例研究) | 4 个具名账户为 52%-100% | 企业(仅具名账户) | 中(公司发布,未经独立审计) | 区分采用深度和续约;索取同四个账户第二年续约状态 |
| 加密变更后的高阶用户 / 数据可携性满意度 | 公开不满有记录;未量化流失影响 | 个人高阶用户 / AI-agent 构建者 | 中(公开反应有记录,但无流失数字) | 向 Granola 索取 Mar 2026 加密变更前后活跃用户队列的专项留存数据 |
空值行记录 Granola 从未公开披露的指标;本表所有非空值均来自独立第三方(YipitData、G2 或独立专家评测),而非 Granola 本身。
[CU030, CU031, CU032, CU033, CU034, CU035]由于 Granola 未披露精确 NRR/GRR/流失数字,这里用示意性留存代理;数值把 YipitData 独立观察到的“接近零流失”发现和四个旗舰案例研究的持续使用转换为 0-100 指数代理。
Granola 未发布精确 NRR、GRR、续约率或 logo 流失率。第 1 行将 YipitData 独立观察到的“接近零流失”和不对称切换发现(Signals B2B 支出面板,2025 年 8 月至 2026 年 2 月)转换为示意性的 97-100 留存指数代理;它不是精确发布百分比。第 2 行假设持续使用与每个旗舰案例研究自身发布叙事一致——没有任何案例披露流失或取消事件——但这不是多年留存的独立证明。参见 evidenceGaps,了解该代理无法补上的精确留存指标缺口。
[CU031, CU032, CU018, CU019, CU020, CU021]6.5 扩张与集中
Granola 自己的 Enterprise 营销页面把其 GTM 动作明确表述为 land-and-expand:“先是一个人爱上 Granola。它会扩散,因为它让每个人工作得更好”(CU038)。四个详细案例都在实践中确认了这种模式——每个都从个人或小团队试点开始,后来由领导层扩展成全公司 Enterprise 计划 rollout(CU039)。除席位增长外,Granola 还打造了刻意用于扩张的产品:用于跨职能知识共享的 Spaces / Team Folders、Personal API、Enterprise Admin API,以及用于第三方 AI-agent 集成的 Granola MCP;这些都旨在加深客户对 Granola 作为组织记忆层的依赖,而不是只把它当个人记笔记工具(CU040)。 几项集中风险削弱了这个扩张故事。第一,Granola 几乎所有量化客户结果证明都压在四个具名账户上,且每个都是资金充足、与 Granola 自身投资方网络有关联的技术公司——这是证据集中风险,区别于但又相关于收入集中:如果任何旗舰账户的指标被修订,或关系终止,本章大多数结果证据都会消失(CU041)。第二,独立报道称 Granola 增长主要靠口碑;审阅过的任何来源都没有记录已披露的 reseller、channel partner 或 systems integrator 分发——相对于可能运行多渠道 GTM 的竞争对手,这是一种单渠道依赖(CU042)。第三,独立分析警告称,Microsoft Copilot 和 Google Gemini 正把相同的会话上下文聚合能力原生嵌入 Teams 和 Workspace 环境,而多数企业员工本来就已有这些席位——这是平台层面的集中风险,即使 YipitData 当前动能很强,也可能压缩 Granola 的扩张跑道(CU043)。最后,Granola 自己的 enterprise 页面称其受到“技术、金融、媒体等”团队信任,但本章每个经独立详细呈现的具名账户都是技术或技术相邻公司;审阅过的来源中没有任何金融或媒体行业账户达到案例研究级细节,已知合规缺口(无 HIPAA BAA、无 EU 托管部署,见产品 / 技术章)也合理解释了为何受监管垂直扩张在此尚未得到证明(CU044)。[CU038, CU039, CU040, CU041, CU042, CU043]
| 扩张驱动因素 | 集中风险 | 影响 | 尽调路径 |
|---|---|---|---|
| 个人先用、再扩到企业(四个案例均如此) | 证据集中:同一组四个账户提供了几乎全部量化结果证明 | 任何一个标杆账户关系终止,或指标被修订,本章大部分结果证据都会消失 | 直接向 Zapier、Vanta、Brex 和 Vercel 客户团队索取第二年最新状态和指标 |
| Spaces/Folders 跨职能扩张 | 客群集中于技术 / SaaS / AI 原生买家(Asana、Cursor、Lovable、Mistral AI、Decagon、Linear、 Replit、Thumbtack、Gusto) | 一旦技术行业渗透见顶,若受监管行业或非技术垂直未跑出牵引力,增长可能受限 | 直接向 Granola 企业销售团队索取具名账户名单和客群结构 |
| Personal/Enterprise API 与 MCP 智能体生态锁定 | 供应商 / 平台集中:扩张越来越依赖外部 AI 生态(Claude、ChatGPT、Cursor)采用 MCP | 如果 MCP 普及停滞,或竞争协议胜出,基于 API 的扩张假设会走弱 | 跟踪 MCP 客户端采用情况,以及企业 MCP beta 转正时间表 |
| 自然口碑驱动增长(未披露渠道 / 经销商计划) | 渠道集中:只有单一分发动作,未记录合作伙伴、经销商或系统集成商网络 | 相较拥有合作伙伴渠道的竞争对手,Granola 快速推进企业扩张的能力受限 | 直接询问 Granola 是否已有或计划推出合作伙伴 / 经销商计划 |
| YipitData 显示中端市场势头(环比增长 25%,流失接近零) | 平台 / 竞争集中风险:Microsoft Copilot 和 Google Gemini 正把原生上下文聚合做进既有席位 | 即使当前势头强劲,也可能压缩 Granola 可触达的扩张空间 | 跟踪 Microsoft 和 Google 的会议 AI 功能发布,以及对 YipitData Granola 替换指标的影响 |
| Granola 自有营销暗示向技术行业之外扩张(金融、媒体、医疗) | 垂直集中:当前具名证据集中在非受监管的技术 / SaaS;HIPAA 与 EU 托管缺口(见产品 / 技术章节) 堵住受监管或主权敏感扩张 | 合规缺口补齐前,总可触达扩张空间受限 | 直接向企业销售索取 Granola 的 HIPAA BAA 和 EU 托管路线图时间表 |
集中风险由分析师基于本章客户、渠道和竞争证据综合得出;均非公司披露的风险因素。
[CU038, CU039, CU040, CU041, CU042, CU043]6.6 展示材料
07风险
7.1 风险排名概览
本章按发生可能性、严重性、缓解成熟度和剩余暴露,对 Granola 的法律、监管、运营、技术、财务、合作伙伴、客户和执行风险排序。依据包括主要法规与监管指南(英国 ICO、EU AI Act、California wiretap statute、Illinois biometric-privacy law、HHS HIPAA Privacy Rule)、同类 AI 会议记录工具的诉讼与律所分析、Granola 自身安全事后复盘和合规页面,以及前文收集的独立财务、竞争和员工数据。综合权衡后,优先级最高的风险类别是:合作伙伴 / 供应商集中与平台捆绑替代、财务和资本充足性披露不透明、法律 / 监管同意与隐私暴露、反复由外部发现的运营 / 安全事故,以及执行 / 治理披露缺口(CR058)。审阅过的来源没有发现 Granola 涉诉、监管执法行动、已披露现金 / burn 数字,或具名非创始高管——这意味着本排名中几个最严重项目目前仍是前瞻性或结构性风险,而非已经发生;每项都有可监测的尽调路径,但没有公司披露的量化阈值(CR059)。下方风险热力图绘制了后续各节展开的八类风险在发生可能性、严重性、缓解成熟度和剩余暴露上的位置。[CR058, CR059]
截至 2026-07-01 运行日,Granola 八个最高优先级风险类别的发生可能性、严重性、缓释成熟度和剩余暴露。
发生可能性 / 严重性 / 缓释成熟度 / 剩余暴露标签,是综合本章风险登记册形成的分析师评估,不是公司披露的风险分数。
[CR058, CR010, CR029, CR038, CR041, CR031]7.2 法律与监管风险
Granola 自己的法律文件只建立了基线合规姿态:Privacy Policy 允许为满足法律要求向政府和执法机构披露(CR001);Data Processing Addendum 承诺在泄露后“不得无故拖延”通知(CR002);Terms of Service 链接到公开 Vulnerability Disclosure Policy 和三份安全事后复盘(CR003)。公司文件之外,适用监管版图相当重。英国 ICO 的 monitoring-workers 指南要求雇主部署任何录音或转写工具时,必须有记录在案的合法依据、比例原则和透明度(CR004);其 AI 专项指南还要求关注公平性、问责和统计准确性(CR005)。考虑到 Granola 依赖 OpenAI 和 Anthropic 模型,EU AI Act 的通用 AI 透明度规则直接相关,并已于 2025 年 8 月适用;更广泛的 AI 系统透明度义务将在 2026 年 8 月跟进(CR006)。美国方面,California 的全方同意 wiretap statute 将未经每位参与者同意录制保密通信定为犯罪(CR007);Illinois 的 Biometric Information Privacy Act 要求任何实体采集声纹前提供书面通知并取得同意(CR008);HHS 的 HIPAA Privacy Rule 要求任何 AI 供应商在代表 covered entity 处理受保护健康信息前签署书面 business-associate agreement(CR009)。 这对 AI 会议记录工具类别并非假设:针对 Otter.ai 的合并联邦集体诉讼指控的正是这些 wiretap 和 biometric-privacy 违规(CR010);Bloomberg Law 报道称,法院在 AI 转写供应商到底是“工具”还是未经授权的窃听者上存在分歧——这一区别可能让客户公司承担替代责任(CR011);Goodwin Procter 2026 年 4 月客户提示把 California 法定赔偿量化为每次违规最高 $5,000,并单独提示生物识别和特权风险(CR012)。九个美国州要求全方同意(CR013)。Granola “无机器人”、设备级采集设计在机械上不同于 Otter.ai 的 bot-based 模式,但关于供应商分类的独立法律评论表明,这一差异本身不能解决其他会议参与者的底层同意问题(CR014)。Granola Enterprise 计划的 SOC 2 Type II 认证被营销为“独立审计”,但 SOC 2 本身不认证 wiretap、UK GDPR、AI Act 或 HIPAA 合规(CR015);公司自家博客还披露其目前没有 HIPAA Business Associate Agreement(CR016)。审阅过的来源没有发现直接点名 Granola 的诉讼或监管行动,因此该类别迄今已实现的风险是同类可比、行业范围风险,而非公司特定风险(CR017, CR018)。[CR001, CR002, CR003, CR004, CR005, CR006]
| 规则 / 许可 / 案件 | 司法辖区 | 状态 | 可能性 | 严重性 | 缓释措施 | 剩余敞口 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| 全员(双方)同意窃听法(如 California Penal Code Sec. 632) | 美国(CA, FL, IL, MA, MD, MT, NH, PA, WA) | 可比竞争对手 Otter.ai 正处于合并后的联邦诉讼;未发现针对 Granola 的案件 | 中 | 高 | 无 bot、设备端采集定位(部分缓释;不能解决非账户持有人同意问题) | 中高 | 确认 Granola 面向全员同意州非账户持有人参会者的应用内同意披露机制 |
| Illinois Biometric Information Privacy Act(声纹采集) | 美国(Illinois) | 未披露 Granola 特有声纹功能;针对 Otter.ai 等公司的品类级诉讼已被引用 | 中低 | 高 | 未披露 | 中 | 确认 Granola 的转写管线是否执行说话人声纹识别或说话人分离,进而可能生成生物识别标识符 |
| HIPAA 业务伙伴协议(BAA)缺口 | 美国医疗垂直 | 已确认缺口——公司披露未持有 BAA | 中 | 中高 | 无(缺口已披露,但未解决) | 医疗相邻账户:高 | 索取 Granola 为受监管医疗客户提供 BAA 的计划或时间表 |
| EU AI Act——通用 AI 透明度与披露义务 | 欧盟 / 欧盟境内用户 | GPAI 规则自 2025 年 8 月适用;更广泛的 AI 系统透明度规则自 2026 年 8 月适用 | 高(Granola 已依赖 GPAI 模型) | 中 | 未披露针对 AI Act 的措施 | 中 | 索取 Granola 针对 AI Act 的合规映射,或覆盖其 OpenAI/Anthropic 依赖的 DPIA 文件 |
| UK ICO / UK GDPR 职场监控合规 | 英国 / 雇有英国或欧盟员工的客户 | 这是持续合规义务,不是特定案件或罚款 | 中 | 中 | DPA、Enterprise 模型训练退出选项、SOC 2 Type II | 中 | 索取 Granola 面向 Enterprise 职场部署的合法依据文件和 DPIA |
严重性和可能性是分析师基于主要法规 / 监管指引及可比供应商诉讼作出的评估,不是公司披露的风险评级;已审阅来源未披露任何直接点名 Granola 的诉讼或执法。
[CR004, CR005, CR006, CR007, CR008, CR009]7.3 运营、质量与安全风险
Granola 在 2025 年公开披露了三份安全事后复盘,涵盖 AssemblyAI API key 暴露、遗留未托管 Google 账户 workspace auto-join 缺口,以及影响 187 名用户的 Google Workspace session-logout 漏洞(CR019)。独立安全研究者 Tenable 从外部技术视角佐证了 AssemblyAI key 暴露,识别出同一个未经认证的端点(CR020)。值得注意的是,三起披露事故全部由外部研究者报告给 Granola,而不是由内部检测捕捉(CR023);审阅过的来源也没有发现公开 Vulnerability Disclosure Policy 之外的独立 penetration-test 报告或 bug-bounty 项目(CR025)——不过,唯一有记录的独立研究发现已经表明,公司服务器端 API 表面,而不只是客户端应用,至少被探测并利用过一次(CR026)。根据 Granola 自己的描述,三起披露事故都没有暴露实际会议音频或转写内容;这比内容保密性泄露的严重程度明显更低(CR027)。 可靠性方面,一家独立 uptime 跟踪器在最近 90 天窗口内记录了 4 起 Granola 事故,中位时长约 2.5 小时(CR022),但 Granola 没有发布自己的 uptime SLA 或事故日志。Granola Enterprise 计划宣传其约三个月拿到 SOC 2 Type II 认证——相较该认证通常需要多季度观察期,这个速度很快——而公司自己的 Security FAQs 又把 HIPAA 和 FERPA 列为开放合规主题,而非已落定认证(CR021, CR024)。一名低声誉层级的独立评测者另称,Granola 默认通过链接分享用户笔记,且在未 opt-out 时用转写训练内部模型;如果这一说法具有代表性,会构成有意义的运营和信任风险,但尚未由技术审计独立验证(CR028)。[CR019, CR020, CR021, CR022, CR023, CR024]
| 失效模式 | 可能性 | 严重性 | 缓释成熟度 | 剩余敞口 | 未解决缺口 |
|---|---|---|---|---|---|
| 外部反复发现访问控制 / 密钥暴露漏洞(2025 年披露 3 起事件,另有 1 起 Tenable 独立发现) | 中高 | 高 | 被动响应——外部研究员报告后才发布事后复盘 | 中 | 未披露主动渗透测试计划或漏洞赏金 |
| 服务可靠性 / 正常运行时间事件 | 中(独立追踪器显示 90 天内 4 起事件) | 中 | 未知——公司未发布正常运行时间 SLA 或事件日志 | 中 | 公司未披露历史正常运行时间或 SLA 承诺 |
| SOC 2 Type II 认证在异常短(约 3 个月)窗口内完成 | N/A——已发生 | 中 | 已披露,但本报告未独立复核 | 中 | SOC 2 报告范围和审计方未获独立确认 |
| 免费层和标准层的默认数据共享 / 模型训练设置 | 中 | 中 | Enterprise 层可退出;免费 / 标准层默认设置未变 | 中 | 客户群中的退出采用率未披露 |
| 没有独立的转写或摘要质量准确率基准 | N/A | 中低 | 未披露 | 中低 | Granola 或第三方实验室未发布受控、独立的准确率对比 |
可能性 / 严重性基于 Granola 自有复盘、一个独立研究员披露和一个独立正常运行时间追踪器综合得出;事件计数仅覆盖 2025 年至 2026-07-01 运行日。
[CR019, CR020, CR021, CR022, CR023, CR024]7.4 合作伙伴、依赖与集中风险
Granola 自己的安全页面确认,产品依赖四家外部 AI 供应商:Deepgram 和 AssemblyAI 负责转写,OpenAI 和 Anthropic 负责摘要;转写稿和笔记只存放在美国托管的 AWS Virtual Private Cloud,且全程加密(CR029,CR060)。这不是理论依赖:AssemblyAI 正是 Granola 2025 年首次披露安全事件的直接入口,说明供应商集中风险已经兑现过一次(CR030)。第二个、规模可能更大的依赖风险来自竞争,而非供应侧:Microsoft 自己的文档确认,Teams 的 AI 会议回顾可通过每用户每月 $10 的 Teams Premium 附加包或 Microsoft 365 Copilot 许可证获得——它把相近功能打包进多数企业客户已经付费的套件,压低了 Granola $14 Business 计划的价格锚(CR031);与此同时,Google 在 Meet 内原生营销 Gemini 驱动的笔记(CR033),Zoom 则称 AI Companion 使用量同比增长超过三倍,独立市场报告显示其上线五个月内达到 510,000 个账户(CR032,CR034)。实体硬件记事设备又提供了一个非软件替代路径(CR035)。 Granola 的资本提供方也呈现集中模式:Index Ventures 和 Kleiner Perkins 的投资文章都点名现有 Granola 客户(Asana、Vanta、Gusto、Thumbtack、Cursor、Lovable)作为证明点,这意味着同一投资人网络既提供资本,也放大客户背书(CR036)。融资结构上,Companies House 文件显示 Granola 的英国母公司只有股权相关记录,没有登记押记(担保债务)文件;这与全股权资本结构、且不存在单独贷款方依赖相一致(CR037)。最后,已审阅来源没有发现经销商、系统集成商或渠道合作协议,因此 Granola 当前的市场进入依赖直销、自助和自然推荐渠道,而不是多元化分销网络(CR038)。[CR029, CR030, CR031, CR032, CR033, CR034]
| 依赖项 | 对手方 | 角色 | 集中度 | 失效情景 | 严重性 | 缓释措施 | 剩余敞口 |
|---|---|---|---|---|---|---|---|
| AI / ASR 计算供应商 | OpenAI、Anthropic、Deepgram、AssemblyAI | 核心转写与摘要 | 高——按功能划分有 4 家单一来源供应商 | 任何供应商涨价、宕机或政策变动,都会削弱产品或利润率;AssemblyAI 已与 2025 年安全事件相关 | 高 | 未披露(未说明多供应商冗余) | 高 |
| 平台捆绑的既有巨头 | 平台:Microsoft Teams Premium/Copilot、Zoom AI Companion、Google Gemini in Meet | 搭载在客户已付费协作套件中的借道分发替代品 | N/A——这是竞争集中,不是供应商集中 | 某个具名参考客户默认使用捆绑的 AI 会议笔记功能,而非续约 Granola | 高 | 差异化 Chat/Briefs/Recipes/MCP 智能体层定位 | 中高 |
| 云托管基础设施 | AWS(美国托管 VPC) | 加密转写稿和笔记的唯一存储地点 | 高——单一云供应商、单一区域 | 区域性 AWS 事故,或仅美国 VPC 无法满足的欧盟 / 英国数据驻留要求 | 中 | 静态和传输中加密、每日备份 | 中 |
| 资本提供方 / 客户背书重叠 | Index Ventures、Kleiner Perkins | Series C 投资者,也公开把 Granola 的具名客户作为证明点 | 中——投资者叙事与参考客户叙事重叠 | 尽调对手方质疑客户证据的独立性 | 中 | 未披露 | 中 |
| 分发渠道集中 | 未发现(仅直接销售 / 自助 / 自然推荐) | 主要获客路径 | 中——未披露经销商或系统集成商网络 | 如果自然 / 病毒式推荐在渠道计划成熟前放缓,增长会减速 | 中低 | 案例研究已记录先落地再扩张路径 | 中低 |
集中度和严重性评级为分析师评估;已审阅来源未披露供应商多元化计划、第二云区域或正式经销商计划。
[CR029, CR030, CR031, CR032, CR033, CR034]截至 2026-07-01 运行日,Granola 的关键第三方供应商、托管、竞争替代和资本依赖。
依赖边来自 Granola 自己的安全页面、Series C 投资方帖子以及独立竞品文档;该图仅覆盖本章审阅来源中识别出的外部可见依赖。
[CR029, CR060, CR030, CR031, CR032, CR033]7.5 财务、模型与执行风险
Granola 于 2026-03-25 宣布完成 $125M Series C,投后估值 $1.5B,公司与独立来源相互印证(CR039);TechCrunch 和 Tracxn 又分别收敛到 $192M 的已披露总融资额(CR040)。但除融资总额外,已审阅来源没有披露 Granola 的现金余额、烧钱速度或现金续航(CR041);作为一家英国私营公司,Granola 在 Companies House 提交了截至 2025 年 3 月 31 日年度的「total exemption full accounts」,近期没有法定普通申报义务去披露营业额、利润率或利润(CR042)。唯一可得的独立员工数是 Tracxn 截至 2026 年 5 月 26 日给出的 116 人(CR043);另有独立 B2B 支出面板称其六个月员工数增长 49%,但没有已披露收入,就无法把这一增速与人均收入基准相对照(CR044)。公开 SaaS 基准数据把订阅业务毛利率中位数放在约 76%,并提示沉重的 AI 推理算力成本常常把毛利率压到 70% 以下——Granola 每场会议要调用四家付费第三方 ASR/LLM 供应商,这一描述有理由适用(CR045)。相比之下,同一类别中更成熟的上市可比公司 Zoom 在 FY2026 Form 10-K 中直接披露 77.0% 毛利率,显示 Granola 作为私营公司还未达到的透明度基准(CR046)。Granola 及其投资人都没有披露未来融资轮的具体收入或使用量触发条件;投资人文章只是把本轮融资定位为支持多年产品建设(CR047)。 执行层面,BusinessCloud 和 TechCrunch 只识别出两位联合创始人 Chris Pedregal 和 Sam Stephenson,没有披露独立董事会名单(CR048);Pedregal 过往履历包括 Socratic(2018 年被 Google 收购)以及在 Google Area 120 内部搭建的 Stack(CR049)。Granola 招聘页确认公司仍在招聘,但没有披露当前总员工数(CR050);快速扩员叠加缺少独立确认的组织架构图,意味着公司面临的扩张复杂度,本章来源尚无法验证是否已有足够管理深度匹配(CR051)。没有任何来源显示 Granola 设有首席财务官、总法律顾问、合规负责人或数据保护官;考虑到本章前文记录的监管与诉讼邻近风险,这一缺口格外醒目(CR052)。[CR039, CR040, CR041, CR042, CR043, CR044]
| 角色 / 职能 | 依赖或缺口 | 可能性 | 严重性 | 缓释措施 | 尽调路径 |
|---|---|---|---|---|---|
| 创始人层领导力与治理 | 仅识别出两名具名联合创始人;未独立披露董事会名单、CFO、总法律顾问或数据保护官 | 中 | 高 | 未披露 | 索取完整高管和董事会名单,并确认是否有专职合规 / 法务负责人 |
| 资本充足性披露 | 未披露现金余额、烧钱速度或 runway;英国小公司审计豁免取消了任何短期法定披露触发点 | 高(结构性) | 高 | 无(结构性豁免,不是公司缓释措施) | 任何投资决定前,直接索取数据室财务资料 |
| 人员扩张和组织深度 | 六个月员工数增长 49%(独立样本),但只对应单个第三方员工数快照(116 名员工),且未披露组织架构图 | 中 | 中 | 招聘页面显示仍在积极招聘 | 按职能索取当前员工数(工程、安全、合规、GTM) |
| 安全检测与响应能力 | 2025 年披露的三起安全事件均由外部研究员报告,而非内部发现 | 中 | 中 | 已有公开漏洞披露政策 | 索取任何内部安全监控或事件响应计划详情 |
可能性 / 严重性为分析师评估;审计豁免和员工数是本表仅有的量化项,且均来自第三方或备案文件,并非公司披露。
[CR041, CR042, CR043, CR044, CR048, CR050]7.6 缓释措施、监测与终止标准
Granola 已披露的主要风险缓释措施,是 SOC 2 Type II 认证、Enterprise 层级默认关闭模型训练,以及 DPA 中的违规通知承诺(CR053)。这些措施都没有直接覆盖本章识别出的三类公司披露最少的监管问题:全方同意录音、EU AI Act 透明度义务,以及 HIPAA Business Associate Agreement;截至运行日,公司没有为这些风险给出明确缓释方案(CR054)。本章为每一类最高风险设定可监测的终止标准,而不是依赖直觉:若出现直接点名 Granola 的诉讼或执法行动,法律 / 监管风险就升级;其最接近的机器人型竞争对手已经越过这一门槛(CR055)。若出现降价融资、Series D 明显延迟,或披露的烧钱额超过 Series C 融资额,财务 / 资本风险就升级;已有两位负面评论者认为,鉴于估值跃升未经验证,这一结果可能发生(CR056)。若核心 AI/ASR 供应商出现有记录的涨价、宕机或政策变化,或有证据显示某个具名参考客户用打包的 Teams/Zoom/Meet 功能替代 Granola,合作伙伴 / 供应商风险就升级(CR057)。 综合前几节建立的风险登记,本章整体风险排序把合作伙伴 / 供应商集中与平台打包替代、财务 / 资本充足度不透明、法律 / 监管同意风险,排在运营 / 安全以及执行 / 治理风险之前;主要原因是前三类风险严重度高,而公司披露的缓释措施薄弱或缺失(CR058)。这五大风险类别每一类都有明确尽调路径,但没有一类配有公司披露的量化阈值,例如烧钱上限或诉讼成本储备——因此持续监测要依靠外部信号(诉讼文件、融资公告、供应商事件披露、具名客户流失),而不是公司报告指标(CR059)。下方风险传导图追踪各类风险预计如何传导到客户留存、利润率、融资,并最终影响投资论点。[CR053, CR054, CR055, CR056, CR057, CR058]
| 风险 | 可监控触发因素 | 阈值 / 事件 | 行动含义 |
|---|---|---|---|
| 法律 / 监管(窃听法、BIPA、AI Act、HIPAA) | 直接点名 Granola 的诉讼或监管执法文件 | 任何点名 Granola 的投诉、传票,或 ICO/FTC/州总检察长问询 | 立即重估法律成本和声誉风险;任何后续投资前升级尽调 |
| 财务 / 资本充足性 | 降价轮、Series D 明显延迟,或披露净烧钱超过 Series C 募资额 | 任何公开报道称估值标记持平 / 下调,或投资者以折价进行二级出售 | 承诺新资本前,重新评估入场估值和清算优先权堆栈 |
| 合作伙伴 / 供应商依赖 | 核心 AI/ASR 供应商涨价、宕机或政策变动,或具名参考客户出现已证实的捆绑既有厂商替换 | 具名账户(如 Brex、Vanta、Zapier、Vercel)确认切换到 Teams/Zoom/Meet 捆绑的 AI 笔记功能 | 重新核算客户论证使用的留存和扩张假设 |
| 运营 / 安全 | 12 个月内披露第四起安全事件,或任何涉及实际会议音频 / 转写内容暴露的事件 | 公开事后复盘或媒体报道显示内容级(不只是访问控制)泄露 | 视为支撑 Enterprise 销售的信任 / 合规叙事发生重大不利变化 |
| 执行 / 治理 | 任一具名联合创始人离职,或下一轮融资后仍没有具名 CFO/总法律顾问/DPO | 公开离职公告,或连续第二轮融资仍未具名合规或财务高管 | 在交易条款中重估关键人风险(vesting 加速、关键人保险、董事会权利) |
阈值是分析师定义的监控代理指标,不是公司披露的契约;基于已审阅来源,截至 2026-07-01 运行日,五个触发因素均未发生。
[CR053, CR054, CR055, CR056, CR057, CR058]Granola 的法律、运营、合作伙伴、财务和执行风险预计如何传导到客户留存、利润率、融资以及整体投资论点。
边将本章风险登记册综合成因果路径;这是分析师构建,不是公司披露的风险模型。
[CR055, CR056, CR057, CR045, CR047]7.7 附录
08估值
8.1 融资与估值背景
Granola 在不到三年内披露了四轮融资:约 $4.25M 种子轮;2024 年 10 月由 Spark Capital 领投的 $20M Series A;2025 年 5 月由 NFDG 领投、估值 $250M 的 $43M Series B;以及 2026-03-25 由 Index Ventures 领投、投后估值 $1.5B 的 $125M Series C,Kleiner Perkins 加入,Lightspeed、Spark 和 NFDG 回投。Series C 之后,TechCrunch 和 The SaaS News 都把已披露总融资额列为 $192M,BusinessCloud 另报约 $188M,累计融资额在发布方层面存在小幅差异。Forbes 在 2026 年 1 月报道称,Granola 正洽谈至少 $100M、估值超过 $1B 的融资;这轮最终两个月后以 $125M、$1.5B 估值完成,说明流程中估值和金额都有上调。所有已披露资本都是风险股权;已审阅来源没有出现债务、venture debt 或可转工具。Companies House 的 PSC 文件显示,一位联合创始人在 2023 年 4 月发生稀释事件,更像第一轮机构融资,而非任何离任。仅凭公开证据很难评估入场纪律:Granola Labs Ltd 的英国小公司审计豁免使其申报账目既不披露营业额,也不披露利润,因此价格无法对照公司自身财务基础核验;公司和投资人也没有公开给出下一轮触发条件、runway 数字或退出时间表。[CV001, CV002, CV003, CV004, CV006, CV007]
8.2 投资论点与反论点
Index Ventures 支持的论点是自下而上的:个人用户成为布道者,使用从个人扩散到团队,再到整个企业。证据包括 Brex 的 land-and-expand 模式,以及横跨 Vanta、Zapier、Vercel、Asana、Gusto、Thumbtack、Cursor、Lovable、Decagon、Mistral AI 的具名参考客户名单,再叠加 Bloomberg 报道的 Series C 前季度收入增长 250%。反论点有三条彼此不同的线。第一,财务完全不透明:本报告审阅的证据中没有任何收入、ARR、毛利率、现金余额或烧钱数字,审计豁免也封住了近期被动披露的路径。第二,产品商品化:Fathom 和 tl;dv 已经匹配 Granola 的无机器人定位,Microsoft Teams、Google Meet 和 Zoom AI Companion 又把会议笔记 AI 分发到买方已经拥有的套件里,且没有增量成本,压窄了论点依赖的差异化窗口。第三,可传导的法律风险:Otter.ai 正面临一项合并联邦集体诉讼,被指在未经同意情况下录音并采集声纹;National Law Review、Bloomberg Law、Goodwin Procter 的法律评论把整个 AI 记事类别——包括更不显眼、无机器人的工具——都视为暴露于相同 wiretap 和生物隐私法规之下。三条反论点单独看都不足以否决投资,但合在一起意味着多头案例依赖的信号(增长率、投资人质量、采用模式)方向上为正,却没有任何一项能对照 Granola 自己的数字独立验证。[CV016, CV017, CV018, CV019, CV020]
| 论点 | 支撑 / 削弱证据 | 改变判断的信号 |
|---|---|---|
| 正向论点:自下而上的病毒式采用转化为持久企业收入 | Index Ventures 将有机、内部布道者主导的采用描述为核心增长驱动;Brex 式先落地再扩张模式在多个章节被引用(CV016, CV017)。 | 披露 logo 级留存或扩张收入数据,确认该模式能从少数具名账户推广到更广客户。 |
| 正向论点:品类级 AI 估值溢价支撑偏高的私募估值 | Carta 数据显示,同阶段 AI 初创公司相较非 AI 同行普遍享有大幅估值溢价(CV024, CV025)。 | 有证据表明该溢价属于 Granola 特有,而不是可能退潮的品类大潮。 |
| 反向论点:财务完全不透明 | 本报告任何位置都未披露收入、ARR、利润率或烧钱数字;英国小公司审计豁免封死了短期强制披露路径(CV010, CV011, CV018)。 | 披露 ARR、经审计账目,或投资者来源的收入估计。 |
| 反向论点:产品商品化 | Fathom 和 tl;dv 已匹配 Granola 的无 bot 定位;Microsoft、Google、Zoom 把会议笔记 AI 捆进买家已拥有的套件(CV019)。 | 有证据表明 Granola 的上下文 / API 层能制造转换成本,且这种成本不是单靠无 bot 采集就能形成。 |
| 反向论点:可传导的法律 / 监管风险 | Otter.ai 面临合并审理的窃听 / 生物识别集体诉讼;法律评论认为整个笔记品类都有敞口,包括不那么显眼的无 bot 工具(CV020, CV042, CV045)。 | 在明确监控窗口(如 12 个月)后,仍没有直接点名 Granola 的窃听 / 同意投诉。 |
每行把正向或反向论点、背后的具体证据,以及需要跟踪的证实 / 证伪信号配对。
[CV016, CV017, CV024, CV025, CV010, CV011]8.3 牛市、基准与熊市情景
牛市情景要求已报道的 250% 季度增长率在近期延续,Brex 式 land-and-expand 模式能推广到具名企业客户名单,并且 Series C 同步推出的个人 / 企业 API 和管理后台能带来真实扩张收入;如果公司最终披露 $100M ARR 基础,套用分析师引用的 8x-20x Applied-AI 倍数区间,将支持 $1.5B-$2.0B 以上结果,与实际 Series C 价格相符。基准情景假设增长放缓到后高速增长 SaaS 常见的 40-60% 同比区间,下一轮定价持平到小幅上行;这与 Carta 2026 年 Q1 数据所示 Series C 一级估值全行业同比上涨 12.5% 相一致。熊市情景假设增长急剧放缓,同时 Carta 显示的 11.4% 全行业 down-round 比例,以及针对最接近直接同业 Otter.ai 的现行 wiretap 诉讼先例,共同把下一轮融资推向持平或下行。两组类别层面数据框定三种情景:Carta 显示 AI Series A 公司定价中位数为 $300M,非 AI 同业为 $55M,证明 Granola 所处轮次内有广泛 AI 溢价;Finro 2026 年 Q1 数据则显示,公开市场「Productivity Tools」倍数在压缩,同一类别的私募轮仍继续拿到溢价——这种背离之所以重要,正因为 Granola 仍是私营公司。[CV016, CV019, CV021, CV022, CV023, CV024]
| 情景 | 关键假设 | 估值 / 回报逻辑 | 关键风险 | 概率信号 |
|---|---|---|---|---|
| 乐观 | 250% 季度增长短期延续;Brex 式先落地再扩张能推广至具名企业客户名单;新 API / 管理后台带来扩张收入(CV016, CV021)。 | 下一轮估值达到或高于 $1.5B-$2.5B;如果最终披露 $150M-$300M ARR 基数,这大致贴合市场分析师引用的 AI 应用倍数区间(8x-20x ARR)(CV021)。 | 没有披露 ARR 可确认增长持续性;Fathom/tl;dv/捆绑既有厂商带来商品化(CV019)。 | 与 Bloomberg 报道的增长延续一致,但这种超高增速通常会随时间减速;也没有第二个数据点确认。 |
| 基准 | 增长放缓至后高增长 SaaS 常见的 40%-60% 同比区间;鉴于 Carta 显示 Series C 估值同比上涨 12.5%,下一轮价格持平至小幅上行(CV022, CV025)。 | 下一次一级或二级交易中,估值守在 $1.5B-$2.0B 区间;公司到 2027-2028 年仍保持私有(CV022)。 | 即使增长维持,功能商品化也会压缩定价权(CV019)。 | 符合 Carta 覆盖的后期估值通胀大趋势,但不能证明 Granola 特有持久性(CV025)。 |
| 悲观 | 增长急剧放缓;窃听 / 同意诉讼先例(Otter.ai)外溢至 Granola;审计豁免备案继续让任何机构后续尽调拿不到财务数据(CV023, CV042, CV010)。 | 下一轮融资估值持平或下调,与 Carta 11.4% 的行业降价轮比例一致;低于 $1.5B 的二级市场标记变得可行(CV023)。 | 在资本获取吃紧前,单位经济性可能始终得不到独立验证(CV011, CV028)。 | 基于 Carta 的实时降价轮基准率,以及针对直接可比 bot 型同行的活跃诉讼先例(CV023, CV042)。 |
估值区间是基于品类级倍数和增长率基准(Carta、Finro、SaaS 基准数据)搭出的情景框架;由于没有披露 ARR / 收入基数,它不是 Granola 特有的贴现模型。
[CV016, CV019, CV021, CV022, CV023, CV024]以市场分析师提到的 8x-20x 应用型 AI 倍数区间,展示不同假设 ARR 下的隐含估值;Granola 未披露 ARR,因此这些只是情景输入,不是拟合模型。
仅为示意敏感性网格,依据分析师评论中讨论的品类倍数区间搭建;Granola 未披露 ARR,因此没有任何一行代表已确认的公司数据点。
[CV026, CV031, CV024]基于品类层面的降轮、估值膨胀与倍数收缩信号,给出 Granola 下一轮定价结果的低 / 基准 / 高情景框架。
由于 Granola 未披露收入基数,悲观 / 基准 / 乐观三行只是与品类基准挂钩的情景框架,不是折现现金流或 ARR 倍数输出;Zoom 行仅作为公开市场规模锚。
[CV023, CV042, CV022, CV025, CV016, CV034]8.4 可比公司集:上市、私营与 M&A
没有单一可比公司能同时匹配 Granola 的类别和阶段,因此本节用一组精选对象三角定位。直接竞争对手 Fireflies.ai 在 2025 年 6 月通过 tender offer 达到 $1B 以上估值,而非一级融资;同时称其自 2023 年以来一直盈利。直接竞争对手 Otter.ai 在 2025 年 3 月披露 ARR 超过 $100M,但本节审阅证据中没有可印证的当前估值。相邻 AI 生产力公司 Notion 在 2025 年 12 月员工股票出售中最新估值 $11B,是一个规模大得多的上限参照。上市可比公司 Zoom Communications 截至 2026 年中市值 $25.31B、过去十二个月收入 $4.93B(约 5.1x 市销率)、毛利率 77.0%——成熟、低增长、盈利的画像与 Granola 所称高速增长私营阶段在结构上不同,但可作为利润率和倍数锚点。M&A 方面,Salesforce 在 2026 年 6 月同意以约 $3.6B 收购 AI 客服平台 Fin,显示战略买家愿意为已验证、已企业部署的 AI agent 平台支付的价格;虽属相邻但不同垂直领域,背景则是 CNBC 所称 2025 年全球 M&A 创下 $4.9T 纪录,并把 AI 驱动的动能带入 2026 年。在这组对象中,Granola 是唯一一家估值建立在零披露收入或 ARR 数字之上的公司。[CV029, CV030, CV031, CV032, CV033, CV034]
| 可比对象 | 指标 | 倍数 / 估值 / 状态 | 参照意义 | 局限 |
|---|---|---|---|---|
| Granola(标的公司) | $125M Series C 轮,2026-03-25 | $1.5B 投后估值;未披露 ARR / 收入,因此无法计算倍数(CV004, CV011)。 | 本报告标的;实际成交价格。 | 没有财务基底可校验价格;估值只靠增长率信号和投资者质量支撑。 |
| Fireflies.ai(直接竞争对手) | 要约收购,2025-06-12 | 估值 $1B+;自 2023 年起盈利;2021 年以来未再募集一级资本(CV029, CV030)。 | 能力最接近的可比公司;通过二级流动性而非一级融资轮达到独角兽估值。 | 估值依据是要约收购,不是公平交易的一级融资;未披露 ARR 倍数。 |
| Otter.ai(直接竞争对手) | ARR 披露,2025 年 3 月 | 已披露 $100M+ ARR;本轮研究未能佐证当前估值(CV031、CV033)。 | 直接同业中披露收入数据最强的品类龙头。 | 没有估值数字,无法计算倍数;最后一次已知估值标记来自 2021 年,本次未确认。 |
| Notion(相邻 AI 生产力) | 员工股份出售,Dec 2025 | $11B 估值(CV034)。 | 相邻 AI 生产力赛道里规模大得多的可比私营公司,可作为品类估值上限参照。 | 产品更宽,覆盖工作区 / 数据库,不是可一一对应的点解决方案;规模差距约一个数量级。 |
| Zoom Communications(上市可比公司) | 市值 / TTM 收入,mid-2026 | $25.31B 市值对应 $4.93B 收入(约 5.1x P/S);毛利率 77.0%(CV035、CV036)。 | 成熟上市协作软件公司,已打包 AI 会议功能。 | 低增长(约 5% YoY)、盈利、上市公司画像,与 Granola 声称的私营阶段超高增长不具可比性。 |
| Salesforce / Fin(M&A 先例) | 收购,2026-06-15 | 服务 30,000+ 家企业的 AI 客服智能体平台,收购价约 $3.6B(CV037)。 | 说明在相邻品类中,已验证且已部署到企业的 AI 智能体平台,战略买家愿意出到什么价格。 | 垂直领域不同(客服 vs. 会议笔记);未披露收购价 / 收入倍数。 |
| Applied AI / Productivity Tools 品类(分析师基准) | Q1 2026 倍数分散度 | 同一品类上市公司倍数压缩时,私营「Productivity Tools」估值仍在上升(CV026、CV027)。 | 品类层面背景:即便公开市场已开始下修估值,私人资本仍愿为生产力 AI 支付溢价。 | 聚合品类数据,不是公司特定倍数;Granola 未披露 ARR 基数,无法映射。 |
覆盖范围是一组经筛选、非穷尽的可比样本,横跨直接竞争对手、相邻私营可比公司、上市可比公司、M&A 先例和品类层面的分析师基准;选择依据是与直接同业的贴合度,以及本章和前文审阅来源中的数据可得性。
[CV004, CV011, CV029, CV030, CV031, CV033]8.5 退出准备度与最终尽调问题
Crunchbase 的 2026 年 IPO 候选方法依赖已披露融资、增长信号、投资人组合和市场时机;其重点候选公司,例如 Databricks(收入 run-rate $4.8B、同比增长 55%+、净留存 140%+)和 Canva(年化收入 $3.3B、月活用户 240M),展示了接近 IPO 准备度所需的规模与披露画像。Granola 没有披露这一框架依赖的任何指标,因此无法按同一基准评估其公开上市准备度,也没有来源证明具体的下一轮融资触发条件或时间线。补上这一缺口,是单项价值最高的尽调动作:一旦披露 ARR 或收入数字,本分析就能用 $1.5B 价格计算实际倍数,而不是只从类别基准推理。除收入披露外,剩余高价值尽调问题还包括 cap table 与 preference stack 细节(理解四轮融资中嵌入的下行保护)、毛利率和单位经济可见度(鉴于依赖第三方 LLM 推理与转写供应商)、持续监测 wiretap/生物识别诉讼暴露(最接近直接同业已有活跃先例)、runway 与下一轮时间的清晰度,以及经验证的当前用户数或客户数,用来把每用户估值与 Fireflies 披露的 20M 用户和 Otter 披露的 35M+ 用户基础做合理性校验。[CV039, CV040, CV011, CV013, CV048, CV042]
| 主题 | 缺失证据 | 为什么重要 | 尽调路径 |
|---|---|---|---|
| 收入 / ARR 披露 | 当前年度 ARR 或过去 12 个月收入,附套餐分层拆分(CV011)。 | 没有财务基数,就算不出倍数,也无法给 250% 增长说法定量。 | 在 NDA 下直接向公司或领投方索取当前 ARR 和同期群收入。 |
| 股权结构和清算优先权堆栈 | seed / Series A / B / C 各轮清算优先权叠加细节(CV013)。 | 优先权悬挂会改变普通股 / 创始人和后续投资人的下行保护与有效入场价。 | 通过公司或 Companies House 文件索取股权结构表摘要或确认声明细节。 |
| 毛利率和单位经济 | 收入成本、毛利率、CAC 和净收入留存数据(CV048)。 | 判断 Granola 在依赖 LLM / 转录供应商时,能否守住 SaaS 常见的 70%+ 毛利率。 | 在 NDA 下向公司索取,或跟踪下一次投资人披露的基准。 |
| 诉讼敞口监控 | 确认是否已有将 Granola 列名的窃听 / CIPA / 生物识别投诉(CV042、CV045)。 | Otter.ai 面临的直接同业诉讼说明,风险已在品类层面活跃,并非假设。 | 每季度跟踪法院文件检索和法律媒体报道。 |
| 下一轮触发点和现金跑道 | 已披露现金跑道、烧钱率或计划中的 Series D 时间(CV028)。 | 判断 $125M Series C 能在下一次重估前买来 18 个月,还是 4+ 年跑道。 | 向公司或领投方索取跑道 / 烧钱披露。 |
| 当前用户 / 客户数 | 经过验证的当前席位或客户数,需超过 Series A 时披露的 5,000 周活用户数字(CV031、CV030 提供竞争对手规模参照)。 | 需要用它和 Fireflies(20M 用户)及 Otter(35M+ 用户)对照,校验每用户估值是否合理。 | 直接向公司索取当前 MAU / 付费席位数。 |
按可能改变估值结论的程度做非正式排序;尽调路径假设公司或领投方配合,因为这些数据今天都未公开披露。
[CV011, CV013, CV048, CV042, CV045, CV028]8.6 建议、风险评级与论点破裂触发器
一边是 $1.5B 估值却零收入披露,一边是私募市场仍普遍支撑 AI 倍数上行,但 Granola 相邻的公开市场「Productivity Tools」类别倍数已经开始压缩,同时全行业 down-round 比例仍有 11.4%。权衡后,证据支持继续研究建议:中等信心、高风险评级、估值偏紧——不是买入,也不是回避,因为增长和采用信号方向上真实,尽管无法验证。六个论点破裂触发器会迫使立即重估:直接点名 Granola 的 wiretap 或生物隐私投诉;披露 down round 或低于 $1.5B 的二级交易;下一轮融资周期仍然零 ARR/收入披露;竞争对手以更低价格推出显著更强的无机器人 + API 技术栈;首份审计财务显示利润率或烧钱承压;或战略收购报价低于上一轮一级融资价格。截至本次运行的证据,这些触发器都尚未触发;但每一项都可以独立监测,一旦任一发生,建议应立刻重审。[CV049, CV042, CV023, CV011, CV019, CV010]
| 维度 | 值 | 理由 |
|---|---|---|
| 建议 | 继续研究 | Granola、其投资者或任何第三方数据提供商均未披露收入、ARR、毛利率或烧钱数字;缺少 Granola 特有财务基底,无法独立为 $1.5B 估值定价(CV004, CV011)。 |
| 置信度 | 中 | 融资历史和可比公司事实得到多个独立来源充分交叉验证,但支撑估值的核心输入(收入、增长基数、利润率)仍完全未披露(CV004, CV011)。 |
| 风险评级 | 高 | 针对直接 / 相邻同行的法律先例(Otter.ai 窃听诉讼)、11.4% 的行业实时降价轮比例,以及财务完全不透明,共同构成高风险画像(CV023, CV042, CV045)。 |
| 估值立场 | 偏高 | 约十个月估值增长六倍,却零收入披露;同时公开市场 'Productivity Tools' 倍数被压缩。相较现有可比公司,这一估值显得激进(CV008, CV026, CV049)。 |
| 决策含义 | 跟踪;在 ARR / 毛利率披露,或有定价后续融资轮缩小估值支撑缺口前,不投入新的一级资本。 | 在下一次披露融资事件,或任何经审计账目备案时重启评估(CV010, CV028, CV049)。 |
Recommendation/confidence/riskRating/valuationStance 与 report-meta.yaml 中记录的枚举值一致;本表是在章节层面、带来源引用推导出的判断。
[CV004, CV011, CV023, CV042, CV045, CV026]| 触发项 | 阈值 | 对投资判断的影响 | 行动含义 |
|---|---|---|---|
| 直接点名 Granola 的窃听 / CIPA 或生物识别隐私诉讼 | 任何将 Granola 列为被告的起诉 | 侵蚀作为差异化核心的无 bot 信任定位;先例显示 Otter.ai 面临同类诉讼(CV042、CV045)。 | 将风险评级下调至 critical;立即重估建议。 |
| 披露降价轮,或可验证的二级交易低于 $1.5B | 任何定价轮或经佐证的二级交易低于 Series C 标记 | 推翻 Series C 定价背后的增长叙事;确认熊市情景下修估值(CV023)。 | 重写估值判断,转为 expensive/avoid。 |
| 下一轮融资周期前仍未披露 ARR / 收入 | 预计 Series D 窗口(约 2027)之前仍零财务披露 | 当前投资判断所依赖的增长说法无法独立核验(CV011、CV028)。 | 维持 research-more/track;不要上调到 buy。 |
| 竞争对手以更低价格推出明显更强的无 bot + 企业 API 组合 | Fathom、tl;dv、Otter 或捆绑型既有厂商在两个价格档内达到功能平价 | 核心价值主张沦为商品,任何溢价倍数都难以支撑(CV019)。 | 下调估值立场;重估竞争护城河。 |
| 首份审计或投资人披露财务显示毛利或烧钱承压 | 首份披露财务显示毛利率显著低于约 70% SaaS 下限,或现金跑道偏短 | 把今天的不透明风险变成硬数据点,且可能为负面(CV010、CV048)。 | 用真实财务重新跑估值模型;建议可能上下调整。 |
| 战略收购接洽报价低于上一轮一级市场融资 | 任何可信报道的收购接洽低于 $1.5B | 说明市场真实出清价低于私募标记(CV037 提供验证充分时战略买家愿付价格的量级参照)。 | 视为熊市情景的外部确认。 |
上述触发项是可监控、前瞻性的条件,不是截至本次研究已经发生的事件;阈值是定性判断,不是合同约定。
[CV042, CV045, CV023, CV011, CV028, CV019]从已披露的规模 / 增长信号出发,串起尚未解决的证明缺口和风险传导,最终落到推荐结论。
[CV016, CV011, CV024, CV025, CV026, CV042]面向 IC 的评分,覆盖市场、验证、护城河、经济性、风险、估值和证据质量。
[CV024, CV025, CV011, CV019, CV048, CV042]8.7 附录
免责声明
本报告是基于公开证据的尽调快照,不构成投资建议。重要的财务、法律、技术和合同事实仍未公开;任何投资决策前,都应向管理层和一手文件直接核验。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | Granola describes itself as an AI notepad for back-to-back meetings that provides notes, actions, and memory without a meeting bot. | 中 | SO001 |
| CO002 | Independent March 2026 coverage places Granola in London, United Kingdom. | 中 | SO011, SO012 |
| CO003 | Granola was founded in March 2023. | 中 | SO009, SO012 |
| CO004 | Granola was founded by Chris Pedregal and Sam Stephenson. | 中 | SO009, SO012 |
| CO005 | BusinessCloud reports that Chris Pedregal founded Socratic before starting Granola and that Socratic was acquired by Google in 2018. | 中 | SO009 |
| CO006 | BusinessCloud also reports that Pedregal later built Stack inside Google’s Area 120 incubator before the product was folded into Google Drive in 2022. | 中 | SO009 |
| CO007 | Public fundraising coverage identifies Sam Stephenson as Granola’s cofounder and a visible product/design leader. | 中 | SO012, SO015 |
| CO008 | Granola says its product combines user-typed notes with AI transcriptions to generate structured recaps and follow-ups. | 高 | SO001, SO014, SO016 |
| CO009 | Granola positions bot-free local or on-device capture as a core product differentiator relative to visible meeting-bot note takers. | 高 | SO001, SO004, SO006 |
| CO010 | As of the run date, Granola is available on macOS, Windows, and iPhone. | 高 | SO001, SO013 |
| CO011 | Granola’s mobile app can sync with Google Calendar and start notes or phone-call transcription from the iPhone app. | 中 | SO013, SO004 |
| CO012 | Granola became a late-stage private unicorn after its March 2026 Series C valued the company at $1.5 billion. | 高 | SO002, SO006, SO009 |
| CO013 | TechCrunch reported that Granola had raised $4.25 million in seed funding before the Series A. | 中 | SO015 |
| CO014 | Granola announced a $20 million Series A led by Spark Capital on 2024-10-23. | 高 | SO014, SO015 |
| CO015 | TechCrunch reported that Granola had about 5,000 weekly users by October 2024. | 中 | SO015 |
| CO016 | TechCrunch reported that about 70% of users who tried Granola in week one returned to use it again for another meeting. | 中 | SO015 |
| CO017 | Granola raised $43 million in Series B funding led by NFDG in May 2025 at a $250 million valuation. | 高 | SO016, SO017 |
| CO018 | The May 2025 Granola 2.0 launch introduced shared team folders and an AI workspace for teams. | 高 | SO016, SO017 |
| CO019 | By May 2025 Granola said it was transcribing and analyzing millions of minutes of conversation each day. | 高 | SO016, SO003 |
| CO020 | TechCrunch reported in May 2025 that Granola’s user base had been growing roughly 10% per week since launch. | 中 | SO017 |
| CO021 | Granola announced a $125 million Series C led by Index Ventures on 2026-03-25 at a $1.5 billion valuation. | 高 | SO002, SO006, SO007 |
| CO022 | Kleiner Perkins, Lightspeed, Spark, and NFDG all participated in Granola’s March 2026 Series C. | 高 | SO002, SO006, SO012 |
| CO023 | TechCrunch and The SaaS News both reported that Granola’s total funding reached $192 million after the Series C. | 高 | SO006, SO012 |
| CO024 | Granola’s March 2026 product expansion included team Spaces or folders with granular access controls. | 高 | SO006, SO008 |
| CO025 | Granola rolled out personal and enterprise APIs in March 2026 to expose meeting data to external workflows and tools. | 高 | SO006, SO010, SO012 |
| CO026 | Granola introduced Recipes in Granola Chat on 2025-09-30 as repeatable AI workflows built on meeting context. | 中 | SO027 |
| CO027 | Granola launched MCP on 2026-02-04 to connect meeting notes with external AI apps. | 中 | SO026 |
| CO028 | Granola launched Briefs on 2026-05-20 to prepare users with live summaries of who they are meeting and what matters now. | 中 | SO028 |
| CO029 | Investor and company materials increasingly frame Granola as a system of record or context layer for company conversations rather than only a personal notepad. | 中 | SO007, SO008, SO016 |
| CO030 | Public March 2026 materials name customers or users such as Vanta, Gusto, Thumbtack, Asana, Cursor, Lovable, Decagon, and Mistral AI. | 高 | SO002, SO006, SO008 |
| CO031 | Granola’s Vanta case study says the product saves more than 260 hours annually per user. | 中 | SO019 |
| CO032 | Granola’s Brex case study says the product saves 427 hours per week. | 中 | SO021 |
| CO033 | Granola’s Zapier case study says the product saves more than 830 hours per month. | 中 | SO022 |
| CO034 | Granola’s Vercel case study says searchable transcripts and AI recaps help GTM staff act on prospect feedback in minutes rather than hours. | 中 | SO020 |
| CO035 | Granola’s privacy policy says third parties such as OpenAI and Anthropic cannot use customer personal data to train their AI models. | 中 | SO023 |
| CO036 | The same privacy policy says Granola may use de-identified data for model training unless users opt out in account settings. | 中 | SO023 |
| CO037 | Granola’s DPA says the company supports GDPR, UK GDPR, Swiss data-protection rules, and U.S. state privacy laws such as CCPA. | 中 | SO024 |
| CO038 | Granola’s privacy policy identifies the operating entity as Granola, Inc. at 1151 Walker Road, Suite 417, Dover, Delaware 19904. | 中 | SO023 |
| CO039 | The reviewed 2024-2026 official and independent sources do not publicly disclose Granola’s current revenue or ARR. | 低 | SO002, SO006, SO017 |
| CO040 | The reviewed official and independent sources do not publicly disclose Granola’s exact current employee count. | 低 | SO003, SO006, SO012 |
| CO041 | The reviewed official and independent sources do not publicly disclose Granola’s current paying-customer count or seat count. | 低 | SO002, SO006, SO012 |
| CO042 | On 2026-03-17 public users criticized Granola for encrypting its local database and breaking local AI-agent workflows. | 中 | SO025, SO006 |
| CO043 | TechCrunch reported that Granola’s 2026 API rollout was partly a response to complaints from users whose local workflows broke after a storage change. | 中 | SO006 |
| CO044 | Public governance disclosure appears limited because the fetched official and news surfaces focus on the two founders and do not identify an independent board roster or clearly named CFO/COO bench. | 低 | SO003, SO009, SO012 |
| CO045 | Customer proof and investor commentary consistently describe bottom-up adoption that starts with individual users and spreads into teams. | 中 | SO007, SO014, SO019 |
| CO046 | Granola says the product works with Google Workspace or Microsoft login and highlights an engineering team with backgrounds from Apple, Amazon, Google, and Meta. | 中 | SO004, SO003 |
| CM001 | Granola markets itself as an AI notepad that captures meeting audio and user-typed notes without adding a visible bot participant to the call. | 中 | SM001 |
| CM002 | Granola's Desktop app works with any videoconferencing platform, including Zoom, Google Meet, and Microsoft Teams, while its iPhone app is built for in-person meetings. | 中 | SM008 |
| CM003 | Granola's March 2026 Series C launch added Spaces, personal and enterprise APIs, and an updated MCP so teams can access shared meeting context outside the core notepad. | 中 | SM007 |
| CM004 | Microsoft 365 Copilot in Teams offers a built-in meeting recap and catch-up capability bundled inside existing Teams/Microsoft 365 subscriptions. | 高 | SM020, SM021 |
| CM005 | Google Workspace bundles a "Take notes for me" AI note-taking feature into Google Meet via Gemini, saving the transcript automatically into Google Docs and Drive. | 中 | SM019 |
| CM006 | Zoom AI Companion includes a Meeting Summary feature available to Zoom Workplace Pro and higher-tier licensed users at no separate charge beyond the existing plan. | 中 | SM022 |
| CM007 | Fireflies.ai advertises usage across more than 1 million companies, indicating a large existing installed base in the adjacent AI meeting-transcription category. | 中 | SM017 |
| CM008 | Otter.ai markets calendar-based auto-join capture for Zoom, Microsoft Teams, and Google Meet, positioning it as a bot-based alternative to Granola's bot-free capture model. | 中 | SM018 |
| CM009 | Fathom markets both bot and bot-free capture modes and positions its product for teams ranging from a single user to 1,000 employees. | 中 | SM015 |
| CM010 | Gong, a dedicated sales conversation-intelligence platform, is priced separately from lightweight AI notepads at roughly $1,300 to $1,600 per user per year plus a mandatory platform fee of $5,000 to $50,000 per year, a materially different job-to-be-done and price point than Granola's notepad category. | 中 | SM012 |
| CM011 | Some privacy-sensitive users have responded to Granola's local-database encryption by switching to plain local markdown-file note-taking workflows (e.g., paired with Obsidian) as a status-quo substitute for a dedicated AI notepad vendor. | 低 | SM031 |
| CM012 | Grand View Research (summarized via GII Research) estimates the global AI meeting assistant market at USD 3.47 billion in 2025, growing to USD 21.48 billion by 2033 at a 25.8% CAGR. | 中 | SM027 |
| CM013 | Statifacts estimates the global AI meeting assistant market at USD 3.50 billion in 2025, USD 4.41 billion in 2026, USD 17.55 billion by 2032, and USD 35.02 billion by 2035, a 25.9% CAGR. | 中 | SM025 |
| CM014 | The Business Research Company estimates the AI-powered meeting assistants market reached USD 3.14 billion in 2025 and will grow to USD 9.33 billion by 2030 at a 24.3% CAGR. | 中 | SM026 |
| CM015 | Precedence Research estimates a narrower "AI note-taking market" at USD 623.50 million in 2025, growing to USD 3,476.74 million by 2035 at an 18.75% CAGR (2026-2035). | 中 | SM024 |
| CM016 | Congruence Market Insights estimates the global AI note-taking solutions market at USD 450.7 million in 2024, growing to USD 1,800.3 million by 2032 at an 18.9% CAGR. | 中 | SM028 |
| CM017 | A PitchBook analysis cited by Startup Fortune valued the global AI transcription and meeting intelligence market at USD 3.5 billion as of a June 2024 analysis, projecting growth to USD 11.8 billion by 2030. | 低 | SM029 |
| CM018 | Broad-lens analyst estimates that bundle transcription, scheduling, and meeting-analytics features (the "AI meeting assistant" category) put the 2025 market roughly 5 to 7 times larger than narrow-lens estimates that isolate standalone "AI note-taking" software specifically. | 中 | SM025, SM026, SM027, SM024, SM028 |
| CM019 | No public source in this research set discloses Granola's own revenue, ARR, or paid-seat count, so a company-specific SAM or SOM figure cannot be derived from public data and must be treated as an open diligence question. | 低 | |
| CM020 | Granola's direct fetch of the primary Grand View Research report page returned a bot-challenge ("Just a moment...") response rather than the underlying report, so the chapter relies on a syndicated GII Research summary of that report rather than the primary source. | 中 | SM034 |
| CM021 | All five analyst market-sizing lenses reviewed (Grand View/GII, Statifacts, The Business Research Company, Precedence Research, Congruence Market Insights) project double-digit compound annual growth for AI meeting-assistant or note-taking software through the early 2030s, ranging from 18.75% to 25.9% CAGR. | 中 | SM024, SM025, SM026, SM027, SM028 |
| CM022 | Analyst publishers differ on both the category boundary ("meeting assistant" vs. "note-taking software") and the base year and forecast horizon used, so their headline figures are not directly interchangeable without adjustment. | 中 | SM024, SM025, SM026, SM027, SM028 |
| CM023 | Brex's Head of IT, Lloyd Cilliers, formalized a company-wide Granola rollout only after discovering organic, employee-led adoption already spanning sales, engineering, and support, and required SOC 2 Type 2 compliance, SSO, and SCIM before approval. | 中 | SM002 |
| CM024 | Vanta's VP of Customer Success and Account Management adopted Granola to replace manual reconciliation of live notes and raw transcripts across daily 1:1s, standups, and cross-functional syncs. | 中 | SM003 |
| CM025 | Zapier's Principal Designer adopted Granola to build a searchable, shared knowledge repository across design and cross-functional meetings, replacing individually-taken, inconsistent notes. | 中 | SM004 |
| CM026 | A Vercel sales engineer adopted Granola after a prior transcription tool proved too slow to generate searchable call recaps in time to send prospects same-day follow-ups. | 中 | SM005 |
| CM027 | Granola's product use-case page for Product teams segments buyers into Product Managers, Product Designers, and Product Leaders, each with a distinct customer-research or roadmap workflow. | 中 | SM013 |
| CM028 | Granola's sales use-case page markets direct CRM sync integrations with HubSpot, Affinity, Attio, and Zapier for Account Executive buyers. | 中 | SM014 |
| CM029 | Granola's own guidance segments the sales-tooling market by company size: founders and SMB sales teams under 15 reps are framed as the fit for a lightweight AI notepad, while teams with 30+ reps and a dedicated sales-enablement manager are directed toward Gong-style conversation intelligence instead. | 中 | SM012 |
| CM030 | Granola's second-brain blog post frames venture capital partners as a distinct buyer segment, using captured meeting context to accelerate investment-committee memo writing and preserve deal context when associates leave the firm. | 中 | SM009 |
| CM031 | Executive search firm Daversa Partners adopted a bot-free meeting-capture approach across 136 of its 150 employees specifically because a visible bot participant was considered intrusive during confidential CEO search conversations. | 中 | SM010 |
| CM032 | Granola's Business plan is priced at $14 per user per month for unlimited meeting history, CRM integrations, and folder-level queries across a team's conversations. | 中 | SM009 |
| CM033 | Fathom's publicly listed pricing spans a free tier up to roughly $34 per user per month across its paid plans, giving buyers a self-serve, low-commitment adoption path comparable to Granola's. | 中 | SM016 |
| CM034 | Granola's enterprise page lists SSO/SAML, SCIM provisioning, domain management, an admin dashboard, and SOC 2 Type II certification as the specific controls that let IT and security buyers formalize a rollout that began through individual, bottom-up adoption. | 中 | SM001 |
| CM035 | Microsoft's telemetry-based Work Trend Index research found that the average Microsoft 365 user is interrupted every two minutes during core work hours (275 times a day) by meetings, emails, or chats. | 中 | SM023 |
| CM036 | The same Microsoft research found that 57-60% of meetings are unscheduled or ad hoc, and nearly a third of meetings now span multiple time zones, up 35% since 2021. | 中 | SM023 |
| CM037 | Gallup survey data cited in Granola's hybrid-teams guide shows 52% of US remote-capable employees work in a hybrid arrangement, 26% work fully remote, and only 22% work entirely on-site, sustaining demand for tools that capture both in-room and remote meeting audio. | 中 | SM010 |
| CM038 | An IDC-sourced estimate cited by Granola places the annual financial cost of lost institutional knowledge at $31.5 billion across Fortune 500 companies, with 42% of institutional knowledge residing solely with individual employees. | 低 | SM009 |
| CM039 | Granola's March 2026 product launch extended the product from a personal notepad into team-wide Spaces, a personal API, an enterprise API, and an updated MCP, explicitly aimed at expanding the addressable workflow beyond individual note-taking into shared organizational context and third-party agent integrations (Claude, ChatGPT, Replit, Figma Make, and others). | 中 | SM007 |
| CM040 | Startup Fortune reported that Granola exposes notes to anyone holding a share link by default and feeds meeting transcripts into an internal AI training pipeline unless a user manually opts out, citing a prior report by The Verge. | 中 | SM029 |
| CM041 | Granola's own security page confirms that its transcript and note data are stored only in a US-hosted AWS Virtual Private Cloud, and that anonymized user data is used to improve the product by default unless a user opts out, with enterprise accounts defaulting to no model training. | 中 | SM008 |
| CM042 | BuildBetter's independent review gave Granola a 3-out-of-10 "Enterprise Readiness Rating," citing US-only AWS hosting with no EU region or on-premises/VPC option as a primary gap for regulated or EU-headquartered buyers. | 中 | SM030 |
| CM043 | In March 2026, Granola encrypted its local notes database, removing local file and API access to stored meeting data; the change was publicly criticized by a16z partner Guido Appenzeller in a widely viewed social media thread, prompting some users to evaluate switching to open, markdown-based alternatives. | 中 | SM031 |
| CM044 | Microsoft, Google, and Zoom each bundle an AI meeting-notes or recap feature directly into their existing Teams, Workspace, and Zoom Workplace subscriptions at no additional charge, capping the price a standalone vendor like Granola can command for note-taking as a standalone feature. | 中 | SM019, SM020, SM021, SM022 |
| CM045 | Dedicated sales conversation-intelligence platforms priced at $1,300-$1,600 per user per year plus a $5,000-$50,000 annual platform fee represent a materially higher-ACV segment that a lightweight AI notepad is not positioned to win outright, capping Granola's addressable spend per seat within the largest, most complex sales organizations. | 中 | SM012 |
| CM046 | No public source in this research set discloses how much of the analyst-estimated AI meeting-assistant market is currently captured by bundled incumbent features (Teams, Google Meet, Zoom) versus standalone vendors such as Granola, Fathom, Fireflies, and Otter, leaving the truly available standalone-paid wedge unquantified. | 低 | |
| CM047 | BuildBetter's enterprise-readiness review was published in 2025 and has not been independently re-verified against Granola's hosting and compliance posture as of the 2026-07-01 run date, so its findings may be stale relative to any post-Series-C infrastructure changes. | 低 | SM030 |
| CM048 | tl;dv's independent review frames Granola as a capable but narrower notepad relative to bot-based competitors that offer built-in video recording and broader integration libraries, an evaluation angle distinct from Granola's own bot-free positioning. | 低 | SM032 |
| CM049 | TechCrunch observed that as meeting note-taking becomes a commonplace feature, competitive value in the category is shifting toward turning notes and transcripts into actions, a trend it noted Read AI, Fireflies, and Quill are also pursuing. | 中 | SM033 |
| CP001 | Granola's own "AI Meeting Assistant - Without the Bot" page frames the whole competitive category around a single axis: whether a rival tool joins a call as a visible bot or captures audio locally without one. | 中 | SP001 |
| CP002 | Granola's own comparison chart on that page claims it is the only listed option that keeps notes private by default and works without a bot joining the call. | 中 | SP001 |
| CP003 | tl;dv's 2026 homepage advertises a "NO BOT REQUIRED" bot-free capture option and reports more than 2 million users, showing Granola's bot-free positioning is no longer unique among direct peers. | 中 | SP016 |
| CP004 | Fathom's 2026 pricing page markets a "Choice of bot-free (in beta) or bot capture" mode across its Free and Premium individual tiers, extending bot-free capture to a mainstream direct-peer product. | 中 | SP010 |
| CP005 | Otter.ai's core Business workflow still relies on a calendar-connected bot that auto-joins Zoom, Microsoft Teams, or Google Meet to record and share notes, keeping the bot-based model dominant at the largest standalone competitor by claimed user count. | 中 | SP013 |
| CP006 | TechCrunch's March 2026 coverage of physical AI notetaking devices names Granola alongside Read AI, Fireflies, and Fathom as digital competitors that hardware notetakers can substitute for in-person capture, a use case none of the software vendors serve natively. | 中 | SP031 |
| CP007 | Fathom's pricing page lists a Free tier (unlimited recordings/transcription, choice of bot-free-beta or bot capture), Team at $19 ($15 billed annually) per user/month, Business at $34 ($25 annually) per user/month, and a custom Enterprise tier with SSO/SCIM. | 中 | SP010 |
| CP008 | Fathom's Business tier also includes a public API, MCP support, and native Claude/ChatGPT integrations, positioning it directly against Granola's own API/MCP expansion. | 中 | SP010 |
| CP009 | Fireflies.ai's 2026 homepage claims usage across more than 1 million companies and advertises SOC 2 Type II, GDPR, and HIPAA compliance plus a "zero data retention" AI-training policy directly on its marketing page. | 中 | SP011 |
| CP010 | Fireflies' 2026 pricing page lists a Free plan at $0, Pro at $18 ($10 billed annually) per seat/month, Business at $29 ($19 annually) per seat/month marked "Most Popular," and an annual-only Enterprise tier at $39 per seat/month. | 中 | SP012 |
| CP011 | Otter.ai's April 2026 relaunch announcement claims more than 35 million users, over 1 billion meetings processed, use by 86% of the Fortune 500 and 88% of the Forbes Cloud 100, and more than $1.5 billion in claimed cumulative customer ROI. | 中 | SP015 |
| CP012 | That same announcement launched a "Conversational Knowledge Engine" with an MCP server and client, AI Chat Connectors into Gmail, Google Drive, Notion, Jira, and Salesforce, and an estimated addressable opportunity of more than $100 billion, directly overlapping Granola's own Spaces/API/MCP expansion strategy. | 中 | SP015 |
| CP013 | Otter.ai's pricing page lists a free Basic tier, Pro at $16.99 per user/month (promotional $8.49), Business at $30 per user/month (promotional $24), and a custom Enterprise tier requiring a minimum 100-user license for SSO. | 中 | SP014 |
| CP014 | An independent 2026 report found Otter.ai facing a consolidated federal class action alleging its meeting agent records audio, transcripts, and voiceprints without participant consent, exposing potential ECPA and state biometric-privacy liability, with a motion-to-dismiss hearing scheduled for May 20, 2026. | 中 | SP027 |
| CP015 | The same report states similar litigation already targets Fireflies.ai and that Read AI faces university-level usage bans over consent concerns, showing the consent/privacy risk spans the bot-based capture segment rather than one vendor. | 中 | SP027 |
| CP016 | tl;dv's 2026 homepage reports more than 2 million users, a 4.7-out-of-5 rating across 450 reviews, and GDPR/SOC 2 compliance badges alongside its bot-free capture option. | 中 | SP016 |
| CP017 | tl;dv's own 2026 review of Granola claims tl;dv reaches 97%-plus transcription accuracy versus roughly 94-95% for Granola, supports 30-plus languages versus roughly 15, and connects to more than 5,000 apps via Zapier plus native HubSpot/Salesforce integrations versus a smaller native integration set it attributes to Granola. | 低 | SP017 |
| CP018 | An independent 2026 pricing comparison estimates tl;dv's paid plans span roughly $29 to $98 per user per month, a wider and higher range than Fathom's, Fireflies', Otter's, or Granola's published tiers. | 中 | SP029 |
| CP019 | Read AI's 2026 homepage promotes an "Ada" digital-twin AI assistant, integrations across Google Meet, Microsoft Teams, Zoom, HubSpot, Salesforce, Notion, and Slack, and states Zoom has selected Read AI as an official "Essential App" with premium features enabled for Zoom customers. | 中 | SP018 |
| CP020 | Read AI's pricing page lists a Free tier (30 meetings measured per month), Pro at $15 ($19.75 monthly) per user/month, a mid tier at $22.50 ($29.75 monthly) per user/month, and an Enterprise tier at $39.75 per user/month requiring 5-plus licenses. | 中 | SP019 |
| CP021 | Gong's 2026 homepage brands itself the "#1 AI OS for Revenue Teams," claims more than 5,000 customers, and is expanding into agentic "Revenue Harness" and "Gong Assistant" capabilities beyond call recording and coaching. | 中 | SP020 |
| CP022 | An independent 2026 pricing breakdown estimates Gong's Foundations plan costs roughly $1,600 per user per year (negotiable to $1,000-$1,349), plus a mandatory $5,000-$50,000 annual platform fee and a $7,500-plus onboarding fee, none of which Gong publishes directly. | 中 | SP028 |
| CP023 | Microsoft's own support documentation confirms intelligent meeting recap in Teams requires either the Teams Premium add-on license or a Microsoft 365 Copilot license, and that Copilot needs live transcription turned on to answer questions about a meeting after it ends. | 高 | SP022, SP023 |
| CP024 | Microsoft's own pricing page lists Teams Premium at $10.00 per user per month billed yearly, a materially lower add-on price than Granola's $14 per user per month Business plan or most standalone competitors' mid tiers. | 中 | SP024 |
| CP025 | Google Workspace's "Take notes for me" feature captures real-time notes via Gemini in Google Meet and automatically organizes the transcript into a Google Doc attached to the Calendar event, at no incremental price beyond an existing Workspace subscription. | 中 | SP021 |
| CP026 | Zoom's own support documentation confirms Meeting Summary with AI Companion is available to any licensed user on a paid Zoom Workplace Pro/Business/Enterprise plan at no separately billed cost, while Zoom's own March 2026 announcement reports AI Companion Monthly Active Users more than tripled year over year in Q4 FY26. | 高 | SP025, SP026 |
| CP027 | Zoom's AI Companion is also available as a standalone offering outside a bundled Workplace subscription, and Zoom states it does not use customer content to train its own or third-party models. | 中 | SP026 |
| CP028 | TechCrunch's March 2026 roundup lists at least eight physical AI notetaking devices on the market -- Plaud Note/Note Pro ($159/$179), Mobvoi TicNote ($159), Comulytic Note Pro ($159), Plaud NotePin/NotePin S ($159/$179), Omi pendant ($89), Viaim RecDot earbuds ($200), Anker Soundcore Work ($159), and Pocket ($199) -- several bundling monthly subscriptions for advanced AI summaries. | 中 | SP031 |
| CP029 | Independent commentary on Granola's March 2026 local-database encryption change notes that some privacy-sensitive users respond to vendor-hosted note databases by reverting to local markdown files or generic note-taking apps rather than any AI meeting-notes vendor, making "no dedicated tool" a persistent status-quo substitute. | 中 | SP033 |
| CP030 | Otter.ai, Fathom, and Granola have each independently shipped a cross-meeting AI chat or agent layer in 2026 (Otter's AI Chat Connectors, Fathom's account-wide "Ask Fathom," and Granola's rebuilt agentic Granola Chat), showing single-meeting transcription alone is no longer a durable differentiator among direct peers. | 中 | SP015, SP010, SP002 |
| CP031 | Granola's own May 2026 blog post on agent design states it chose to have its rebuilt Chat read up to 100 meetings at a high level rather than 10 meetings in full transcript detail -- a specific architecture tradeoff the company frames as central to its agent strategy. | 中 | SP004 |
| CP032 | Granola's security documentation requires enterprise buyers to request Trust Center access to view its SOC 2 report, while Fireflies and Otter both surface SOC 2/HIPAA-related compliance claims directly on public marketing or pricing pages without a gated request step. | 中 | SP005, SP011, SP014 |
| CP033 | Across reviewed 2026 pricing pages, standalone AI meeting-notes vendors cluster into a similar $0-$40 per user per month band (Granola's own $14 Business plan, Fathom $15-34, Fireflies $10-39, Otter $17-30, Read AI $15-40), while Microsoft's Teams Premium add-on undercuts all of them at $10 per user per month and Gong's conversation-intelligence tier sits an order of magnitude higher at roughly $1,000-$1,600 per user per year plus mandatory platform fees. | 中 | SP006, SP010, SP012, SP014, SP019, SP024, SP028 |
| CP034 | Granola's own sales buyer's guide estimates Gong's Foundations plan costs $1,300-$1,600 per user per year plus a $5,000-$50,000 mandatory annual platform fee, based on third-party buyer estimates since Gong does not publish pricing directly. | 中 | SP006 |
| CP035 | Independent third-party pricing research corroborates Granola's own Gong cost estimate, separately reporting a similar $1,000-$1,600-per-user-per-year range plus the same $5,000-$50,000 mandatory platform fee. | 中 | SP028 |
| CP036 | Otter.ai's and Fireflies' 2026 pricing pages both list steep promotional discounts off list price (Otter Pro $16.99 discounted to $8.49, Business $30 discounted to $24; Fireflies Pro $18 discounted to $10, Business $29 discounted to $19), evidence of active price competition among standalone vendors that could pressure Granola's own flat $14 Business list price. | 中 | SP014, SP012 |
| CP037 | Granola's March 2026 decision to encrypt its local notes database was publicly criticized by a prominent investor on social media and covered by an independent blog arguing that once notes live behind vendor-controlled encryption, "the app now has zero value" for users who wanted local-first data portability -- a switching-cost and lock-in risk relative to rivals. | 中 | SP033 |
| CP038 | Fathom, Fireflies, and Otter each publish native integrations into CRM, project-management, and automation tools (Fathom: CRM field sync via its pricing page; Fireflies: CRM, project-management, ATS, and Slack integrations; Otter: Salesforce/HubSpot sync plus Zapier), lowering the switching cost of moving structured meeting data into a company's system of record independent of any single vendor. | 中 | SP010, SP011, SP014 |
| CP039 | Otter's 2026 relaunch explicitly positions itself as an MCP client and server that can pull from and push to Notion, Salesforce, Gmail, and Google Drive while also serving Claude and ChatGPT as external AI tools, mirroring Granola's own MCP strategy and reducing the durability of "agent connectivity" as a standalone moat. | 中 | SP015 |
| CP040 | Microsoft's Teams Premium/Copilot recap features, Google's Gemini-powered "Take notes for me," and Zoom's AI Companion are all distributed through collaboration-suite subscriptions that buyers already pay for, giving each incumbent a default-adoption advantage no standalone vendor, including Granola, can replicate without its own distribution channel. | 中 | SP022, SP021, SP026 |
| CP041 | A 2026 market-state report -- published by a competing AI meeting-assistant vendor -- states word-level transcription accuracy has reached 95%-plus on clean audio industry-wide and frames the market as stratified into platform-bundled summarizers, standalone note-takers, and vertical specialists, with the functional gap between an AI note-taker and a junior human notetaker now largely closed. | 低 | SP030 |
| CP042 | The same report separately states platform-bundled players lead by raw user count (Zoom AI Companion reaching 510,000 accounts within five months of launch and generating 7.2 million meeting summaries; Microsoft Teams Premium crossing 3 million seats) and that Fathom holds the highest reported G2 rating in the standalone category at 5.0 out of 5 across more than 6,000 reviews. | 低 | SP030 |
| CP043 | TechCrunch's 2026 competitive commentary observed that as meeting note-taking becomes a commonplace feature, competitive value in the category is shifting toward turning notes and transcripts into actions, a trend it noted Read AI, Fireflies, and Quill are also pursuing alongside Granola. | 中 | SP032 |
| CP044 | Granola's own "AI Meeting Assistant" marketing page states its notes are "private by default," with "no third-party server logging your conversations." | 中 | SP001 |
| CP045 | Independent 2026 reporting -- citing a prior investigative report -- found that Granola's default settings actually expose meeting notes to anyone holding a share link and train its models on transcripts unless the user manually opts out, directly contradicting Granola's own "private by default" marketing claim. | 中 | SP035 |
| CP046 | The same report states the broader AI meeting-assistant category -- including Otter.ai, Fireflies.ai, and Fellow -- shares a similar pattern of permissive default settings that favor feature development and model training over locked-down security, extending the privacy-default risk beyond Granola alone. | 中 | SP035 |
| CP047 | An independent enterprise-readiness review scored Granola 3 out of 10, citing US-only AWS hosting with no EU region or on-premises/VPC option -- a gap that Fireflies partially closes by advertising private, dedicated cloud storage and Fathom addresses via enterprise custom data-retention controls, but that Otter's public pricing page does not explicitly resolve either. | 中 | SP034, SP011, SP010 |
| CP048 | The 2026 competitive landscape for Granola spans five distinct classes: direct bot-free/bot-optional peers (Fathom, Fireflies, Otter.ai, tl;dv, Read AI), bundled incumbents (Microsoft Teams/Copilot, Google Workspace/Gemini, Zoom AI Companion), an adjacent higher-ACV conversation-intelligence category (Gong), likely hardware entrants (Plaud, Omi, and similar wearables), and a persistent status-quo/internal-build substitute of manual or plain-text note-taking. | 中 | SP001, SP020, SP024, SP031, SP033 |
| CP049 | Granola's March 2026 Series C raised $125 million at a $1.5 billion valuation, led by Index Ventures with Kleiner Perkins participation, bringing total disclosed funding to $192 million -- the scale reference point this chapter uses when comparing competitor funding and revenue signals. | 中 | SP032 |
| CP050 | Granola's own sales buyer's guide states its Business plan costs $14 per user per month and includes unlimited meeting history, CRM integrations, and folder-level queries across a team's conversations. | 中 | SP006 |
| CP051 | Granola's own security page states notes are stored in a US-hosted AWS Virtual Private Cloud, with no EU region disclosed in the reviewed source set. | 中 | SP008 |
| CP052 | Granola's own sales buyer's guide references CRM-integration support (such as HubSpot, Affinity, Attio, and Zapier) as part of its Business plan. | 中 | SP006 |
| CP053 | No source reviewed in this chapter discloses a public user count, customer count, or meeting-volume figure for Granola comparable to the scale metrics Otter.ai, Fireflies, and tl;dv publish about themselves. | 低 | |
| CI001 | Granola’s Business plan is publicly listed at $14 per user per month and includes unlimited meeting history, CRM integrations, and folder-level queries across a team’s conversations. | 高 | SI002, SI016 |
| CI002 | SiliconANGLE’s March 2026 coverage describes Granola’s pricing as a free tier plus two paid tiers: a $14-per-month Business plan and a $35-per-month Enterprise plan with expanded cybersecurity controls and priority support, though the article does not state whether the Enterprise figure is per seat or a flat organizational rate. | 中 | SI016 |
| CI003 | Granola’s own Enterprise page lists enterprise-only features (SSO/SAML, SCIM provisioning, admin dashboard, flexible data retention, SOC 2 Type II) but does not publish a specific Enterprise contract price. | 中 | SI003 |
| CI004 | Granola’s Series C announcement states a personal API for accessing notes is bundled into the Business and Enterprise plans, and a separate enterprise API for admins is bundled into the Enterprise plan only, with neither API priced as a standalone line item. | 中 | SI004 |
| CI005 | No source reviewed in this chapter discloses what share of Granola’s revenue comes from the Business tier, the Enterprise tier, or API/add-on usage. | 低 | |
| CI006 | SiliconANGLE reported that Bloomberg found Granola’s revenue grew by 250% in the quarter preceding its March 2026 Series C, though neither outlet disclosed the absolute revenue base that growth rate was measured against. | 中 | SI016 |
| CI007 | Granola’s own sales buyer’s guide positions its product for founders and SMB sales teams under 15 reps, explicitly ceding large SDR organizations (30+ reps) to conversation-intelligence platforms like Gong, which frames Granola’s addressable per-account contract value as structurally smaller than an enterprise conversation-intelligence deal. | 中 | SI010 |
| CI008 | Granola’s own buyer’s guide cites third-party buyer estimates that Gong’s Foundations plan costs $1,300–$1,600 per user per year plus a mandatory $5,000–$50,000 annual platform fee, used as an adjacent-category price anchor rather than a Granola price. | 中 | SI010 |
| CI009 | An independent CloudTalk analysis corroborates a similar Gong cost structure: roughly $1,600 per user per year, a mandatory $5,000+ platform fee, and add-ons such as Gong Forecast (~$700/user/year) and Gong Engage (~$800/user/year). | 中 | SI028 |
| CI010 | Fireflies.ai lists Pro at $10, Business at $19, and Enterprise at $39 per seat per month (billed annually) on its public pricing page. | 中 | SI026 |
| CI011 | Otter.ai lists Pro at $16.99 per user per month ($8.49 promotional) and Business at $30 ($24 promotional) on its public pricing page. | 中 | SI027 |
| CI012 | Fathom lists Team at $19/$15 and Business at $34/$25 per user per month (2-user minimum), with Enterprise priced only through a sales conversation, matching Granola’s own pattern of gating top-tier pricing. | 中 | SI029 |
| CI013 | Brex’s Granola case study states implementation began as a trial on the Business plan with select teams and moved to Granola Enterprise within weeks once adoption spread organically, a concrete land-and-expand upgrade path from self-serve to a paid enterprise contract. | 中 | SI006 |
| CI014 | Vercel’s case study describes its entire go-to-market team adopting Granola, indicating expansion within a function rather than a single-user deployment. | 中 | SI009 |
| CI015 | Zapier’s published case study claims Granola saves the company 830+ hours per month, a time-savings ROI proxy rather than a verified revenue, margin, or contract-value figure. | 中 | SI008 |
| CI016 | Vanta’s published case study claims Granola saves 260+ hours annually per user, another company-published time-savings proxy with no disclosed seat count or contract value behind it. | 中 | SI007 |
| CI017 | Index Ventures’ investment post describes a bottom-up adoption pattern -- individual users becoming evangelists, with usage spreading grassroots-style from individuals to teams to entire enterprises -- as the growth mechanism it backed in the Series C. | 中 | SI022 |
| CI018 | Granola’s own careers page states the company transcribes millions of minutes of conversation every day and that adoption “spreads quickly from leaders to entire organizations,” a qualitative usage-scale and virality claim without an accompanying seat, customer, or revenue count. | 中 | SI005 |
| CI019 | No source reviewed in this chapter -- company-published or third-party -- discloses a customer acquisition cost (CAC), lifetime value (LTV), or CAC-payback figure for Granola. | 低 | |
| CI020 | A 2026 external SaaS benchmark report puts median CAC payback for product-led-growth/bottom-up SaaS companies at 12–15 months, with top-quartile companies near 9–12 months, and flags any payback above 24 months as a warning sign for venture-funded SaaS -- offered here strictly as comp-set context, not a measured Granola figure. | 低 | SI025 |
| CI021 | The same 2026 benchmark report states median sales-and-marketing spend as a share of revenue for venture-backed SaaS fell from more than 50% in 2021 to roughly 35–40% in 2024–2026, again offered only as external comp-set context. | 低 | SI025 |
| CI022 | Granola’s own security page states it uses third-party transcription providers such as Deepgram and AssemblyAI and third-party AI providers such as OpenAI and Anthropic to transcribe and summarize meetings, meaning Granola pays external vendors on a usage basis for its two core cost-of-revenue inputs. | 中 | SI011 |
| CI023 | Because transcription and LLM summarization are licensed from third-party vendors rather than built on owned infrastructure, Granola’s cost of revenue is likely a usage-linked variable cost that scales with meeting volume and note length rather than a mostly fixed hosting cost. | 中 | SI011 |
| CI024 | The 2026 SaaS benchmark report states median subscription gross margin across SaaS companies is roughly 76%, with top-quartile companies above 84%, and explicitly flags that companies below 70% gross margin usually carry unusual variable costs such as heavy compute or AI inference -- a pattern directly relevant to an AI-transcription product like Granola’s, though not a measured Granola figure. | 低 | SI025 |
| CI025 | Zoom Communications’ FY2026 Form 10-K (fiscal year ended January 31, 2026) reports gross margin increased to 77.0% from 75.8% a year earlier, driven by lower stock-based compensation and hosting efficiencies -- cited here only as a public-company collaboration-software margin benchmark, not a Granola fact. | 中 | SI014 |
| CI026 | Granola requires enterprise buyers to adopt SSO/SAML, SCIM provisioning, an admin dashboard, and encryption, and states it holds SOC 2 Type II certification, all of which represent ongoing security-engineering and compliance investment layered on top of its core transcription cost base. | 中 | SI003 |
| CI027 | No source reviewed in this chapter discloses Granola’s cost of revenue, gross margin, or an R&D/sales-and-marketing/G&A operating-expense breakdown; TheNextWeb’s Series C coverage explicitly notes that Granola has not disclosed revenue figures, user counts, or retention metrics publicly. | 中 | SI030 |
| CI028 | Companies House filing history shows Granola Labs Ltd filed “total exemption full accounts” for the periods ended 31 March 2024 and 31 March 2025, a UK small-company audit exemption that does not require public disclosure of turnover, cost of revenue, or profit. | 高 | SI013, SI012 |
| CI029 | Companies House records show Granola Labs Ltd (company number 14703652) was incorporated on 3 March 2023 under SIC code 62012 (business and domestic software development) and remains an active private limited company as of the latest filing reviewed. | 高 | SI012, SI013 |
| CI030 | A Companies House PSC filing records that co-founder Christopher Jonathan Pedregal ceased to be a “person with significant control” of Granola Labs Ltd on 14 April 2023, consistent with cap-table dilution following the company’s first institutional funding rather than any disclosed departure. | 中 | SI013 |
| CI031 | Granola’s March 2026 decision to encrypt its local notes database -- publicly criticized by an a16z partner for breaking third-party agent workflows -- represents an unquantified engineering and trust-remediation cost incurred to protect its data model rather than a feature customers directly pay more for. | 中 | SI033 |
| CI032 | A March 2026 report states that by October 2024, when Granola raised its $20 million Series A, it had approximately 5,000 weekly users, but the company has not disclosed a current user count as of mid-2026. | 中 | SI032 |
| CI033 | The same report notes Fireflies.ai has surpassed 16 million users and a $1 billion valuation, giving a concrete independent scale comparison against Granola’s undisclosed current user count. | 中 | SI032 |
| CI034 | Tracxn’s company profile estimates Granola had 116 employees as of late May 2026, the most specific independent headcount figure identified for this chapter. | 中 | SI024 |
| CI035 | Tracxn’s public profile masks Granola’s per-round revenue-multiple and several funding-amount fields behind a paid-data paywall, meaning even a dedicated data vendor does not surface Granola’s revenue or valuation multiple for free. | 中 | SI024 |
| CI036 | Granola’s published customer roster (Brex, Vanta, Zapier, Vercel, and -- per its Series C post and Kleiner Perkins’ writeup -- Asana, Gusto, Thumbtack, Cursor, Lovable, Decagon, and Mistral AI) demonstrates reference-customer quality but none of the reviewed case studies disclose seat counts, contract value, or renewal/expansion rates. | 中 | SI004, SI023 |
| CI037 | TheNextWeb’s Series C coverage explicitly states Granola has not disclosed retention metrics publicly, and no other source reviewed in this chapter discloses a net revenue retention, logo retention, or churn figure for the company. | 中 | SI030 |
| CI038 | Granola announced a $125 million Series C on 2026-03-25 led by Index Ventures, with participation from Kleiner Perkins and returning investors Lightspeed, Spark, and NFDG, at a $1.5 billion post-money valuation. | 高 | SI004, SI015 |
| CI040 | TechCrunch and The SaaS News both reported Granola’s total funding reached $192 million after the Series C, while BusinessCloud instead reported approximately $188 million (£141 million), a roughly $4 million cross-publisher discrepancy on the exact cumulative total. | 中 | SI015, SI017, SI018 |
| CI041 | StartupRise’s coverage states only that the fresh Series C funding “will be used to further scale its product and support continued growth,” a generic use-of-funds statement with no numeric allocation across headcount, infrastructure, or go-to-market spend. | 中 | SI021 |
| CI042 | Granola’s own Series C blog post frames the raise around shipping Spaces, personal and enterprise APIs, an updated MCP server, and an admin usage-visibility dashboard, implying the capital is being directed at product/engineering expansion rather than disclosed as a numeric budget split. | 中 | SI004 |
| CI043 | SiliconANGLE reported that Bloomberg found Granola plans to hire more people following the Series C, consistent with the open roles listed on Granola’s own careers page. | 中 | SI016, SI005 |
| CI044 | No source reviewed in this chapter discloses Granola’s cash on hand, monthly burn rate, or runway in months; as a private UK company filing exempt small-company accounts, Granola has no near-term legal obligation to publish any of these figures. | 低 | |
| CI045 | The Companies House filing history reviewed for Granola Labs Ltd lists confirmation statements, accounts, PSC, and officer filings but no registered charge (secured-debt) filings, consistent with an equity-only capital structure to date, though this chapter reviewed the filing-history summary rather than the full charges register. | 中 | SI013 |
| CI046 | Neither Granola nor its investors state a public next-round trigger; Index Ventures’ and Kleiner Perkins’ investment posts instead frame the Series C around a multi-year pivot from meeting notes to enterprise “context layer” infrastructure, implying the round is meant to fund a multi-year build rather than bridge to an imminent follow-on. | 低 | SI022, SI023 |
| CI047 | TheNextWeb reported that Granola’s valuation rose sixfold from $250 million to $1.5 billion in about ten months without the company disclosing revenue figures, user counts, or retention metrics, which TheNextWeb frames as inviting scrutiny. | 中 | SI030 |
| CI048 | A low-reputation opinion outlet, BritishWire, amplified social-media comparisons to the dot-com bubble and argued a market correction is likely once highly valued AI startups like Granola are “forced to show their actual revenue sheets,” reflecting public sentiment rather than a substantiated financial finding. | 低 | SI031 |
| CI049 | All disclosed capital raised by Granola to date -- a seed round plus Series A, B, and C -- has been venture equity; no source reviewed in this or prior chapters identifies a debt facility, venture-debt instrument, or project-finance obligation. | 中 | SI015, SI013 |
| CI050 | Granola’s revenue quality signal is directionally positive (250% quarterly growth per Bloomberg, land-and-expand contract upgrades at Brex) but unverifiable in absolute terms because no dollar-denominated revenue, ARR, or seat-count figure has been disclosed by the company, its investors, or any third-party data provider reviewed. | 中 | SI016, SI006, SI024 |
| CI051 | Granola’s margin path is structurally exposed to third-party LLM-inference and transcription vendor costs in a way comparable SaaS benchmarks associate with sub-70% gross margin, versus the 76%-84% median-to-top-quartile range for subscription SaaS and Zoom’s 77.0% FY2026 comparable, making margin trajectory a genuine open question rather than a presumed SaaS-level outcome. | 中 | SI011, SI025, SI014 |
| CI052 | Capital intensity looks low in the classic sense (no debt, no capex-heavy infrastructure disclosed) but capital *dependency* is high: Granola has raised four rounds in under three years, all equity, with no disclosed cash balance, burn, or runway to show how long the $125 million Series C proceeds are expected to last. | 中 | SI004, SI013 |
| CI053 | The primary diligence blocker is structural, not just company reticence: Granola Labs Ltd’s UK small-company audit exemption means no public filing will surface turnover, margin, or profit absent a future statutory trigger, so underwriting Granola’s financials requires negotiated data-room access rather than waiting on a future public filing. | 高 | SI013, SI012 |
| CE001 | Granola markets its core mechanism as bot-free, device-level capture: the app runs on the user's own computer, captures meeting audio locally, and never joins a call as a visible bot participant. | 高 | SE001, SE004 |
| CE002 | A user must manually start Granola before a meeting; the app does not auto-join, auto-record, or otherwise activate itself for a call. | 中 | SE004 |
| CE003 | Granola's current product lineup comprises the AI Notepad (macOS, Windows, iOS), an agentic Granola Chat, Briefs, Recipes, Team Spaces/Folders, a Personal API, an Enterprise Admin API, and Granola MCP. | 高 | SE005, SE007 |
| CE004 | Granola for Enterprise adds SSO/SAML, SCIM provisioning, an admin dashboard, domain management, flexible data-retention windows, and SOC 2 Type II documentation as an enterprise-only control layer on top of the base product. | 中 | SE003 |
| CE005 | Granola's own careers page lists an 'Android Engineer' among current team members, even though no Android app is listed on Granola's sitemap or on any public app store as of the run date. | 中 | SE006 |
| CE006 | Granola is distributed as a free tier plus paid Business and Enterprise plans that are differentiated primarily by feature access (meeting-history depth, CRM integrations, admin controls) rather than by a different underlying product architecture. | 中 | SE003, SE005 |
| CE007 | A personal API for accessing one's own notes is bundled into the Business and Enterprise plans, and a separate enterprise admin API for organization-wide context is bundled into the Enterprise plan only, with neither priced as a standalone product. | 中 | SE007 |
| CE008 | Granola's own /integrations marketing URL returned an HTTP 404 at the time of this review, leaving no single official, current page cataloguing the company's full integration list. | 中 | SE037 |
| CE009 | Zapier's Principal Designer adopted Granola for the whole design/product/engineering org after other AI transcription tools felt invasive; the published case study reports 830+ hours saved per month and 52% organization-wide adoption. | 中 | SE031 |
| CE010 | Vanta's VP of Customer Success adopted Granola for internal 1:1s, standups, and cross-functional syncs, citing SOC 2 Type II certification, encryption, and notice-and-consent recording controls as prerequisites for a security-and-compliance company; the case study reports 69% organization-wide adoption and a 61% increase in meeting visibility. | 中 | SE030 |
| CE011 | Brex's Head of IT rolled Granola out company-wide after employees adopted it organically, valuing SSO/SCIM delivery and a no-public-model-training commitment for a regulated fintech; the case study reports over 50% adoption across knowledge workers within six months and 427 hours saved per week. | 中 | SE029 |
| CE012 | Vercel's sales-engineering team replaced a slower transcription tool with Granola, using its 'chat with the transcript' capability to retrieve specific technical objections raised by prospects; the case study reports 100% GTM team adoption, 11 hours saved weekly per user, and a 25% GTM productivity increase. | 中 | SE032 |
| CE013 | Executive search firm Daversa Partners deployed Granola across 136 of its 150 employees specifically because device-level, bot-free capture felt less intrusive than visible-bot tools for confidential candidate and client conversations. | 中 | SE011 |
| CE014 | An independent Product Hunt community thread cites in-person, multilingual conversation capture -- for example, translating a Japanese conversation into English notes -- as a distinguishing real-world use case for Granola's mobile app. | 低 | SE027 |
| CE015 | None of Granola's four published customer case studies (Brex, Vanta, Zapier, Vercel) discloses a seat count, contract value, renewal rate, or churn figure; all quantified benefits are limited to adoption percentages and hours-saved estimates. | 中 | SE029, SE030, SE031, SE032 |
| CE016 | Granola's desktop client is built as an Electron application packaged for macOS and Windows, confirmed both by Granola's own engineering blog and independently by a security researcher who unpacked the macOS .asar archive to inspect its JavaScript. | 高 | SE019, SE025 |
| CE017 | Granola's transcription pipeline runs on third-party ASR vendors Deepgram and AssemblyAI, and its AI note-enhancement, chat, and summarization pipeline runs on third-party OpenAI and Anthropic LLM APIs, per the company's own security page and corroborated by an independent researcher's discovery of AssemblyAI references inside the client. | 高 | SE004, SE025 |
| CE018 | Independent security research firm Tenable found that an unauthenticated api.granola.ai endpoint ('get-feature-flags') returned a live AssemblyAI API key that could be used to enumerate and pull meeting transcript IDs and content directly from AssemblyAI's own API. | 高 | SE025, SE022 |
| CE019 | Granola's own postmortem states the AssemblyAI key exposure affected roughly 300-333 iOS TestFlight beta testers between November 2024 and March 2025, that the production macOS app was unaffected, and that no evidence of exploitation was found beyond the reporting researcher's own testing. | 高 | SE022, SE025 |
| CE020 | Granola's local desktop application cache stores meeting metadata, transcripts, and AI summaries in a nested 'double-JSON' local file (cache-v3.json); an independent open-source Python project reverse-engineered this structure without making any calls to Granola's own servers. | 中 | SE026 |
| CE021 | The same open-source project documents an undocumented alternative access path -- extracting session tokens from a local supabase.json configuration file to call Granola's cloud API directly -- which its maintainer explicitly chose not to implement, citing privacy, performance, and reliability reasons. | 中 | SE026 |
| CE022 | In March 2026, Granola encrypted its local on-device notes database; an a16z partner publicly criticized the change in a tweet viewed roughly 327,000 times for cutting off third-party agents and APIs that previously read the plaintext cache file directly. | 中 | SE034 |
| CE023 | Granola's co-founder responded to the local-database-encryption criticism by pointing to Granola MCP as the replacement access path; the critic countered that MCP requires MCP-specific client support rather than open file or API access, so the substitution did not resolve the underlying data-portability complaint. | 中 | SE034 |
| CE024 | Granola's own engineering blog describes a deliberate architecture tradeoff behind its April 2026 agentic rebuild of Granola Chat: reading full meeting transcripts is roughly 10x more token-expensive than reading condensed AI-generated notes, forcing an explicit depth-versus-cost/latency choice when the agent answers cross-meeting questions. | 高 | SE018, SE005 |
| CE025 | Granola MCP, launched February 4, 2026, implements the open Model Context Protocol standard so external AI tools such as Claude, ChatGPT, Cursor, and Claude Code can query a user's Granola notes as context; enterprise MCP access remains an early-access beta that is off by default for admins to enable. | 高 | SE017, SE005 |
| CE026 | Granola's transcripts, notes, and AI summaries are stored in an AWS-hosted, US-region Virtual Private Cloud, encrypted at rest and in transit and backed up daily, while raw meeting audio is never stored -- it is transcribed in real time and then discarded. | 高 | SE004, SE011 |
| CE027 | Granola authenticates users through Google Workspace or Microsoft OAuth login tied to the user's calendar rather than through a Granola-specific username/password system. | 中 | SE004 |
| CE028 | Granola's public release history shows a near-monthly cadence of shipped features from its July 2024 desktop launch through May 2026, spanning a $20M raise alongside the original notepad (Nov 2024), iOS launch (May 2025), Windows availability paired with SOC 2 Type II completion (Jul 2025), Recipes (Sep 2025), Granola MCP (Feb 2026), a $125M Series C (Mar 2026), an agentic Chat rebuild (Apr 2026), and Briefs (May 2026). | 高 | SE005, SE007 |
| CE029 | Granola's own changelog documents native or Zapier-mediated integrations added over time: Slack posting (Oct 2024), Notion sharing (Mar 2025), Attio and Zapier connecting to 8,000+ apps (Jul-Sep 2025), and HubSpot (Sep 2025). | 中 | SE005 |
| CE030 | A third-party uptime-monitoring aggregator recorded four Granola incidents in the 90 days before the run date, with a median resolution time of about 2 hours 34 minutes, including an iOS transcription outage (Jun 25, 2026), a Granola Chat outage from a 'broken deployment' (Mar 10, 2026), and a separate incident in which Claude users could not connect to Granola MCP (Mar 9, 2026). | 中 | SE028 |
| CE031 | No source reviewed discloses a Granola-published uptime SLA or a company-maintained public status page; the only incident and uptime visibility identified comes from an unaffiliated third-party monitoring aggregator. | 低 | |
| CE032 | Despite an Android-engineer role on Granola's team roster, no Android application appears on Granola's sitemap or on Google Play as of the run date, indicating Android support is an internal roadmap item rather than a shipped product. | 中 | SE006 |
| CE033 | Granola's own MCP announcement states the company is 'also working on additional enterprise capabilities: advanced access controls, usage monitoring and analytics, higher rate limits, and flexible pricing options,' each explicitly described as forthcoming rather than shipped. | 中 | SE017 |
| CE050 | Granola launched its native iOS app in May 2025, extending capture from desktop-only to in-person, on-the-go meetings and phone calls. | 中 | SE005 |
| CE051 | Granola launched Windows availability in July 2025, the same month it first announced SOC 2 Type II compliance. | 中 | SE005 |
| CE052 | Granola's iOS distribution depends entirely on Apple's App Store review and policy environment, which its own App Store listing identifies as the only public iOS distribution channel. | 中 | SE013 |
| CE034 | Granola states that its bot-free, zero-audio-retention architecture -- local capture, no server-side audio storage, immediate audio deletion after transcription -- is the direct reason it completed SOC 2 Type II certification in roughly three months versus a typical 12-18 months. | 高 | SE011, SE004 |
| CE035 | Granola frames its Chat product around a 'CODE' framework (Capture, Organize, Distill, Express), arguing that keyword search over raw transcripts alone does not produce reusable organizational memory the way structured, relationship-aware synthesis does. | 中 | SE009 |
| CE036 | Recipes, launched September 2025, differentiates Granola's prompt-library approach by sourcing shareable, reusable prompt templates from named external experts (e.g., Lenny Rachitsky, Matt Mochary) rather than relying only on generic, company-authored prompts. | 中 | SE010 |
| CE037 | An independent, competitor-published review (tl;dv) estimates Granola's transcription accuracy at roughly 94-95% versus tl;dv's own claimed 97%+, and states Granola supports about 15 languages versus tl;dv's 30+, framing Granola's core differentiation as workflow and capture design rather than raw transcription performance. | 中 | SE036 |
| CE038 | Granola's 'bot-free' and 'private by default' differentiation claims have each been publicly and independently contested: one independent reviewer rates Granola's enterprise readiness at 3 out of 10, citing weak default sharing settings and no EU or on-prem hosting option, and a separate outlet reports that notes are shareable via link to anyone who obtains it by default unless a user changes that setting. | 中 | SE033, SE035 |
| CE039 | Granola states it is SOC 2 Type II certified (first announced July 7, 2025) and GDPR compliant, with independent audit documentation available to enterprise procurement teams through a Trust Center. | 高 | SE005, SE003 |
| CE040 | Granola's Data Processing Addendum designates Granola as a data processor for customer-submitted personal data, sets a Personal Data Breach notification duty 'without undue delay,' and separately designates Granola an independent controller for its own business-contact and usage data. | 高 | SE014, SE015 |
| CE041 | Granola does not currently hold a HIPAA Business Associate Agreement, a gap the company discloses directly on its own blog rather than in marketing copy, even though its help center separately maintains a dedicated Healthcare & HIPAA FAQ section. | 高 | SE011, SE020 |
| CE042 | Granola's Enterprise plan turns off AI model training by default and contractually prohibits its third-party AI providers (OpenAI, Anthropic) from training on customer meeting data; free and standard consumer accounts are instead opted into anonymized model training by default, with a manual opt-out available in Settings. | 高 | SE004, SE011 |
| CE043 | Granola maintains a public Vulnerability Disclosure Policy and has published three detailed security postmortems since mid-2025: an AssemblyAI API key exposure, a Google Workspace auto-join account-matching gap, and a Google Workspace session-logout gap. | 高 | SE016, SE022, SE023, SE024 |
| CE044 | The Google Workspace auto-join postmortem discloses that, until a fix shipped May 23, 2025, workspace auto-join checked only whether a user's Google email domain matched the workspace domain, which could let a personal ('unmanaged') Google account sharing that domain automatically join an enterprise workspace; Granola's own database check found no workspace had actually been affected before the fix. | 中 | SE023 |
| CE045 | The Google Workspace session-logout postmortem discloses that 187 users who were removed, suspended, or deleted from their linked Google Workspace retained active Granola sessions and could keep accessing historical notes and transcripts until session-validity checks were deployed between May 28 and June 2, 2025. | 中 | SE024 |
| CE046 | Granola's security page and its own local-first architecture blog both state that meeting audio is never stored on any server -- only the resulting text transcript and notes persist -- which narrows the blast radius of any future data breach to text rather than stored audio or biometric voice data. | 高 | SE004, SE011 |
| CE047 | An independent open-source project confirms Granola's local cache can be read entirely offline without any Granola server interaction, technically corroborating the company's on-device-data claim, though the same ecosystem shows that path was partially foreclosed once Granola began encrypting the local database in March 2026. | 中 | SE026, SE034 |
| CE048 | No source reviewed publishes an independent, company-run, or third-party audited transcription-accuracy benchmark for Granola; the only accuracy figure identified (approximately 94-95%) comes from a single competitor-authored review and should be treated as directional rather than verified. | 低 | SE036 |
| CE049 | Granola's enterprise transparency controls include an optional on-screen recording watermark and an automatic in-meeting chat notification when transcription starts, both configurable by admins to match regional consent requirements. | 中 | SE003 |
| CU001 | Granola sells three plans relevant to buyer segmentation: a free Basic tier, a Business plan priced at $14 per user per month per Granola's own sales guide, and an Enterprise plan estimated by an independent reviewer at roughly $35 per user per month (Granola does not publish an Enterprise list price). | 高 | SU028, SU012 |
| CU002 | Granola's own Enterprise marketing page states the product is trusted by teams in tech, finance, media, and more. | 中 | SU007 |
| CU003 | G2 reviewers of Granola span Small-Business (<=50 employees), Mid-Market (51-1,000 employees), and Enterprise (>1,000 employees) company sizes, across verticals including Information Technology and Services, Financial Services, and clinical/healthcare-adjacent roles. | 中 | SU010 |
| CU004 | Independent reviewers and G2 reviewers repeatedly identify recruiters, lawyers, consultants, VCs, sales engineers, and IT/customer-success leaders as Granola's primary user personas. | 中 | SU010, SU012, SU009 |
| CU005 | Granola's four detailed company-published case studies name distinct buyer/user functions: design/product/engineering (Zapier), customer-success leadership (Vanta), IT (Brex), and sales engineering/GTM (Vercel). | 中 | SU001, SU002, SU003, SU004 |
| CU006 | Executive search firm Daversa Partners' president adopted Granola across 136 of its 150 employees specifically because bot-free, device-level capture felt less intrusive than visible-bot tools for confidential candidate and client conversations. | 中 | SU005 |
| CU007 | Press and investor coverage names Asana, Gusto, Thumbtack, Cursor, Lovable, Decagon, Mistral AI, Intercom, Linear, and Replit as additional Granola customers, spanning work management, HR/payroll, marketplaces, developer tools, and AI-native startups. | 高 | SU021, SU024, SU020, SU006 |
| CU008 | No source reviewed discloses Granola's per-customer revenue, seat count, contract value, or an official geographic breakdown of its customer base beyond its UK/US home markets. | 中 | SU001, SU002, SU003, SU004, SU020 |
| CU009 | Granola reported 5,000 weekly users as of its October 2024 Series A. | 中 | SU020 |
| CU010 | Granola has not disclosed an updated total user or customer count since the 5,000-weekly-user figure from its October 2024 Series A; independent analysis states this explicitly. | 中 | SU020 |
| CU011 | Granola's own careers page states the company transcribes millions of minutes of conversation every day, without providing an exact figure or historical trend. | 低 | SU006 |
| CU012 | Independent press reporting states Granola reported 10% weekly user growth since launch as of its March 2026 Series C. | 中 | SU024 |
| CU013 | YipitData's Signals B2B spend panel, covering real transaction data across 900+ mid-market companies, shows Granola's observed mid-market customer count grew 25% month-over-month in February 2026. | 中 | SU014 |
| CU014 | The same YipitData panel shows Granola's gross new-customer additions increased 9x from January to February 2026, and net-new customers in February 2026 reached 4x the prior six-month average. | 中 | SU014 |
| CU015 | YipitData's panel shows Granola's observed mid-market spend reached approximately $6,000 in February 2026, a 3x increase over the prior six months. | 中 | SU014 |
| CU016 | YipitData notes that three competitors -- Fireflies.ai (~40% more customers), Otter.ai, and Fathom (~20% more each) -- still maintain larger total mid-market customer bases than Granola, though their growth is comparatively slowing. | 中 | SU014 |
| CU017 | Granola's headcount grew 49% over the six months ending April 2026, according to YipitData's tracking. | 中 | SU014 |
| CU018 | Zapier's case study reports 830+ hours saved per month and 52% organization-wide adoption after a Principal Designer's individual trial spread virally through the design/product/engineering organization. | 中 | SU001 |
| CU019 | Vanta's case study reports 260+ hours saved annually per user, 69% organization-wide adoption, and a 61% increase in meeting visibility, driven initially by SOC 2/compliance requirements. | 中 | SU002 |
| CU020 | Brex's case study reports 427 hours saved per week and 25,000+ cumulative hours saved, with over 50% adoption across knowledge workers within six months, initiated through organic employee adoption before IT formalized an enterprise rollout. | 中 | SU003 |
| CU021 | Vercel's case study reports 100% adoption across its GTM team, 11 hours saved weekly per user, and a 25% GTM productivity increase, centered on a 'chat with the transcript' sales-enablement use case. | 中 | SU004 |
| CU022 | None of Granola's four published customer case studies discloses a seat count, contract value, renewal rate, or churn figure. | 中 | SU001, SU002, SU003, SU004 |
| CU023 | Granola's G2 profile hosts 19 reviews with an aggregate 4.9-out-of-5 rating as of a late-2025 snapshot, spanning Small-Business, Mid-Market, and Enterprise company sizes. | 中 | SU010 |
| CU024 | Several G2 reviewers' identities were validated through business email, Google One Tap, or LinkedIn and are labeled organic (unincentivized) reviews, while at least two others disclose receiving a nominal gift card for completing the review. | 中 | SU010 |
| CU025 | G2 reviewers identify recurring product limitations: no audio/video playback, inconsistent speaker identification in larger meetings, no dedicated web app, and slow customer support on setup issues. | 中 | SU010 |
| CU026 | Aggregated Product Hunt reviewer commentary describes Granola as reliable and structured but flags the absence of an Android app, iPad/iOS capture limits, account setup friction, and requests for more integrations. | 低 | SU009 |
| CU027 | Three independent expert reviews score Granola materially lower and more variably than its G2 rating: thebusinessdive.com scored it 3.6/5 citing limited collaboration and analytics features, nubiapage.com scored it 4.5/5 citing the missing Android app and unclear free-tier limits, and bestreviews.net highlighted missing language support and iOS/speaker-identification gaps without assigning a numeric score. | 中 | SU011, SU012, SU013 |
| CU028 | An independent Product Hunt community thread cites in-person, multilingual conversation capture -- for example, translating a Japanese conversation into English notes -- as a distinguishing real-world use case for Granola's mobile app. | 低 | SU008 |
| CU029 | Press and investor coverage names Asana, Gusto, Thumbtack, Cursor, Lovable, Decagon, Mistral AI, Intercom, Linear, and Replit as additional Granola customers, without independent verification of usage depth, contract status, or outcomes for any of them. | 中 | SU021, SU024, SU020, SU006 |
| CU030 | No source reviewed discloses Granola's Net Revenue Retention (NRR), Gross Revenue Retention (GRR), logo-churn rate, or average contract renewal rate. | 中 | SU014, SU020 |
| CU031 | YipitData's independent B2B spend-panel analysis found near-zero churn among Granola's observed mid-market customers between roughly August 2025 and February 2026. | 中 | SU014 |
| CU032 | YipitData's panel found an asymmetric switching pattern: customers who adopt Granola frequently drop competing tools (Fathom, Fireflies.ai, Otter.ai), while the reverse -- dropping Granola for a competitor -- is rare in the observed data. | 中 | SU014 |
| CU033 | Granola's G2 profile shows an aggregate 4.9-out-of-5 rating from 19 reviews, while three independently conducted expert reviews scored the product 3.6/5 and 4.5/5 (with a third giving a caveat-heavy unscored assessment), indicating a wide satisfaction-measurement spread depending on review methodology. | 中 | SU010, SU011, SU012, SU013 |
| CU034 | A March 2026 change encrypting Granola's local notes database triggered public criticism from a prominent a16z partner in a tweet viewed roughly 327,000 times, liked 630 times, bookmarked 577 times, and drawing 177 replies, in which the poster asked the community for alternative-product recommendations. | 中 | SU015 |
| CU035 | Independent commentary describes the local-database-encryption backlash as surfacing 'a problem that had been quietly frustrating Granola users for months,' tied to the loss of local file/API access for personal automation and AI-agent workflows. | 中 | SU015 |
| CU036 | TechCrunch reporting confirms the encryption backlash included multiple users beyond the a16z partner, and states Granola's co-founder responded by promising new personal and enterprise APIs to restore programmatic data access. | 中 | SU018 |
| CU037 | Case-study adoption percentages (52%, 69%, over 50%, 100%) measure organization-wide usage depth at a point in time, not renewal, repeat-purchase, or multi-period retention. | 中 | SU001, SU002, SU003, SU004 |
| CU038 | Granola's own Enterprise marketing page frames its go-to-market motion explicitly as land-and-expand: it starts with one person who loves Granola, and it spreads because it makes every individual better at their job. | 中 | SU007 |
| CU039 | All four detailed case studies (Zapier, Vanta, Brex, Vercel) describe an individual or small-team pilot that leadership later expanded into a company-wide Enterprise-plan rollout. | 中 | SU001, SU002, SU003, SU004 |
| CU040 | Granola's expansion products beyond per-seat growth include Spaces/Team Folders for cross-functional knowledge sharing, a Personal API, an Enterprise Admin API, and Granola MCP for third-party AI-agent integration. | 中 | SU019, SU025, SU027 |
| CU041 | Nearly all of Granola's quantified customer-outcome proof rests on four named accounts, each a well-funded technology company with investor-network ties to Granola's own backers, creating an evidentiary concentration risk distinct from revenue concentration. | 中 | SU001, SU002, SU003, SU004 |
| CU042 | Independent reporting states Granola's growth has happened largely through word of mouth, with no disclosed reseller, channel-partner, or systems-integrator distribution documented in any source reviewed. | 中 | SU024 |
| CU043 | Independent analysis warns that Microsoft Copilot and Google Gemini are building the same conversational-context-aggregation capability natively into Teams and Workspace environments where most enterprise employees already have seats, a platform-level concentration risk to Granola's expansion path. | 中 | SU020 |
| CU044 | Granola's own enterprise page states it is trusted by teams in tech, finance, media, and more, but every independently detailed named account in this chapter is a technology or technology-adjacent company; no finance- or media-sector named account with case-study-level detail exists in any source reviewed. | 中 | SU007, SU001, SU002, SU003, SU004, SU006 |
| CR001 | Granola's Privacy Policy states personal data may be collected, used, and disclosed to governments, including law enforcement, or other parties to meet certain legal requirements. | 中 | SR002 |
| CR002 | Granola's Data Processing Addendum commits the company to inform the customer without undue delay in the event of a Personal Data Breach. | 中 | SR001 |
| CR003 | Granola's Terms of Service directly link to three published security post-mortems as part of a public Vulnerability Disclosure Policy. | 中 | SR003 |
| CR004 | The UK ICO's monitoring-workers guidance requires employers that record or monitor staff (including via AI transcription tools) to establish a documented lawful basis, apply necessity and proportionality, and maintain transparency toward the people being monitored. | 中 | SR004 |
| CR005 | The ICO's guidance on AI and data protection requires organizations deploying AI systems to address fairness, accountability, lawfulness, transparency, and statistical accuracy under the UK GDPR. | 中 | SR005 |
| CR006 | The EU AI Act's rules for general-purpose AI models -- the category that includes the OpenAI and Anthropic models Granola uses for summarization -- became applicable in August 2025, with the Act's broader AI-system transparency obligations following in August 2026. | 中 | SR006 |
| CR007 | California Penal Code Section 632 makes it a crime, punishable by a fine of up to $2,500 per violation (or up to $10,000 for repeat violations) plus possible imprisonment, to record a confidential communication without the consent of all parties to it. | 中 | SR007 |
| CR008 | The Illinois Biometric Information Privacy Act requires any private entity that collects biometric identifiers or information, including voiceprints, to provide written notice of collection and purpose and obtain a written release beforehand, and bars selling or profiting from that data. | 中 | SR008 |
| CR009 | HHS's HIPAA Privacy Rule requires a covered entity to impose specified written safeguards in a business associate agreement whenever it uses a contractor, including an AI vendor, that will use or disclose protected health information on its behalf. | 中 | SR009 |
| CR010 | A consolidated federal class action, In re Otter.ai Privacy Litigation (N.D. Cal.), alleges Otter.ai's AI notetaker records and analyzes meeting audio -- including creating voiceprints -- without the consent of non-account-holder participants, asserting claims under the federal Electronic Communications Privacy Act, California's Invasion of Privacy Act, and Illinois's Biometric Information Privacy Act. | 高 | SR010, SR013 |
| CR011 | Bloomberg Law reports that courts are split on whether an AI transcription vendor should be treated as a tool used by meeting participants or as an unauthorized third-party eavesdropper, a distinction that determines wiretap liability and can expose the customer company to vicarious liability for a vendor's statutory violations. | 中 | SR011 |
| CR012 | Goodwin Procter's April 2026 client alert states that California Invasion of Privacy Act violations can expose organizations to damages of up to $5,000 per violation plus class-action, common-law, and in egregious cases criminal exposure, and separately flags voiceprint/BIPA liability and attorney-client-privilege risk as distinct exposure categories for AI transcription tools. | 中 | SR012 |
| CR013 | All-party ("two-party") consent states -- including California, Florida, Illinois, Maryland, Massachusetts, Montana, New Hampshire, Pennsylvania, and Washington -- require every meeting participant's consent before an AI tool may record or transcribe a conversation, while most other US states require only one party's consent. | 中 | SR011 |
| CR014 | Granola markets itself as a "bot-free" notetaker whose AI notepad runs on a single participant's device and captures microphone plus system audio locally rather than joining the call as a visible third-party bot, a design choice that differs mechanically from Otter.ai's bot-based capture but does not, per independent legal commentary on vendor classification, resolve the underlying question of whether every other participant in an all-party-consent state has validly consented. | 中 | SR016, SR011 |
| CR015 | Granola's Enterprise plan markets SSO/SAML, SCIM provisioning, and SOC 2 Type II certification as "independently audited," but none of those controls is itself a certification of compliance with wiretap consent statutes, the UK GDPR, the EU AI Act, or HIPAA. | 中 | SR023, SR005, SR006 |
| CR016 | Granola's own blog post on local-first architecture states directly that the company does not currently hold a HIPAA Business Associate Agreement. | 中 | SR017 |
| CR017 | No source reviewed for this chapter identifies a lawsuit, regulatory investigation, or enforcement action naming Granola directly, indicating the company's litigation and enforcement exposure is currently prospective and category-wide rather than realized. | 低 | SR010, SR013 |
| CR018 | The wiretap and biometric-privacy litigation risk documented against Otter.ai is a category-wide comparable risk rather than a company-specific one, since Granola, Otter.ai, Fireflies.ai, Fathom, tl;dv, and Read AI all process the same underlying class of multi-party meeting audio. | 中 | SR010, SR011, SR012, SR013 |
| CR019 | Granola disclosed three public security post-mortems in 2025: an AssemblyAI API key exposure affecting iOS TestFlight beta testers, a legacy unmanaged Google-account workspace auto-join gap, and a Google Workspace session-logout vulnerability affecting 187 users. | 中 | SR018, SR019, SR020 |
| CR020 | Independent security research firm Tenable identified the AssemblyAI API key exposure via an unauthenticated "get-feature-flags" endpoint, corroborating Granola's own post-mortem account from an independent technical source. | 高 | SR021, SR018 |
| CR021 | Granola's own Security, Privacy & Data FAQs list "Healthcare & HIPAA" and "Education & FERPA" as separate compliance topic headings, indicating the company itself frames HIPAA and FERPA as open compliance questions rather than settled certifications. | 中 | SR024 |
| CR022 | Independent uptime tracker IsDown recorded four Granola service incidents in the 90 days ending on the run date, with a median incident duration of roughly two hours thirty-four minutes. | 低 | SR022 |
| CR023 | All three of Granola's disclosed 2025 security incidents were reported to the company by external security researchers rather than detected internally, based on the company's own post-mortem accounts. | 中 | SR018, SR019, SR020 |
| CR024 | Granola's Enterprise plan advertises SOC 2 Type II certification as "independently audited," and the company's own blog states this certification was achieved in roughly three months -- a timeline that is fast relative to the multi-quarter observation periods typical of SOC 2 Type II audits. | 中 | SR023, SR017 |
| CR025 | No source reviewed in this chapter identifies an independent penetration-test report or bug-bounty program at Granola beyond the vulnerability-disclosure policy referenced in its Terms of Service. | 低 | SR003 |
| CR026 | The one publicly documented piece of independent vulnerability research at Granola (Tenable) targeted an authentication gap on a company API endpoint, indicating the company's server-side attack surface -- not just its client applications -- has already been probed and found exploitable at least once. | 中 | SR021 |
| CR027 | None of Granola's three disclosed 2025 incidents involved unauthorized access to actual meeting audio or transcript content, per the company's own post-mortem descriptions, distinguishing an access-control/key-exposure incident pattern from a content-confidentiality breach pattern. | 中 | SR018, SR019, SR020 |
| CR028 | Independent reviewer Startup Fortune states Granola exposes user notes to anyone holding a share link by default and trains internal models on customer transcripts unless a user manually opts out, a data-handling posture that is a business risk if opt-out uptake is low, though this is a low-reputation-tier outlet rather than a technical audit. | 低 | SR037 |
| CR029 | Granola's own Security page states the company depends on third-party ASR vendors Deepgram and AssemblyAI for transcription and third-party AI vendors OpenAI and Anthropic for meeting summarization, concentrating four external vendor dependencies at the core of the product. | 中 | SR016 |
| CR030 | AssemblyAI, one of Granola's two named transcription vendors, was the direct source of the company's first disclosed 2025 security incident (an exposed API key), showing that vendor-concentration risk had already materialized once. | 中 | SR016, SR018 |
| CR031 | Microsoft's own support documentation confirms Teams' AI-generated meeting recap requires either a Teams Premium add-on license (US$10.00 per user per month) or a Microsoft 365 Copilot license, undercutting Granola's $14 Business-plan price point by bundling a comparable feature into an incumbent collaboration suite at a lower incremental cost. | 中 | SR038, SR039 |
| CR032 | Zoom's own announcement states AI Companion monthly active users more than tripled year over year in fiscal Q4 2026, and a third-party market report states Zoom AI Companion reached 510,000 accounts within five months of launch, showing a large, fast-growing bundled substitute distributed at no separate charge inside an existing collaboration subscription. | 中 | SR040, SR042 |
| CR033 | Google's Workspace page markets Gemini-powered "Take Notes for Me" directly inside Google Meet, a third major collaboration-suite incumbent bundling AI meeting-notes functionality at no separate charge. | 中 | SR041 |
| CR034 | A 2026 market-state report published by a competing AI meeting-assistant vendor states that platform-bundled players lead the AI note-taking category by raw user count, implying standalone vendors like Granola compete against free or bundled incumbents for the same workplace-meeting use case. | 低 | SR042 |
| CR035 | TechCrunch's March 2026 hardware roundup lists multiple wearable and physical AI notetaking devices as alternative capture form factors to app-based tools such as Granola, introducing a distribution-channel substitute risk beyond software competitors. | 中 | SR043 |
| CR036 | Both of Granola's Series C investors, Index Ventures and Kleiner Perkins, publish investment posts naming existing Granola customers (including Asana, Vanta, Gusto, Thumbtack, Cursor, and Lovable) as proof points, showing overlap between Granola's capital providers and its customer reference base. | 中 | SR032, SR033 |
| CR037 | Granola's UK parent, Granola Labs Ltd, has filed only equity-related Companies House records (confirmation statements, accounts, PSC filings) with no registered charge (secured-debt) filing identified, consistent with an all-equity capital structure rather than a bank or venture-debt lender as an additional financing-partner dependency. | 中 | SR014 |
| CR038 | No source reviewed in this chapter identifies a reseller, systems-integrator, or channel-partner distribution agreement for Granola, indicating its go-to-market currently depends on direct, self-serve, and organic-referral channels rather than diversified partner distribution. | 低 | SR025 |
| CR060 | Granola's own Security page states meeting notes are stored in a US-hosted AWS Virtual Private Cloud, encrypted at rest and in transit with daily backups, and that the company does not store raw meeting audio -- only transcripts and user notes. | 中 | SR016 |
| CR039 | Granola announced a $125 million Series C on 2026-03-25 at a $1.5 billion post-money valuation, led by Index Ventures with participation from Kleiner Perkins and returning investors. | 高 | SR025, SR026 |
| CR040 | TechCrunch and Tracxn both report Granola's total disclosed funding reached $192 million after the Series C. | 高 | SR026, SR029 |
| CR041 | No source reviewed discloses Granola's cash balance, monthly burn rate, or runway; as a UK private company that filed total-exemption accounts, Granola has no near-term statutory obligation to disclose any of these figures. | 中 | SR014 |
| CR042 | Granola's UK parent, Granola Labs Ltd, filed "total exemption full accounts" at Companies House for the year to 31 March 2025, meaning no profit-and-loss or turnover figures are publicly available through ordinary company filings absent a future statutory trigger. | 中 | SR014 |
| CR043 | Third-party analyst platform Tracxn states Granola employed 116 people as of 26 May 2026, the only independent total-headcount figure identified in any source reviewed for this chapter. | 中 | SR029 |
| CR044 | YipitData's B2B spend-panel tracking shows Granola's headcount grew 49% over the six months ending April 2026, a growth rate that -- absent disclosed revenue -- cannot be benchmarked against revenue-per-employee or any other efficiency ratio. | 中 | SR034 |
| CR045 | SaaSCalcHub's 2026 SaaS benchmark data places median subscription gross margin at approximately 76% (top-quartile 84%+) and states that companies below 70% gross margin typically carry unusual variable costs such as heavy AI-inference compute -- a description that plausibly applies to Granola given its reliance on four paid third-party ASR/LLM vendors per meeting. | 中 | SR030, SR016 |
| CR046 | Zoom's FY2026 Form 10-K, filed with the SEC, reports a 77.0% gross margin, illustrating the level of disclosure and margin performance a mature public comparable in the same collaboration-software category provides -- a transparency benchmark Granola does not yet meet as a private company. | 中 | SR031 |
| CR047 | Neither Granola nor its Series C investors state a specific revenue, ARR, or usage trigger for a future financing round; investor posts instead frame the raise as funding a multi-year "context layer" product build-out rather than a bridge to an imminent follow-on. | 高 | SR025, SR032, SR033 |
| CR048 | BusinessCloud reports Granola was founded in 2023 by Chris Pedregal and Sam Stephenson, and no source reviewed for this chapter names a third co-founder or discloses an independent board roster. | 中 | SR044 |
| CR049 | TechCrunch's October 2024 coverage of Granola's Series A states co-founder Chris Pedregal previously built Socratic (acquired by Google in 2018) and Stack inside Google's Area 120 before founding Granola, giving the leadership team a track record that includes at least one prior successful exit. | 中 | SR045 |
| CR050 | Granola's own careers page advertises active hiring across engineering, design, and go-to-market roles following the Series C, without disclosing a current total headcount figure on the page itself. | 中 | SR046 |
| CR051 | The only independent total-headcount figure available (116 employees, Tracxn, May 2026) combined with YipitData's reported 49% six-month headcount growth implies Granola is scaling organizational complexity quickly without any independently confirmed VP-level organizational chart or named non-founder executive in the sources reviewed. | 中 | SR029, SR034 |
| CR052 | No source reviewed in this chapter identifies a Chief Financial Officer, General Counsel, Head of Compliance, or Data Protection Officer at Granola, despite the regulatory and litigation-adjacent exposure documented elsewhere in this chapter. | 低 | SR044, SR045 |
| CR053 | Granola's SOC 2 Type II certification, default-off model training on Enterprise, and DPA breach-notification commitment together represent the company's primary disclosed compliance mitigations against the legal, regulatory, and operational risks catalogued in this chapter. | 中 | SR023, SR001, SR017 |
| CR054 | None of Granola's disclosed mitigations -- SOC 2 certification, its DPA, or its model-training opt-out -- directly addresses all-party wiretap consent, EU AI Act transparency obligations, or a HIPAA Business Associate Agreement, leaving those three regulatory categories without a company-disclosed mitigation as of the run date. | 中 | SR023, SR006, SR017 |
| CR055 | A monitorable kill-criterion for Granola's legal/regulatory risk is the filing of any wiretap, biometric-privacy, or GDPR/UK-GDPR enforcement action naming Granola directly, given that comparable litigation has already been filed against its closest bot-based competitor. | 中 | SR010, SR013 |
| CR056 | A monitorable kill-criterion for Granola's financial/model risk is a down round, a materially delayed Series D, or a disclosed net burn exceeding Series C proceeds ahead of a revenue-disclosure event -- none of which any source reviewed currently indicates, though two adverse commentators already flag the valuation as unverified. | 中 | SR027, SR028 |
| CR057 | A monitorable kill-criterion for Granola's partner/dependency risk is a documented price increase, outage, or policy change at OpenAI, Anthropic, Deepgram, or AssemblyAI that materially degrades the product or margin, or a demonstrated case of a named reference account substituting a bundled Teams/Zoom/Meet AI-notes feature for Granola. | 中 | SR016, SR038, SR040, SR041 |
| CR058 | Weighing likelihood, impact, and mitigation maturity across all evidence gathered, this chapter ranks Granola's top risk categories as: (1) partner/vendor concentration and platform-bundling substitution, (2) financial/capital-adequacy disclosure opacity, (3) legal/regulatory consent and privacy exposure, (4) recurring externally-discovered operational/security incidents, and (5) execution/governance disclosure gaps. | 中 | SR010, SR011, SR014, SR016, SR038, SR040, SR018 |
| CR059 | Every top-five risk category in this chapter's ranking has an identified monitorable diligence path or kill-criterion, but none has a company-disclosed quantitative threshold (such as a stated burn-rate ceiling or litigation-cost reserve), meaning monitoring currently depends on external indicators rather than company-reported metrics. | 中 | SR014, SR017 |
| CV001 | Granola raised approximately $4.25 million in seed funding before its Series A, per TechCrunch's October 2024 reporting. | 中 | SV005 |
| CV002 | Granola announced a $20 million Series A led by Spark Capital on 2024-10-23. | 高 | SV004, SV005 |
| CV003 | Granola raised $43 million in a Series B round led by NFDG in May 2025 at a $250 million valuation. | 高 | SV006, SV007 |
| CV004 | Granola announced a $125 million Series C led by Index Ventures on 2026-03-25 at a $1.5 billion post-money valuation, with participation from Kleiner Perkins and returning investors Lightspeed, Spark, and NFDG. | 高 | SV001, SV002, SV008 |
| CV006 | TechCrunch and The SaaS News both reported Granola's total disclosed funding reached $192 million after the Series C, while BusinessCloud instead reported approximately $188 million (£141 million), roughly $4 million below the $192 million figure. | 中 | SV002, SV009, SV003 |
| CV007 | Forbes reported on 2026-01-30 that Granola was in talks to raise a new round of at least $100 million led by Index Ventures at a valuation of over $1 billion, roughly two months before the round closed at $125 million and a $1.5 billion valuation. | 中 | SV031 |
| CV008 | TheNextWeb characterized Granola's valuation rising sixfold from $250 million to $1.5 billion in about ten months, absent any disclosed revenue, user, or retention figures, as inviting scrutiny. | 中 | SV015 |
| CV009 | BritishWire amplified social-media comparisons between Granola's valuation trajectory and the dot-com bubble, arguing a market correction is likely once highly valued AI startups are forced to disclose actual revenue. | 低 | SV016 |
| CV010 | Companies House filing history shows Granola Labs Ltd filed 'total exemption full accounts' for the periods ended 31 March 2024 and 31 March 2025, a UK small-company audit exemption that does not require public disclosure of turnover, cost of revenue, or profit. | 高 | SV010, SV011 |
| CV011 | No source reviewed in this chapter discloses Granola's revenue, ARR, gross margin, cash balance, or burn rate, meaning no valuation multiple can be independently computed against a Granola-specific financial base. | 中 | SV002, SV015 |
| CV012 | All capital Granola has raised to date -- seed plus Series A, B, and C -- is venture equity; no source reviewed identifies a debt facility, venture-debt instrument, or convertible note outstanding. | 中 | SV004, SV006, SV001 |
| CV013 | A Companies House PSC filing records that co-founder Christopher Jonathan Pedregal ceased to be a 'person with significant control' of Granola Labs Ltd on 14 April 2023, consistent with cap-table dilution following the company's first institutional funding rather than any disclosed departure. | 中 | SV011 |
| CV014 | Neither Granola nor its investors state a public next-round trigger or target IPO timeline; Index Ventures' investment post instead frames the Series C around a multi-year pivot from meeting notes to an enterprise 'context layer'. | 中 | SV008 |
| CV015 | Tracxn's public profile confirms Granola's total funding at $192 million over four rounds but masks per-round revenue-multiple fields behind a paid-data paywall, meaning even a dedicated data vendor does not surface a free valuation multiple for Granola. | 中 | SV013 |
| CV016 | Granola's investment thesis rests on Bloomberg-reported 250% quarterly revenue growth ahead of its Series C, a land-and-expand enterprise motion evidenced at Brex, and a reference-customer roster spanning Vanta, Zapier, Vercel, Asana, Gusto, Thumbtack, Cursor, Lovable, Decagon, and Mistral AI. | 中 | SV002, SV008 |
| CV017 | Index Ventures frames Granola's growth mechanism as bottom-up adoption -- individual users becoming evangelists who spread usage from individuals to teams to entire enterprises -- as the basis for its Series C investment. | 中 | SV008 |
| CV018 | Granola's anti-thesis centers on total disclosure opacity: no revenue, ARR, user count, retention, gross margin, or cash-runway figure has been disclosed by the company, its investors, or any third-party data provider reviewed across this report. | 中 | SV002, SV013 |
| CV019 | A second anti-thesis strand is product commoditization: direct peers Fathom and tl;dv now also offer bot-free capture, and bundled incumbents Microsoft Teams, Google Meet, and Zoom AI Companion distribute meeting-notes functionality inside subscriptions buyers already own, narrowing Granola's differentiation window. | 中 | SV017 |
| CV020 | A third anti-thesis strand is legal/regulatory exposure: Otter.ai faces a consolidated federal class action over recording and voiceprint capture without consent, and legal commentary treats bot-free, less-visible capture tools like Granola as squarely within the same wiretap and biometric-privacy risk category. | 中 | SV018, SV021 |
| CV021 | A bull scenario for Granola requires the Bloomberg-reported 250% quarterly growth rate to persist or decelerate only gradually, enterprise pilots at Brex-style accounts to convert into durable multi-year contracts, and the new personal/enterprise APIs and admin dashboard shipped with the Series C to drive expansion revenue. | 低 | |
| CV022 | A base scenario assumes Granola's growth decelerates toward the 40-60% year-over-year band SaaS benchmarks associate with typical post-hypergrowth SaaS companies, with the next round priced flat to modestly up from $1.5 billion given Carta data showing Series C primary valuations up 12.5% year-over-year in Q1 2026. | 中 | SV014, SV026 |
| CV023 | A bear scenario assumes growth decelerates sharply and Granola's next financing prices flat or down, a tail risk grounded in Carta's Q1 2026 data showing an 11.4% down-round rate still active across the market even as headline AI valuations rise. | 中 | SV026 |
| CV024 | Carta's Q1 2026 data shows more than 60% of all venture capital invested went to AI companies (83% within SaaS specifically), and that an AI foundational-model startup at Series A raises at a $300 million median pre-money valuation versus $55 million for a non-AI startup at the same stage, evidencing a broad AI valuation premium Granola's Series C sits within. | 中 | SV026 |
| CV025 | Carta's Q1 2026 data shows Series B and Series C primary pre-money valuations up 17.2% and 12.5% year-over-year respectively, indicating late-stage private pricing was still expanding industry-wide at the time of Granola's Series C. | 中 | SV026 |
| CV026 | Finro's Q1 2026 AI valuation update reports that public multiples in the 'Productivity Tools' category compressed even as private-round valuations in the same category continued to rise, a pattern of private/public divergence directly relevant to a still-private company like Granola. | 中 | SV029 |
| CV027 | Finro's Q1 2026 update notes that M&A multiples stayed broadly stable quarter over quarter across most AI segments, describing acquirers as the 'most pragmatic' market because deals must clear integration cost and ROI rather than narrative. | 中 | SV029 |
| CV028 | No source reviewed in this or prior chapters discloses a specific next-round trigger, runway estimate, or burn rate for Granola, so the timing and pricing of any Series D cannot be forecast from public evidence. | 低 | |
| CV029 | Fireflies.ai announced on 2025-06-12 that it reached a valuation of over $1 billion through a tender offer providing liquidity to early employees, while stating it has been profitable since 2023 and has not raised primary capital since a 2021 Series A. | 中 | SV025 |
| CV030 | Fireflies.ai states it serves 20 million people across more than 500,000 organizations and processes over 2 billion meeting minutes, with usage at 75% of Fortune 500 companies. | 中 | SV025 |
| CV031 | Forbes reported that Otter.ai, the top-downloaded note-taking app in the United States per Data.ai, announced in March 2025 that it had surpassed $100 million in annual recurring revenue. | 中 | SV031 |
| CV032 | Otter.ai's own April 2026 relaunch announcement claims more than 35 million users and over one billion meetings processed, without disclosing a current valuation or revenue figure alongside those user metrics. | 中 | SV017 |
| CV033 | No source reviewed in this run establishes a corroborated, current valuation figure for Otter.ai; scale claims (users, ARR) are disclosed but a post-2021 valuation mark is not, making a like-for-like valuation comparison to Granola impossible for this comparable. | 低 | |
| CV034 | Forbes reported that productivity software Notion was last valued at $11 billion in an employee share sale completed in December 2025, positioning it as a much larger adjacent AI-productivity comparable than Granola. | 中 | SV031 |
| CV035 | Zoom Communications (NASDAQ: ZM) had a market capitalization of $25.31 billion as of June 30, 2026, against trailing-twelve-month revenue of $4.93 billion (up 4.98% year-over-year), implying a price-to-sales ratio of roughly 5.1x. | 中 | SV027, SV028 |
| CV036 | Zoom's FY2026 Form 10-K reports gross margin of 77.0%, up from 75.8% the prior year, providing a mature public-company collaboration-software margin benchmark against which Granola's undisclosed, vendor-dependent cost structure cannot yet be compared. | 中 | SV012 |
| CV037 | Salesforce agreed on 2026-06-15 to acquire Fin (formerly Intercom), an AI customer-service agent platform serving over 30,000 businesses, for approximately $3.6 billion, illustrating the scale of price a strategic acquirer will pay for a proven, enterprise-deployed AI-agent platform in an adjacent category. | 高 | SV030, SV032 |
| CV038 | CNBC reported that global M&A deal value surged nearly 40% to a record $4.9 trillion in 2025, with AI-driven demand cited as a primary catalyst carrying into 2026, indicating a generally receptive strategic-acquirer environment for well-positioned AI software assets. | 中 | SV030 |
| CV039 | Crunchbase's 2026 IPO-candidate list highlights companies like Databricks ($4.8 billion revenue run-rate, 55%+ year-over-year growth, 140%+ net retention) and Canva ($3.3 billion annualized revenue, 240 million monthly users) as the disclosure and scale profile associated with near-term IPO readiness. | 中 | SV033 |
| CV040 | Granola discloses none of the revenue, retention, or user-count metrics that Crunchbase's IPO-prediction methodology relies on, making Granola's public-listing readiness unassessable by the same framework applied to disclosed 2026 IPO candidates. | 中 | SV033, SV002 |
| CV041 | Across the comparable set reviewed, Granola is the only company whose valuation (a $1.5 billion post-money mark) rests on zero disclosed revenue or ARR figure; Fireflies, Otter, Notion, Zoom, and Fin all have at least one disclosed revenue, ARR, or run-rate data point underpinning their respective valuation or purchase price. | 中 | SV025, SV031, SV027, SV030 |
| CV042 | The National Law Review states that AI notetakers are increasingly entangled in high-stakes litigation under federal and state wiretapping laws, citing In re Otter.ai Privacy Litigation as an active precedent. | 中 | SV018 |
| CV043 | Bloomberg Law states a vendor with data-mining rights over recorded meeting content can expose its customer company to vicarious liability for the vendor's own wiretap-statute violations, a mechanism relevant to any enterprise buyer of Granola's bot-free capture product. | 中 | SV019 |
| CV044 | Goodwin Procter states that California Invasion of Privacy Act violations can carry damages of up to $5,000 per violation, a per-incident exposure figure relevant to any consent-based class action targeting a meeting-recording vendor. | 中 | SV020 |
| CV045 | CompleteAI Training reports a federal class action in San Jose consolidating four lawsuits against Otter.ai alleging its tool records conversations and captures biometric voiceprint data without participant consent, the closest direct-peer litigation precedent to Granola's own bot-free, less-visible capture model. | 中 | SV021 |
| CV046 | An independent enterprise-readiness review scored Granola 3 out of 10, citing US-only AWS hosting with no EU region or on-premises/VPC option, a gap that could slow the large-enterprise contract expansion the bull case depends on. | 中 | SV022 |
| CV047 | The addressable market Granola competes in is sized very differently across analysts -- Precedence Research puts the narrower AI note-taking software market at $623.5 million in 2025 while Statifacts puts the broader AI meeting-assistant market at $4.41 billion in 2026 -- meaning total-addressable-market assumptions embedded in any valuation model carry wide, publisher-dependent uncertainty. | 中 | SV023, SV024 |
| CV048 | Because Granola's cost of revenue is dependent on third-party LLM-inference and transcription vendors rather than owned infrastructure, and because SaaS benchmark data flags sub-70% gross margin as typical for AI-inference-heavy businesses, Granola's margin trajectory versus Zoom's disclosed 77.0% and the broader 76% median SaaS benchmark remains an open, unverified question rather than a presumed SaaS-level outcome. | 中 | SV014, SV012 |
| CV049 | Weighing a $1.5 billion valuation with zero disclosed revenue against a private market that still broadly supports rising AI multiples (Carta) but has already begun compressing public 'Productivity Tools' multiples (Finro) and carries an active 11.4% down-round base rate (Carta), the evidence supports a research-more recommendation with low-to-medium confidence, a high risk rating, and a stretched valuation stance rather than a buy, hold, or avoid call. | 中 | SV002, SV026, SV029 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | Granola | Granola — The AI Notepad for back-to-back meetings | |
| SO002 | Granola | Granola raises $125M to put your company's context to work | |
| SO003 | Granola | Careers - Come build the future of work | |
| SO004 | Granola | Security at Granola | |
| SO005 | Granola | Pricing plans | |
| SO006 | TechCrunch | Granola raises $125M, hits $1.5B valuation as it expands from meeting notetaker to enterprise AI app | TechCrunch | |
| SO007 | Index Ventures | Index Ventures Leads Granola's $125M Series C | Index Ventures | |
| SO008 | Kleiner Perkins | Granola: Putting Your Company’s Context to Work | |
| SO009 | BusinessCloud | Granola, founded in 2023, becomes latest UK unicorn | |
| SO010 | SiliconANGLE | Granola raises $125M at $1.5B valuation for its AI note-taking app - SiliconANGLE | |
| SO011 | UKTN | AI notepad Granola secures unicorn status after Series C - UKTN | |
| SO012 | The SaaS News | Granola Raises $125M Series C at $1.5B Valuation | |
| SO013 | Apple App Store | Granola - AI Meeting Notes App - App Store | |
| SO014 | Granola | Granola raises $20M to build the AI notepad that makes you smarter | |
| SO015 | TechCrunch | VCs love using the AI meeting notepad Granola, so they gave it $20M | TechCrunch | |
| SO016 | Business Wire | Granola Launches AI Workspace for Teams and Raises $43M Series B | |
| SO017 | TechCrunch | AI notetaking app Granola raises $43M at $250M valuation, launches collaborative features | TechCrunch | |
| SO018 | The SaaS News | Granola Raises $43 Million in Series B | |
| SO019 | Granola | Vanta saves 260+ hours annually per user with Granola | |
| SO020 | Granola | Vercel's entire GTM team boosts meeting productivity with Granola | |
| SO021 | Granola | Brex saves 427 hours a week with Granola | |
| SO022 | Granola | Zapier saves 830+ hours per month with Granola | |
| SO023 | Granola Docs & Help Center | Privacy Policy - Granola Docs & Help Center | |
| SO024 | Granola Docs & Help Center | Data Processing Addendum - Granola Docs & Help Center | |
| SO025 | Shadow | Granola Encrypted Its Local Database. Here's Why That Matters and What to Use Instead. | |
| SO026 | Granola | Introducing Granola MCP | |
| SO027 | Granola | Introducing Recipes, in the all-new Granola Chat | |
| SO028 | Granola | New: Briefs prepare you for your next meeting as you join | |
| SM001 | Granola | Granola for Enterprise | Every meeting is full of decisions, context, and insights. Granola captures all of it and makes it searchable so the right information is always there when someone needs it. |
| SM002 | Granola | Brex saves 427 hours a week with Granola | Granola is one of the fastest tools the business has adopted organically, and the intuitive, friction-free experience has resulted in minimal administrative burden for IT. |
| SM003 | Granola | Vanta saves 260+ hours annually per user with Granola | When you're moving between nonstop meetings, it's not always possible to take immediately shareable notes. |
| SM004 | Granola | Zapier saves 830+ hours per month with Granola | Granola quickly evolved into a powerful knowledge repository for our whole team. |
| SM005 | Granola | Vercel's entire GTM team boosts meeting productivity with Granola | I can't afford to wait five hours to access my meeting notes. |
| SM006 | Granola | Careers - Come build the future of work | We've raised a $125M Series C at a $1.5bn valuation, led by Index Ventures and Kleiner Perkins. |
| SM007 | Granola | Granola raises $125M to put your company's context to work | Today we're announcing our $125M Series C at a $1.5bn valuation. |
| SM008 | Granola | Security at Granola | Notes are stored in our US-hosted AWS Virtual Private Cloud. |
| SM009 | Granola | Second brain for meetings: Why meeting search alone is not enough | The financial cost of this knowledge loss reaches $31.5 billion annually across Fortune 500 companies, according to IDC research. |
| SM010 | Granola | Choosing an AI meeting assistant for hybrid teams: The hybrid-specific criteria most lists ignore | Daversa Partners, an executive search firm, adopted a bot-free approach across 136 of 150 employees specifically because traditional tools were "intrusive" for CEO searches where discretion matters. |
| SM011 | Granola | Is it safe to record meetings with AI tools? A buyer's safety checklist for 2026 | Most frequent enterprise requirements are no audio storage after transcription, explicit AI training opt-outs, bot-free capture, SOC 2 Type 2 certification, and GDPR compliance. |
| SM012 | Granola | Sales call recording software: A buyer's guide (when you need Gong vs when notes are enough) | Foundations plan costs between $1,300 and $1,600 per user per year...plus a mandatory platform fee ranging from $5,000 to $50,000 per year. |
| SM013 | Granola | The AI notepad for Product teams - Granola | Whatever your role, we’ve got you covered: Product Managers, Product Designers, Product Leaders. |
| SM014 | Granola | The AI notepad for Sales teams - Granola | Pipe notes directly through to HubSpot, Affinity, Attio or Zapier. |
| SM015 | Fathom | Fathom AI Notetaker - Never Take Notes Again | Capture notes your way – bot or no bot – so you can stay focused on the meeting. |
| SM016 | Fathom | Pricing to supercharge every meeting | Free plan available alongside paid tiers spanning roughly $15 to $34 per user per month. |
| SM017 | Fireflies.ai | Fireflies.ai | #1 AI Assistant for Meetings, Email, Chat & CRM | Used across 1 million+ companies. |
| SM018 | Otter.ai | Meeting Notes - Real-time, Shareable, Secure | Otter.ai | Connect your Google or Microsoft calendar and schedule your Otter to auto-join Zoom, Microsoft Teams, or Google Meet. |
| SM019 | AI Note taking and Take Notes for Me in Google Meet with Gemini | Google Workspace | Gemini in Meet captures real-time notes and action items, so you can stay focused on the discussion. | |
| SM020 | Microsoft | Catch up on meetings with Microsoft 365 Copilot in Teams | Copilot in Teams can help you catch up on meetings you missed or need a recap of. |
| SM021 | Microsoft | Recap in Microsoft Teams | Microsoft Support | In a recap, you can find the event recording, transcript, shared files, notes, custom summary, agenda, and follow-up tasks. |
| SM022 | Zoom | Using Meeting Summary with AI Companion | Meeting summary with Zoom AI Companion...allows meeting hosts to initiate an AI-generated summary of their meeting. |
| SM023 | Microsoft WorkLab | Breaking down the infinite workday | Employees are interrupted every two minutes during core work hours — 275 times a day — by meetings, emails, or chats. |
| SM024 | Precedence Research | AI Note Taking Market Size to Hit USD 3476.74 million by 2035 | Revenue, 2025 USD 623.50 Mn; Forecast Year, 2035 USD 3476.74 Mn; CAGR, 2026-2035 18.75%. |
| SM025 | Statifacts | AI Meeting Assistant Market Size to Lead USD 35.02 Billion by 2035 | Market Size in 2025 USD 3.50 Billion; Market Size in 2026 USD 4.41 Billion; Market Size in 2032 USD 17.55 Billion. |
| SM026 | The Business Research Company | AI-Powered Meeting Assistants Market Size Report 2026 to 2035 | Market size has reached to $3.14 billion in 2025...Expected to grow to $9.33 billion in 2030 at a CAGR of 24.3%. |
| SM027 | GII Research (syndicating Grand View Research) | AI Meeting Assistant Market Size, Share & Trends Analysis Report | The global AI meeting assistant market size was estimated at USD 3.47 billion in 2025 and is projected to reach USD 21.48 billion by 2033. |
| SM028 | Congruence Market Insights | AI Note-taking Solutions Market Intelligence | The Global AI Note-taking Solutions Market was valued at USD 450.7 Million in 2024 and is anticipated to reach a value of USD 1,800.3 Million by 2032. |
| SM029 | Startup Fortune | Granola AI Meeting Notes Are Not as Private as You Might Think | A June 2024 analysis by PitchBook valued the global AI transcription and meeting intelligence market at $3.5 billion, projecting it to reach $11.8 billion by 2030. |
| SM030 | BuildBetter | Do they own your data? Granola.ai Privacy Policy Reviewed. | Final Enterprise Readiness Rating: 3/10. US-only, AWS hosted; no EU region or on-prem/VPC options. |
| SM031 | Shadow | Granola Encrypted Its Local Database. Here's Why That Matters — and What to Use Instead. | In a world where notes are managed by agents, the app now has zero value. |
| SM032 | tl;dv | Granola AI Review: My Honest Thoughts After 20+ Meetings (2026) | Granola AI Review: My Honest Thoughts After 20+ Meetings. |
| SM033 | TechCrunch | Granola raises $125M, hits $1.5B valuation as it expands from meeting notetaker to enterprise AI app | Users might not like bots in meetings visibly taking notes, but a lot of them don’t mind if an app on someone’s computer is doing the transcription. |
| SM034 | Grand View Research | AI Meeting Assistant Market Size & Share Report, 2026-2033 | |
| SP001 | Granola | AI Meeting Assistant - Without the Bot | Granola | Notes private by default. There's no third-party server logging your conversations. |
| SP002 | Granola | Granola — AI that already knows what you're working on (Chat) | Ask away — Answer questions about progress you're making, decisions you've made and pull user insights in a flash. |
| SP003 | Granola | Updates - Latest features and improvements from Granola | Granola Chat just got smarter — rebuilt to be agentic: faster across your notes, Team Space, and what's shared with you. |
| SP004 | Granola | How Granola thinks about designing agents | When rebuilding Granola Chat, we had a choice: should the agent read 10 meetings in detail, or 100 at a high level? We chose 100. |
| SP005 | Granola Docs | Security, Privacy & Data FAQs - Granola Docs & Help Center | For our full SOC 2 report, data handling practices, and compliance documentation, use our Trust Center (click "Request Access" in the top right). |
| SP006 | Granola | Sales call recording software: A buyer's guide (when you need Gong vs when notes are enough) | Conversation intelligence platforms like Gong cost roughly $1,300 to $1,600 per user per year based on third-party buyer estimates, plus a mandatory platform fee. |
| SP007 | Granola | Granola for Enterprise | Every meeting is full of decisions, context, and insights. |
| SP008 | Granola | Security at Granola | Notes are stored in our US-hosted AWS Virtual Private Cloud. |
| SP009 | Fathom | Fathom AI Notetaker - Never Take Notes Again | Capture notes your way – bot or no bot – so you can stay focused on the meeting. |
| SP010 | Fathom | Pricing to supercharge every meeting | Team $19/$15 per user/month; Business $34/$25 per user/month; Public API & MCP; Claude & ChatGPT Integrations. |
| SP011 | Fireflies.ai | Fireflies.ai | #1 AI Assistant for Meetings, Email, Chat & CRM | USED ACROSS 1 MILLION+ COMPANIES. SOC 2 Type II. GDPR. HIPAA Compliant. Zero Data Retention. |
| SP012 | Fireflies.ai | Pricing & Plans | Fireflies.ai AI App for Smarter Meeting Notes | Pro $18 $10 per seat/month billed annually; Business $29 $19 (Most Popular); Enterprise $39 per seat/month, annual only. |
| SP013 | Otter.ai | Meeting Notes - Real-time, Shareable, Secure | Otter.ai | Connect your Google or Microsoft calendar and schedule your Otter to auto-join Zoom, Microsoft Teams, or Google Meet. |
| SP014 | Otter.ai | Pricing | Otter.ai | Pro $16.99/user/month ($8.49 promotional); Business $30 ($24 promotional); Otter MCP server integration for AI assistants. |
| SP015 | Otter.ai | Otter.ai Evolves from AI Notetaker to Create $100B Enterprise Conversational Knowledge Engine Market | Otter.ai estimates the total opportunity exceeds $100 billion... more than 35 million people, across over one billion meetings. |
| SP016 | tl;dv | tl;dv - The AI Notetaker built for Team Collaboration | 2M+ users. GDPR & SOC2. NO BOT REQUIRED. |
| SP017 | tl;dv | Granola AI Review: My Honest Thoughts After 20+ Meetings (2026) | Granola vs tl;dv: Bot-free vs Bot — Which Wins on Accuracy? Transcription accuracy 97%+ (tl;dv) vs ~94-95% (Granola). |
| SP018 | Read AI | Meeting Summaries, Transcripts, AI Notetaker & Enterprise Search | read.ai | Meet Ada, Your Digital Twin. Zoom selects Read as an Essential App, Premium Features Enabled for Zoom Customers. |
| SP019 | Read AI | Plans & Pricing | Pro $15/$19.75 per user/month; Team $22.50/$29.75; Enterprise $39.75, requires 5+ licenses. |
| SP020 | Gong | Gong - Revenue AI OS | Trusted by 5,000+ customers. #1 AI OS for Revenue Teams. |
| SP021 | AI Note taking and Take Notes for Me in Google Meet with Gemini | Google Workspace | Gemini in Meet captures real-time notes and action items, so you can stay focused on the discussion. | |
| SP022 | Microsoft | Catch up on meetings with Microsoft 365 Copilot in Teams | Intelligent recap experiences are available as part of Teams Premium... or as part of the Microsoft 365 Copilot license. |
| SP023 | Microsoft | Recap in Microsoft Teams | Microsoft Support | Intelligent recap experiences are available as part of Teams Premium, an add-on license... or as part of the Microsoft 365 Copilot license. |
| SP024 | Microsoft | Microsoft Teams Premium | Microsoft Teams | Teams Premium pricing: $10.00 user/month, paid yearly. |
| SP025 | Zoom | Using Meeting Summary with AI Companion | A Licensed user on a Zoom Workplace Pro, Business, or Enterprise account... Meeting Summary enabled. |
| SP026 | Zoom | Work smarter, create faster: Zoom AI Companion drives conversations to completion | AI Companion Monthly Active Users more than tripled year over year in Q4 FY26. |
| SP027 | CompleteAITraining | Otter.ai Faces Consolidated Class Action Over Meeting Recording Practices | A federal class action in San Jose is consolidating four lawsuits against Otter.ai... Similar litigation has already targeted Fireflies.ai. Other AI meeting tools, including Read AI, face university bans due to consent concerns. |
| SP028 | CloudTalk | Gong Pricing in 2026: Full Guide and Plan Breakdown | Gong Foundations (base plan) — ~$1,600/user/year... Mandatory platform fee — $5,000-$50,000/year. |
| SP029 | CostBench | Best AI Meeting Assistants for Enterprise 2026: Top 6 Ranked & Reviewed | tl;dv ($0-$98/user/month)... Otter.ai ($0-$30/user/month)... Read AI ($0-$39.75/user/month). |
| SP030 | Laxis | The State of Meeting Note-Taking 2026: AI Adoption, ROI & Market Benchmarks | 75% of professionals now use an AI note-taker... Zoom AI Companion reached 510,000 accounts within five months of launch... Fathom holds the highest G2 rating in the category (5.0/5 across 6,000+ reviews). |
| SP031 | TechCrunch | These AI notetaking devices can help you record and transcribe your meetings | Digital meeting notetakers like Read AI, Fireflies.ai, Fathom, and Granola help record and transcribe online meetings. But for in-person or more versatile options, many people prefer physical recording devices. |
| SP032 | TechCrunch | Granola raises $125M, hits $1.5B valuation as it expands from meeting notetaker to enterprise AI app | As meeting note-taking becomes a commonplace feature... Read AI, Fireflies, and Quill have already started working in this direction. |
| SP033 | Shadow | Granola Encrypted Its Local Database. Here's Why That Matters — and What to Use Instead. | In a world where notes are managed by agents, the app now has zero value. |
| SP034 | BuildBetter | Do they own your data? Granola.ai Privacy Policy Reviewed. | Final Enterprise Readiness Rating: 3/10. US-only, AWS hosted; no EU region or on-prem/VPC options. |
| SP035 | Startup Fortune | Granola AI Meeting Notes Are Not as Private as You Might Think | Granola, a popular AI meeting notepad, exposes user notes to anyone with a link by default and trains internal AI models on your transcripts unless you manually opt out. |
| SI001 | Granola | Pricing plans | Granola | Get the Granola plan that is right for you and your team, so everyone can be on the same page. |
| SI002 | Granola | Second brain for meetings: organizational memory | Granola blog | In the Business plan, you pay $14 per user per month for unlimited meeting history, CRM integrations, and folder-level queries across all conversations your team has had. |
| SI003 | Granola | Granola for Enterprise | SSO / SAML ... SCIM provisioning ... SOC 2 Type II certified. Independently audited. Security documentation available for procurement. |
| SI004 | Granola | Granola raises $125M to put your company's context to work | A personal API for people to access their notes...(available on Business and Enterprise plans). An enterprise API, for admins to work with team context (available on Enterprise plan). |
| SI005 | Granola | Careers - Come build the future of work | Granola | Each day Granola transcribes millions of minutes of conversation. Teams that adopt it tend to stick around -- and it spreads quickly from leaders to entire organizations. |
| SI006 | Granola | Brex saves 427 hours a week with Granola | Implementation started with a trial run on the Business plan... within weeks, Brex was live with Granola Enterprise. |
| SI007 | Granola | Vanta saves 260+ hours annually per user with Granola | Vanta saves 260+ hours annually per user with Granola. |
| SI008 | Granola | Zapier saves 830+ hours per month with Granola | Granola quickly evolved into a powerful knowledge repository for our whole team... cut back on the total number of meetings we need to have. |
| SI009 | Granola | Vercel's entire GTM team boosts meeting productivity with Granola | It's actually unbelievable how good granola.ai is. It replaces writing documents. |
| SI010 | Granola | Sales call recording software: A buyer's guide (when you need Gong vs when notes are enough) | If you're a founder or running a team under 15, an AI notepad solves the actual bottleneck. Gong's Foundations plan costs between $1,300 and $1,600 per user per year... plus a mandatory platform fee ranging from $5,000 to $50,000 per year. |
| SI011 | Granola | Security at Granola | Granola uses best-in-class transcription providers (like Deepgram and Assembly) and AI providers (like OpenAI and Anthropic) to summarize your meeting. |
| SI012 | Companies House (UK Government) | GRANOLA LABS LTD overview - Find and update company information | GRANOLA LABS LTD. Incorporated on 3 March 2023. Nature of business (SIC) 62012 - Business and domestic software development. |
| SI013 | Companies House (UK Government) | GRANOLA LABS LTD filing history - Find and update company information | AA Total exemption full accounts made up to 31 March 2025 ... Cessation of Christopher Jonathan Pedregal as a person with significant control on 14 April 2023. |
| SI014 | U.S. Securities and Exchange Commission (EDGAR) / Zoom Communications, Inc. | Zoom Communications, Inc. Form 10-K for fiscal year ended January 31, 2026 | Gross margin increased to 77.0% for the fiscal year ended January 31, 2026 from 75.8% for the fiscal year ended January 31, 2025. |
| SI015 | TechCrunch | Granola raises $125M, hits $1.5B valuation as it expands from meeting notetaker to enterprise AI app | With this round, which comes less than a year after its $43 million round, the startup has raised $192 million. |
| SI016 | SiliconANGLE | Granola raises $125M at $1.5B valuation for its AI note-taking app | Bloomberg reported that the investment follows a quarter in which Granola’s revenue grew by 250%. ... The $14-per-month Business plan... The top-end Enterprise subscription, which is priced at $35 per month. |
| SI017 | The SaaS News | Granola Raises $125M Series C at $1.5B Valuation | bringing Granola’s total funding to $192 million. |
| SI018 | BusinessCloud | Granola, founded in 2023, becomes latest UK unicorn | It takes total funding to around $188m (£141m at today’s exchange rate). |
| SI019 | UKTN | AI notepad Granola secures unicorn status after Series C | Granola, a London-based AI-powered notepad startup has been valued at $1.5bn after securing a $125m (£93.4m) Series C investment. |
| SI020 | TechFundingNews | Granola raises $125M Series C to turn company conversations into context | The London startup just raised $125 million in a Series C round at a $1.5 billion valuation, led by Danny Rimer from Index Ventures and Mamoon Hamid from Kleiner Perkins. |
| SI021 | StartupRise | London-based Granola hits unicorn status with $125M funding | the fresh funding will be used to further scale its product and support continued growth. |
| SI022 | Index Ventures | Index Ventures Leads Granola's $125M Series C | These moments turn every user into an evangelist, and it’s a pattern we’ve seen before: starting in the individual grassroots, spreading across teams, and quickly becoming instrumental to how an enterprise functions. |
| SI023 | Kleiner Perkins | Granola: Putting Your Company's Context to Work | Teams at Asana, Vanta, Gusto, Thumbtack, Cursor, and Lovable rely on it daily, and what started as a personal productivity tool is quickly becoming company-wide infrastructure. |
| SI024 | Tracxn | Granola - 2026 Company Profile & Team | Granola has 116 employees as of May 26 [2026]. Granola has raised a total funding of $192M over 4 rounds. |
| SI025 | SaaSCalcHub | 2026 SaaS Industry Benchmarks (Growth, CAC, NDR, Rule of 40) | Subscription gross margin has remained stable around 76% at the median, with top-quartile companies at 84%+... Companies below 70% gross margin are usually selling something that has unusual variable costs (heavy compute, payments interchange, AI inference) or are pricing too low. |
| SI026 | Fireflies.ai | Pricing & Plans | Fireflies.ai AI App for Smarter Meeting Notes | Pro $18 $10 per seat/month billed annually; Business $29 $19 (Most Popular); Enterprise $39 per seat/month, annual only. |
| SI027 | Otter.ai | Pricing | Otter.ai | Pro $16.99/user/month ($8.49 promotional); Business $30 ($24 promotional). |
| SI028 | CloudTalk | Gong Pricing in 2026: Full Guide and Plan Breakdown | Gong Foundations (base plan) — ~$1,600/user/year... Mandatory platform fee — $5,000–$50,000/year. |
| SI029 | Fathom | Pricing to supercharge every meeting | Fathom | Team $19/$15 per user/month (2 user min); Business $34/$25; Enterprise: Talk to sales. |
| SI030 | TheNextWeb | London's Granola raises $125M to turn meeting recordings into enterprise AI infrastructure | A $1.5 billion valuation for a company that was worth $250 million ten months ago invites scrutiny. Granola has not disclosed revenue figures, user counts, or retention metrics publicly. |
| SI031 | BritishWire | Granola $125M Series C: Is The $1.5B AI Notetaker A Bubble Or A Bargain? | A brutal market correction feels absolutely inevitable once these highly valued startups are finally forced to show their actual revenue sheets. |
| SI032 | Laffaz | Granola went from 5,000 weekly users to a $1.5B unicorn — without a meeting bot | By October 2024, when Granola raised its $20 million Series A, it had 5,000 weekly users... The company has not disclosed its current user count. Fireflies.ai has surpassed 16 million users. |
| SI033 | Shadow | Granola Encrypted Its Local Database. Here's Why That Matters -- and What to Use Instead. | Granola encrypted its local database... Granola’s co-founder Chris Pedregal responded in the thread: "Hey Guido, we’re not going closed. Our MCP lets you access everything." |
| SE001 | Granola | AI Meeting Assistant - Without the Bot | Granola | Granola's meeting assistant takes a different approach than bot-centered tools. Our AI notepad runs on your device, captures your meeting audio locally. |
| SE002 | Granola | Granola — AI that already knows what you're working on (Chat) | Recipes are simple shortcuts for complex questions or your most-used prompts. Share your Recipes with others or borrow one from an expert. |
| SE003 | Granola | Granola for Enterprise | SSO / SAML ... SCIM provisioning ... SOC 2 Type II certified. Independently audited. Security documentation available for procurement. |
| SE004 | Granola | Security at Granola | Granola uses best-in-class transcription providers (like Deepgram and Assembly) and AI providers (like OpenAI and Anthropic) to summarize your meeting. |
| SE005 | Granola | Updates - Latest features and improvements from Granola | Introducing Granola MCP. Granola MCP connects your meeting notes to Claude, ChatGPT, Cursor, and other AI tools via Model Context Protocol. |
| SE006 | Granola | Careers - Come build the future of work | Granola | We're growing. Each day Granola transcribes millions of minutes of conversation. |
| SE007 | Granola | Granola raises $125M to put your company's context to work | A personal API for people to access their notes...(available on Business and Enterprise plans). An enterprise API, for admins to work with team context (available on Enterprise plan). |
| SE008 | Granola | New: Briefs prepare you for your next meeting as you join | Overnight, Granola starts preparing for your day ahead. First, it checks who you're meeting. |
| SE009 | Granola | Second brain for meetings: Why meeting search alone is not enough | A second brain for meetings applies the CODE framework (Capture, Organize, Distill, Express) to build compounding organizational memory. |
| SE010 | Granola | Introducing Recipes, in the all-new Granola Chat | Today, we're launching with Recipes written by Lenny Rachitsky, Matt Mochary, Ridd, and lots of other experts. |
| SE011 | Granola | Local-first AI notetakers vs. cloud-based: Enterprise security & data control | Granola's zero-audio-retention model is why the company completed SOC 2 Type 2 certification in three months. ... Granola does not currently hold a HIPAA Business Associate Agreement. |
| SE012 | Granola | Data residency & deployment models for enterprise AI notetakers: On-premise vs. cloud | Granola achieved SOC 2 Type 2 certification in three months rather than the typical twelve to eighteen. |
| SE013 | Apple | Granola - AI Meeting Notes App - App Store | Granola is the AI notepad for people in back-to-back meetings, trusted by top founders, VCs, and recruiters. And now it's available on your iPhone. |
| SE014 | Granola | Data Processing Addendum | In the event of a Personal Data Breach, Granola shall, without undue delay, inform Customer of the Personal Data Breach. |
| SE015 | Granola | Privacy Policy | Personal Data may be collected, used, and disclosed with the government, including law enforcement, or other parties to meet certain legal requirements. |
| SE016 | Granola | Platform Terms of Service (Terms & Policies index) | Vulnerability Disclosure Policy ... Security Reports: Post-Mortem: AssemblyAI API Key Exposure, Post-Mortem: Legacy Unmanaged Google Accounts Workspace Auto-Join, Post-Mortem: Google Workspace Session Logout Vulnerability. |
| SE017 | Granola | Introducing Granola MCP | MCP (Model Context Protocol) is a standard that lets AI tools connect to external data sources... If you're on a Granola Enterprise plan, MCP is available in early access beta. It's off by default. |
| SE018 | Granola | How Granola thinks about designing agents | Transcripts are the most vivid reenactment of what happened in the meeting, but they are messy... roughly 10x the number of tokens compared to meeting notes. |
| SE019 | Granola | So, you think it's easy to change an app icon? | Granola is an Electron app, so we started with the obvious. |
| SE020 | Granola | Security, Privacy & Data FAQs - Granola Docs & Help Center | Security & Compliance Overview, Data Storage & Residency, Healthcare & HIPAA, Education & FERPA, Note Privacy & Sharing, AI & Model Training. |
| SE021 | Granola | Our Security Standards - Granola Docs & Help Center | We've made our security information easily accessible across several dedicated resources: Security Page, Trust Center, Privacy and Security Policies. |
| SE022 | Granola | Post-Mortem: AssemblyAI API Key Exposure | An unauthenticated feature-flag endpoint exposed an AssemblyAI API key; impact was limited to iOS TestFlight beta testers between November 2024 and March 2025. |
| SE023 | Granola | Post-Mortem: Legacy Unmanaged Google Accounts Workspace Auto-Join | Auto-provisioning for workspaces was based only on a user's Google OAuth email domain, until a fix on May 23, 2025 required explicit Google Workspace membership. |
| SE024 | Granola | Post-Mortem: Google Workspace Session Logout Vulnerability | 187 Granola Workspace users removed from their linked Google Workspace were not automatically logged out of the Granola app. |
| SE025 | Tenable | Granola API Endpoint Information Disclosure (TRA-2025-07) | One of those endpoints, "get-feature-flags," returns data without authentication. This includes a value called "assembly_key." |
| SE026 | pedramamini (GitHub) | GranolaMCP: Programmatic, CLI, and MCP access to Granola.ai data | GranolaMCP operates entirely on local cache files - it reads meeting data directly from Granola's local cache file (cache-v3.json) without making any API calls to Granola's servers. |
| SE027 | Product Hunt (via Wayback Machine) | Granola Forums on Product Hunt | I've tested so many AI Note takers as of late... Recently I've been using Granola's new mobile app for in-person convos... in Japanese... Granola captures all key points and WRITES THE NOTES IN ENGLISH. |
| SE028 | IsDown | Is Granola Down? Check current status and user reports | In the last 90 days, Granola had 4 incidents with a median duration of 2 hours 34 minutes. |
| SE029 | Granola (customer case study) | Brex saves 427 hours a week with Granola | Over 50% adoption rate across knowledge workers in 6 months; 25k+ hours saved since adopting Granola; 427 hours saved per week. |
| SE030 | Granola (customer case study) | Vanta saves 260+ hours annually per user with Granola | 69% adoption rate across the organization; 61% increase in meeting visibility with Granola. |
| SE031 | Granola (customer case study) | Zapier saves 830+ hours per month with Granola | 830+ hours saved per month; 52% organization-wide adoption rate. |
| SE032 | Granola (customer case study) | Vercel's entire GTM team boosts meeting productivity with Granola | 100% adoption rate across GTM teams; 11 hours saved weekly per user; 25% GTM team productivity increase. |
| SE033 | BuildBetter | Do they own your data? Granola.ai Privacy Policy Reviewed. | Final Enterprise Readiness Rating: 3/10. Not enterprise-ready (Reviewed 2025). |
| SE034 | Shadow | Granola Encrypted Its Local Database. Here's Why That Matters -- and What to Use Instead. | Sorry to see Granola @meetgranola going closed. They encrypted their local db, no local and no cloud API... In a world where notes are managed by agents, the app now has zero value. |
| SE035 | Startup Fortune | Granola AI Meeting Notes Are Not as Private as You Might Think | Granola, a popular AI meeting notepad, exposes user notes to anyone with a link by default and trains internal AI models on your transcripts unless you manually opt out. |
| SE036 | tl;dv | Granola AI Review: My Honest Thoughts After 20+ Meetings (2026) | Transcription accuracy: tl;dv 97%+, Granola ~94-95%. Languages supported: tl;dv 30+, Granola ~15. |
| SE037 | Granola | Integrations page (granola.ai/integrations) | HTTP 404 returned; page title falls back to the generic Granola homepage title with no integration catalog content rendered. |
| SU001 | Granola (customer case study) | Zapier saves 830+ hours per month with Granola | 830+ hours saved per month. 52% organization-wide adoption rate. |
| SU002 | Granola (customer case study) | Vanta saves 260+ hours annually per user with Granola | 260+ hours saved annually per user. 69% adoption rate across the organization. 61% increase in meeting visibility with Granola. |
| SU003 | Granola (customer case study) | Brex saves 427 hours a week with Granola | Over 50% adoption rate across knowledge workers in 6 months. 25k+ hours saved since adopting Granola. 427 hours saved per week. |
| SU004 | Granola (customer case study) | Vercel's entire GTM team boosts meeting productivity with Granola | 100% adoption rate across GTM teams. 11 hours saved weekly per user. 25% GTM team productivity increase. |
| SU005 | Granola | Local-first AI notetakers vs. cloud-based: Enterprise security & data control | Daversa Partners adopted Granola across 136 of 150 employees because, in the words of president Laura Kinder, traditional tools felt "intrusive" for sensitive conversations. |
| SU006 | Granola | Careers - Come build the future of work | Granola | Granola is trusted by teams we admire -- from companies like Vanta, Brex, and Intercom to fast-growing startups like Cursor, Linear, and Replit... Teams that adopt it tend to stick around -- and it spreads quickly from leaders to entire organizations. |
| SU007 | Granola | Granola for Enterprise | Trusted by teams in tech, finance, media, and more. It starts with one person who loves Granola. It spreads because it makes every individual better at their job. |
| SU008 | Product Hunt (via Wayback Machine) | Granola Forums on Product Hunt | Community discussion citing in-person, multilingual conversation capture (e.g., translating a Japanese conversation into English notes) as a real-world mobile use case. |
| SU009 | Product Hunt | Granola Reviews (2026) | Product Hunt | Users repeatedly praise the clear, structured output, strong search and cross-meeting recall... A few drawbacks come up: no Android app, limits around iPad/iOS meeting capture, some account setup friction, occasional UI lag. |
| SU010 | G2 (via Wayback Machine) | Granola Reviews & Product Details | Reviews (19) ... 4.9 ... Small-Business (50 or fewer emp.), Enterprise (>1000 emp.), Mid-Market (51-1000 emp.) reviewers across Information Technology and Services, Financial Services, Clinical/healthcare-adjacent roles. |
| SU011 | The Business Dive | Granola Review 2026: Where It Shines And Where It Fails | Why Granola scored 3.6 out of 5 ... limited feature depth, reflected in lower scores for collaboration features (2.5/5) and especially other capabilities (2/5). |
| SU012 | Nubia Magazine | Granola AI Review in 2026: windows, android, founder, funding, AI & FAQs | Final Nubia Magazine rating: 4.5 out of 5. Recommended for founders, consultants, account executives, podcasters, journalists, coaches and anyone whose calendar looks like a wall of meetings. |
| SU013 | Best Reviews | Granola Reviews 2026 by Experts & Users | There are a few downsides, such as the lack of support for additional languages, the limited iOS app, and the lack of speaker labeling on the desktop app. |
| SU014 | YipitData | Is Granola the new leader in AI notetaking? | Signals data — drawn from real software transactions across 900+ mid-market companies — shows Granola posting 3x spend growth over the last six months, near-zero churn, and a pattern of displacing competitors like Fathom, Otter, and Fireflies.ai. |
| SU015 | Shadow | Granola Encrypted Its Local Database. Here's Why That Matters -- and What to Use Instead. | "Sorry to see Granola @meetgranola going closed... Any recommendations for good alternatives? What are you switching to?" 630 people liked it. 577 bookmarked it. 177 replied. |
| SU016 | Startup Fortune | Granola AI Meeting Notes Are Not as Private as You Might Think | Granola, a popular AI meeting notepad, exposes user notes to anyone with a link by default and trains internal AI models on your transcripts unless you manually opt out. |
| SU017 | BuildBetter | Do they own your data? Granola.ai Privacy Policy Reviewed. | Final Enterprise Readiness Rating: 3/10. Not enterprise-ready (Reviewed 2025). |
| SU018 | TechCrunch | Granola raises $125M, hits $1.5B valuation as it expands from meeting notetaker to enterprise AI app | The API launch comes after a bunch of users, including an a16z partner, were mad at Granola for locking down its local database and breaking on-device AI agent workflows they had set up. |
| SU019 | Businesswire (Granola press release) | Granola Launches AI Workspace for Teams and Raises $43M Series B | Shared Team Folders: Create dedicated spaces for Sales Calls, Customer Feedback, Hiring Loops, Weekly Syncs, and more. |
| SU020 | Laffaz | Granola went from 5,000 weekly users to a $1.5B unicorn — without a meeting bot | By October 2024 ... it had 5,000 weekly users ... The company has not disclosed its current user count. ... Granola has not disclosed revenue, growth rates, or retention data publicly. |
| SU021 | Kleiner Perkins | Granola: putting your company's context to work | Teams at Asana, Vanta, Gusto, Thumbtack, Cursor, and Lovable rely on it daily, and what started as a personal productivity tool is quickly becoming company-wide infrastructure. |
| SU022 | Index Ventures | Granola: investing in what makes us human | One customer told us they love how Granola plays on the notion of memory, and how re-reading notes surfaced different things each time. |
| SU023 | Hacker News | Open-source granola (meetings summary) | Granola in title is reference to https://www.granola.ai/ |
| SU024 | Reworked | Granola Hits Unicorn Status With $125M in Funding | Granola's growth has happened largely through word of mouth, particularly among founders and investors in the tech industry... The company has reported 10% weekly user growth since launch. |
| SU025 | MarTech360 | Granola Launches AI Workspace for Teams and Raises $43M Series B | Granola 2.0 ... transforms the company's popular AI meeting assistant into a powerful, intelligent workspace for teams. |
| SU026 | tl;dv (competitor blog) | Granola AI Review: My Honest Thoughts After 20+ Meetings (2026) | Transcription accuracy ~94-95% (Granola) vs 97%+ (tl;dv). Languages supported ~15 (Granola) vs 30+ (tl;dv). |
| SU027 | Granola | Granola raises $125M to put your company's context to work | A personal API for people to access their notes...(available on Business and Enterprise plans). An enterprise API, for admins to work with team context (available on Enterprise plan). |
| SU028 | Granola | Sales call recording software: a buyer's guide -- when you need Gong vs. when notes are enough | Granola's Business plan costs $14 per user per month. |
| SR001 | Granola | Data Processing Addendum | In the event of a Personal Data Breach, Granola shall, without undue delay, inform Customer of the Personal Data Breach. |
| SR002 | Granola | Privacy Policy | Personal Data may be collected, used, and disclosed with the government, including law enforcement, or other parties to meet certain legal requirements. |
| SR003 | Granola | Platform Terms of Service (Terms & Policies index) | Vulnerability Disclosure Policy ... Security Reports: Post-Mortem: AssemblyAI API Key Exposure, Post-Mortem: Legacy Unmanaged Google Accounts Workspace Auto-Join, Post-Mortem: Google Workspace Session Logout Vulnerability. |
| SR004 | Information Commissioner's Office (UK) | Employment practices and data protection: monitoring workers | This guidance discusses the monitoring of workers by employers, and how this interacts with data protection... help provide greater regulatory certainty; protect workers' data protection rights. |
| SR005 | Information Commissioner's Office (UK) | Guidance on AI and data protection | The Guidance on AI and Data Protection has been updated after requests from UK industry to clarify requirements for fairness in AI. |
| SR006 | European Commission (Digital Strategy) | AI Act | The AI Act defines 4 levels of risk for AI systems... Transparency rules of the AI Act will come into effect in August 2026. |
| SR007 | California Legislative Information (via Wayback Machine) | California Penal Code Section 632 (Invasion of Privacy) | A person who, intentionally and without the consent of all parties to a confidential communication, uses an electronic amplifying or recording device to eavesdrop upon or record the confidential communication... shall be punished by a fine not exceeding two thousand five hundred dollars ($2,500) per violation. |
| SR008 | Illinois General Assembly | 740 ILCS 14/15 -- Biometric Information Privacy Act: Retention, collection, disclosure, destruction | No private entity may collect, capture, purchase, receive through trade, or otherwise obtain a person's...biometric identifier or biometric information, unless it first informs the subject in writing...and receives a written release. |
| SR009 | U.S. Department of Health and Human Services (Office for Civil Rights) | Summary of the HIPAA Privacy Rule | When a covered entity uses a contractor or other non-workforce member to perform 'business associate' services...the Rule requires that the covered entity include certain protections for the information in a business associate agreement. |
| SR010 | National Law Review / Foley & Lardner | When AI Notetakers Take the Stand -- The Legal Risks Lurking in Your Virtual Meetings | These AI notetakers are increasingly finding themselves, and their users, entangled in high-stakes litigation under federal and state wiretapping laws... In re Otter.ai Privacy Litigation. |
| SR011 | Bloomberg Law | AI Notetaking Poses Wiretapping, Discovery, and Ethical Pitfalls | A vendor with data-mining rights can expose a customer company to vicarious liability for the vendor's CIPA violations... the case is In re Otter.AI Privacy Litigation, N.D. Cal., No. 25-cv-06911. |
| SR012 | Goodwin Procter LLP | AI Transcription Tools Under Scrutiny: Navigating Privacy Risks and Practical Mitigation Strategies | Failing to comply with these wiretap laws may expose organizations and their vendors to civil liability... violations of the California Invasion of Privacy Act can result in organizations facing damages of up to $5,000 per violation. |
| SR013 | CompleteAI Training | Otter.ai Faces Consolidated Class Action Over Meeting Recording Practices | A federal class action in San Jose is consolidating four lawsuits against Otter.ai, alleging the company's meeting transcription tool records conversations and captures biometric data without participant consent. |
| SR014 | Companies House (UK Government) | GRANOLA LABS LTD filing history - Find and update company information | AA Total exemption full accounts made up to 31 March 2025 ... Cessation of Christopher Jonathan Pedregal as a person with significant control on 14 April 2023. |
| SR015 | Companies House (UK Government) | GRANOLA LABS LTD overview - Find and update company information | GRANOLA LABS LTD. Incorporated on 3 March 2023. Nature of business (SIC) 62012 - Business and domestic software development. |
| SR031 | U.S. Securities and Exchange Commission (EDGAR) / Zoom Communications, Inc. | Zoom Communications, Inc. Form 10-K for fiscal year ended January 31, 2026 | Gross margin increased to 77.0% for the fiscal year ended January 31, 2026 from 75.8% for the fiscal year ended January 31, 2025. |
| SR016 | Granola | Security at Granola | Granola uses best-in-class transcription providers (like Deepgram and Assembly) and AI providers (like OpenAI and Anthropic) to summarize your meeting. |
| SR017 | Granola | Local-first AI notetakers vs. cloud-based: Enterprise security & data control | Granola's zero-audio-retention model is why the company completed SOC 2 Type 2 certification in three months... Granola does not currently hold a HIPAA Business Associate Agreement. |
| SR023 | Granola | Granola for Enterprise | SSO / SAML ... SCIM provisioning ... SOC 2 Type II certified. Independently audited. Security documentation available for procurement. |
| SR025 | Granola | Granola raises $125M to put your company's context to work | A personal API for people to access their notes...(available on Business and Enterprise plans). An enterprise API, for admins to work with team context (available on Enterprise plan). |
| SR046 | Granola | Careers - Come build the future of work | We're growing fast and hiring across engineering, design, and go-to-market. |
| SR018 | Granola | Post-Mortem: AssemblyAI API Key Exposure | An unauthenticated feature-flag endpoint exposed an AssemblyAI API key; impact was limited to iOS TestFlight beta testers between November 2024 and March 2025. |
| SR019 | Granola | Post-Mortem: Legacy Unmanaged Google Accounts Workspace Auto-Join | Auto-provisioning for workspaces was based only on a user's Google OAuth email domain, until a fix on May 23, 2025 required explicit Google Workspace membership. |
| SR020 | Granola | Post-Mortem: Google Workspace Session Logout Vulnerability | 187 Granola Workspace users removed from their linked Google Workspace were not automatically logged out of the Granola app. |
| SR021 | Tenable | Granola API Endpoint Information Disclosure (TRA-2025-07) | One of those endpoints, "get-feature-flags," returns data without authentication. This includes a value called "assembly_key." |
| SR024 | Granola | Security, Privacy & Data FAQs - Granola Docs & Help Center | Security & Compliance Overview, Data Storage & Residency, Healthcare & HIPAA, Education & FERPA, Note Privacy & Sharing, AI & Model Training. |
| SR022 | IsDown | Is Granola Down? Check current status and user reports | In the last 90 days, Granola had 4 incidents with a median duration of 2 hours 34 minutes. |
| SR026 | TechCrunch | Granola raises $125M, hits $1.5B valuation as it expands from meeting notetaker to enterprise AI app | With this round, which comes less than a year after its $43 million round, the startup has raised $192 million. |
| SR027 | TheNextWeb | London's Granola raises $125M to turn meeting recordings into enterprise AI infrastructure | A $1.5 billion valuation for a company that was worth $250 million ten months ago invites scrutiny. Granola has not disclosed revenue figures, user counts, or retention metrics publicly. |
| SR028 | BritishWire | Granola $125M Series C: Is The $1.5B AI Notetaker A Bubble Or A Bargain? | A brutal market correction feels absolutely inevitable once these highly valued startups are finally forced to show their actual revenue sheets. |
| SR037 | Startup Fortune | Granola AI Meeting Notes Are Not as Private as You Might Think | Granola, a popular AI meeting notepad, exposes user notes to anyone with a link by default and trains internal AI models on your transcripts unless you manually opt out. |
| SR043 | TechCrunch | These AI notetaking devices can help you record and transcribe your meetings | Digital meeting notetakers like Read AI, Fireflies.ai, Fathom, and Granola help record and transcribe online meetings. But for in-person or more versatile options, many people prefer physical recording devices. |
| SR044 | BusinessCloud | Granola, founded in 2023, becomes latest UK unicorn | |
| SR045 | TechCrunch | VCs love using the AI meeting notepad Granola, so they gave it $20M | |
| SR029 | Tracxn | Granola - 2026 Company Profile & Team | Granola has 116 employees as of May 26 [2026]. Granola has raised a total funding of $192M over 4 rounds. |
| SR030 | SaaSCalcHub | 2026 SaaS Industry Benchmarks (Growth, CAC, NDR, Rule of 40) | Subscription gross margin has remained stable around 76% at the median, with top-quartile companies at 84%+... Companies below 70% gross margin are usually selling something that has unusual variable costs (heavy compute, payments interchange, AI inference). |
| SR034 | YipitData | Is Granola the new leader in AI notetaking? | Signals data -- drawn from real software transactions across 900+ mid-market companies -- shows Granola posting 3x spend growth over the last six months, near-zero churn, and a pattern of displacing competitors like Fathom, Otter, and Fireflies.ai. |
| SR042 | Laxis | The State of Meeting Note-Taking 2026: AI Adoption, ROI & Market Benchmarks | 75% of professionals now use an AI note-taker... Zoom AI Companion reached 510,000 accounts within five months of launch... Fathom holds the highest G2 rating in the category (5.0/5 across 6,000+ reviews). |
| SR032 | Index Ventures | Index Ventures Leads Granola's $125M Series C | These moments turn every user into an evangelist, and it's a pattern we've seen before: starting in the individual grassroots, spreading across teams, and quickly becoming instrumental to how an enterprise functions. |
| SR033 | Kleiner Perkins | Granola: Putting Your Company's Context to Work | Teams at Asana, Vanta, Gusto, Thumbtack, Cursor, and Lovable rely on it daily, and what started as a personal productivity tool is quickly becoming company-wide infrastructure. |
| SR035 | G2 (via Wayback Machine) | Granola Reviews & Product Details | Reviews (19) ... 4.9 ... Small-Business (50 or fewer emp.), Enterprise (>1000 emp.), Mid-Market (51-1000 emp.) reviewers across Information Technology and Services, Financial Services, Clinical/healthcare-adjacent roles. |
| SR036 | Shadow | Granola Encrypted Its Local Database. Here's Why That Matters -- and What to Use Instead. | "Sorry to see Granola @meetgranola going closed... Any recommendations for good alternatives? What are you switching to?" 630 people liked it. 577 bookmarked it. 177 replied. |
| SR038 | Microsoft | Catch up on meetings with Microsoft 365 Copilot in Teams | Intelligent recap experiences are available as part of Teams Premium... or as part of the Microsoft 365 Copilot license. |
| SR039 | Microsoft | Microsoft Teams Premium | Teams Premium pricing: $10.00 user/month, paid yearly. |
| SR040 | Zoom | Work smarter, create faster: Zoom AI Companion drives conversations to completion | AI Companion Monthly Active Users more than tripled year over year in Q4 FY26. |
| SR041 | AI Note taking and Take Notes for Me in Google Meet with Gemini | Gemini in Meet captures real-time notes and action items, so you can stay focused on the discussion. | |
| SV001 | Granola | Granola raises $125M to put your company's context to work | We raised $125M in Series C funding led by Index Ventures. |
| SV002 | TechCrunch | Granola raises $125M, hits $1.5B valuation as it expands from meeting notetaker to enterprise AI app | Granola has raised $125 million in a Series C round at a $1.5 billion valuation. |
| SV003 | BusinessCloud | Granola, founded in 2023, becomes latest UK unicorn | Total funding raised is now approximately $188m (£141m). |
| SV004 | Granola | Granola raises $20M to build the AI notepad that makes you smarter | We raised a $20M Series A led by Spark Capital. |
| SV005 | TechCrunch | VCs love using the AI meeting notepad Granola, so they gave it $20M | Granola has around 5,000 weekly active users. |
| SV006 | Business Wire | Granola Launches AI Workspace for Teams and Raises $43M Series B | Granola raised $43M in Series B funding. |
| SV007 | TechCrunch | AI notetaking app Granola raises $43M at $250M valuation, launches collaborative features | Granola raised $43 million in a Series B round at a $250 million valuation. |
| SV008 | Index Ventures | Index Ventures Leads Granola's $125M Series C | Granola's growth has been driven by individual users becoming evangelists inside their organizations. |
| SV009 | The SaaS News | Granola Raises $125M Series C at $1.5B Valuation | Granola's total funding now stands at $192 million. |
| SV010 | Companies House (UK Government) | GRANOLA LABS LTD overview - Find and update company information | GRANOLA LABS LTD. Incorporated on 3 March 2023. |
| SV011 | Companies House (UK Government) | GRANOLA LABS LTD filing history - Find and update company information | AA Total exemption full accounts made up to 31 March 2025. |
| SV012 | U.S. Securities and Exchange Commission (EDGAR) / Zoom Communications, Inc. | Zoom Communications, Inc. Form 10-K for fiscal year ended January 31, 2026 | Gross margin increased to 77.0% for the fiscal year ended January 31, 2026. |
| SV013 | Tracxn | Granola - 2026 Company Profile & Team | Granola has raised a total funding of $192M over 4 rounds. |
| SV014 | SaaSCalcHub | 2026 SaaS Industry Benchmarks (Growth, CAC, NDR, Rule of 40) | Subscription gross margin has remained stable around 76% at the median, with top-quartile companies at 84%+. |
| SV015 | TheNextWeb | London's Granola raises $125M to turn meeting recordings into enterprise AI infrastructure | A $1.5 billion valuation for a company that was worth $250 million ten months ago invites scrutiny. |
| SV016 | BritishWire | Granola $125M Series C: Is The $1.5B AI Notetaker A Bubble Or A Bargain? | A brutal market correction feels absolutely inevitable once these highly valued startups are finally forced to show their actual revenue sheets. |
| SV017 | Otter.ai | Otter.ai Evolves from AI Notetaker to Create $100B Enterprise Conversational Knowledge Engine Market | More than 35 million people, across over one billion meetings. |
| SV018 | National Law Review / Foley & Lardner | When AI Notetakers Take the Stand -- The Legal Risks Lurking in Your Virtual Meetings | These AI notetakers are increasingly finding themselves, and their users, entangled in high-stakes litigation under federal and state wiretapping laws. |
| SV019 | Bloomberg Law | AI Notetaking Poses Wiretapping, Discovery, and Ethical Pitfalls | A vendor with data-mining rights can expose a customer company to vicarious liability for the vendor's CIPA violations. |
| SV020 | Goodwin Procter LLP | AI Transcription Tools Under Scrutiny: Navigating Privacy Risks and Practical Mitigation Strategies | Violations of the California Invasion of Privacy Act can result in organizations facing damages of up to $5,000 per violation. |
| SV021 | CompleteAI Training | Otter.ai Faces Consolidated Class Action Over Meeting Recording Practices | A federal class action in San Jose is consolidating four lawsuits against Otter.ai. |
| SV022 | BuildBetter | Do they own your data? Granola.ai Privacy Policy Reviewed. | Final Enterprise Readiness Rating: 3/10. US-only, AWS hosted; no EU region or on-prem/VPC options. |
| SV023 | Precedence Research | AI Note Taking Market Size to Hit USD 3476.74 million by 2035 | Revenue, 2025 USD 623.50 Mn; Forecast Year, 2035 USD 3476.74 Mn. |
| SV024 | Statifacts | AI Meeting Assistant Market Size to Lead USD 35.02 Billion by 2035 | Market Size in 2026 USD 4.41 Billion. |
| SV025 | Fireflies.ai | Fireflies reaches $1 billion valuation, partners with Perplexity to bring real-time web search to meetings | Fireflies.ai... today announced it has reached a valuation of over $1 billion following its first tender offer... maintains profitability since 2023. |
| SV026 | Carta | State of Private Markets: Q1 2026 | More than 60% of all venture capital raised by companies on Carta in Q1 went to AI companies... An AI foundational model startup at Series A might be raising at a $300 million median valuation. A non-AI startup at the same stage is at $55 million. |
| SV027 | Stock Analysis (S&P Global Market Intelligence data) | Zoom Communications (ZM) Market Cap & Net Worth | Zoom Video has a market cap or net worth of $25.31 billion as of June 30, 2026. |
| SV028 | Stock Analysis (S&P Global Market Intelligence data) | Zoom Communications (ZM) Revenue 2017-2026 | Zoom Video... revenue in the last twelve months to $4.93B, up 4.98% year-over-year... P/S Ratio 5.13. |
| SV029 | Finro Financial Consulting | AI Valuation Multiples Q1 2026: Investors Reprice Quality | Data intelligence, marketing tech, and productivity tools compressed in public multiples... the most notable private increases came in marketing tech and productivity tools, even as public comps in those areas did not move the same way. |
| SV030 | CNBC | The global M&A boom is rolling into 2026 as AI sparks deal frenzy — but cash is getting tight | The total value of deal-making activity surged nearly 40% to a record of $4.9 trillion in 2025... as central banks cut interest rates, valuations improved and companies increased spending on artificial intelligence. |
| SV031 | Forbes | VCs' Favorite Note-Taking App Granola In Talks To Hit $1 Billion Valuation | Now it's in talks with investors to raise at a valuation of over $1 billion... Otter AI... announced last March that it was bringing in over $100 million in annual recurring revenue... Notion, which was last valued at $11 billion in an employee share sale last month. |
| SV032 | TechCrunch | Salesforce acquires AI customer service platform Fin for $3.6B | Salesforce announced on Monday that it will acquire AI customer service platform Fin for $3.6 billion. |
| SV033 | Crunchbase News | Crunchbase Predicts: 15 Companies That Could Go Public In 2026 As The IPO Market Gains Momentum | Databricks... growing more than 55% year over year, with an over $4.8 billion revenue run rate... Net retention was above 140%... Canva... annualized revenue has reached $3.3 billion. |