EvenUp
PI 法律 AI 赛道的品类龙头,客户证据扎实,但软件加服务的经济性仍不透明。
EvenUp 看起来是原告侧法律 AI 的品类龙头,但缺少经审计 ARR、毛利率和留存数据,>$2B 估值很难支撑。
封面要素
公司概况
EvenUp 是一家总部位于 San Francisco 的法律 AI 公司,2019 年由 Rami Karabibar、Raymond Mieszaniec 和 Saam Mashhad 创立。公司向原告方人身伤害律所销售 Claims Intelligence Platform,借助 Piai 模型和人工复核流程,自动化处理医疗记录分析、索赔函起草、案件监控、和解基准比较和托管式诉前运营。公司已从软件扩展到 PLAAS,该订阅服务接触度更高,会把 EvenUp 法律运营人员嵌入律所工作流。公开证据显示客户采用强劲、后期融资到位,但经济性披露有限。
- 创始人
- Rami Karabibar, Raymond Mieszaniec, Saam Mashhad
- 创立地点
- San Francisco, CA
- 总部
- San Francisco, CA
- 产品
- Claims Intelligence Platform 以 Piai AI 模型为核心,覆盖索赔函生成、医疗时间线、案件陪伴工作流、和解分析、沟通智能体,以及 PLAAS 托管式诉前服务。
- 客户
- 美国原告方人身伤害律所,尤其在高案量 PI 业务中渗透更深。
- 商业模式
- Claims Intelligence Platform 按案件计价的软件收入,加上 PLAAS 及配套诉前支持带来的订阅式托管服务收入。
- 阶段
- Series E
- 融资情况
- 2025 年 10 月完成 $150M Series E 轮,估值超过 $2B;已披露累计融资 $385M。
执行摘要
主要优势
- 在原告人身伤害 AI 领域具备品类领导地位,覆盖 2,000+ 家律所,约占前 100 大 PI 律所的 30%。
- 产品深度覆盖索赔需求函起草、病历时间线、工作流自动化、和解分析和托管服务。
- 自有 PI 数据和人工复核闭环,使其垂直工作流比通用 LLM 包装层更可防守。
- $150M Series E 和累计披露融资 $385M,说明投资人支持力度强。
主要风险
- 未公开 ARR、毛利率、烧钱速度或 NRR,>$2B 估值难以对标。
- PLAAS 可能从结构上压缩利润率,把 EvenUp 从纯 SaaS 变成软件加服务运营商。
- HIPAA、律师伦理和 AI 质量一旦出错,可能带来法律、声誉和客户留存损伤。
- 对 CMS / 平台的依赖,以及 Supio、Filevine、Clio、Harvey 和通用 LLM 带来的 AI 商品化,可能压窄差异化。
未决问题
- 核心软件业务和 PLAAS 的经审计 ARR、收入增长和毛利率披露。
- PLAAS 单位经济,包括每份合同收入、人员配置比例和混合毛利率。
- 客户留存、流失和扩张指标,包括按队列拆分的 NRR / GRR。
- 治理和投资人权利细节,包括清算优先权,以及 RELX / REV 的任何战略权利。
- 2025 年 10 月 Series E 后的准确员工数、现金余额和现金跑道。
目录
01公司概况
1.1 身份、总部与使命
EvenUp 于 2019 年末在 California 州 San Francisco 创立,三位联合创始人的法律、运营和技术背景互补。公司公开表述的使命,是为人身伤害索赔人拉平竞争条件——创业动机来自 CEO Rami Karabibar 的个人经历:一位家庭成员处理 PI 案件时,他亲眼看到资源受限的律所很难组织足够证据和文档来争取公平和解。EvenUp 的平台处在法律科技与 AI 驱动工作流的交汇处,为 PI 律所提供端到端工具,覆盖案件 intake、医疗记录分析、损害金额量化和索赔函生成。公司注册并总部位于 San Francisco,采用灵活的混合办公模式。EvenUp 所在的广义法律 AI 与法律科技市场,在 2023–2025 年生成式 AI 浪潮后吸引了大量机构资本。EvenUp 持有 SOC 2 Type II 和 HIPAA 合规认证,使其能在以客户数据隐私为核心要求的律所环境中部署。公司仍为私营,未披露年度收入;它是一家高增长、后期风投支持的创业公司,估值已达到 $2B+ 的独角兽以上水平。[CO001, CO002, CO025, CO015, CO010, CO026]
EvenUp 平台如何通过 AI 驱动的产品层,把案件接收连接到和解结果。
[CO012, CO019, CO027, CO033, CO034]1.2 创始人与领导团队
EvenUp 由三人共同创立,截至本报告日期,三人仍担任高管。CEO Rami Karabibar 推动公司战略愿景,也是主要对外发言人;他的创业动机来自目睹一位家庭成员在法律资源不足的情况下处理人身伤害索赔。COO Raymond(Ray)Mieszaniec 具备运营和创业经验,此前共同创办金融科技创业公司 EquitySim,并在 PwC 香港风险咨询业务工作。第三位联合创始人 Saam Mashhad 的头衔少见,同时负责产品和法律运营,体现了公司的判断:在这个市场,产品决策离不开法律运营专业能力。截至 2025 年末,EvenUp 投资人页面和高管访谈均确认三位联合创始人身份。EvenUp 还保有一支 300+ 人法律运营团队,支撑其 human-in-the-loop 复核流程,确保 AI 输出达到原告律师要求的证据标准。研究期间未发现重大领导层离职。[CO002, CO015, CO016, CO017, CO019]
| 姓名 | 职务 | 过往经历 | 创始人 | 关键人物备注 |
|---|---|---|---|---|
| Rami Karabibar | CEO 兼联合创始人 | 有 PI 诉讼亲身经历;2019 年参与创立 EvenUp | 是 | 主要对外发言人;对战略至关重要 |
| Raymond (Ray) Mieszaniec | COO 兼联合创始人 | 联合创立 EquitySim(金融科技);曾在香港 PwC 做风险咨询 | 是 | 负责运营;截至 Nov 2025,媒体访谈已确认 |
| Saam Mashhad | 产品与法律运营负责人兼联合创始人 | 法律运营背景;兼具产品和法律双重角色 | 是 | 罕见的产品-法律混合画像,在法律科技中稀缺 |
截至报告日期,三位联合创始人都仍担任活跃高管职位。未发现重大领导层变动。董事会构成未公开披露。
[CO002, CO015, CO016, CO017, CO019]1.3 产品、技术与平台
EvenUp 的旗舰产品是 Claims Intelligence Platform,一套 AI 驱动工具,面向人身伤害律所自动化诉前工作流。核心是 Piai™,这是 EvenUp 用数十万起 PI 案件训练的自研 AI 模型,驱动自动化医疗记录分析、损害金额量化和索赔函起草。截至 2025 年 5 月,公司推出 AI Drafts Suite,把 Smart Workflows、Medical Bill Summary 和 Express Demands 打包在按案件计价模式下。其他产品包括 Mirror Mode(让输出适配律所风格)、Case Companion(案内 AI 助手)、Voice Agent(用于客户 intake)和 AI Playbooks(标准化律所级自动化)。2025 年 12 月,EvenUp 增加 Medical Management,这一工具识别进行中 PI 案件的治疗缺口,帮助律师优化客户赔付。最新产品是 2026 年 5 月宣布的 PLAAS(Pre-Litigation as a Service);在该模式下,EvenUp 实际上成为外包法律运营伙伴,而不只是软件供应商。PLAAS 发布时预售订阅达到 $10M,表明市场对更高接触度服务模式有强需求。平台在整个数据管道中执行 SOC 2 Type II 和 HIPAA 控制。[CO012, CO027, CO022, CO023, CO032, CO033]
截至 2026 年 5 月 EvenUp 的关键运营和牵引力指标,突出吞吐量、渗透率和累计规模。
前 100 大律所渗透率和每周案件量来自 LawnNext 报道(May 2026);所有数字均受公司披露限制约束。
[CO003, CO004, CO008, CO009, CO013, CO005]1.4 客户规模与市场位置
截至 2026 年 5 月,EvenUp 服务超过 2,000 家人身伤害律所,包括美国前 100 大 PI 律所中的约 30%;该比例高于 2025 年 5 月 AI Drafts Suite 发布时披露的 20%,当时整体客户为 1,500+ 家律所。公司累计解决超过 200,000 起 PI 案件,处理原告损害赔偿金额超过 $14 billion。PLAAS 发布时,EvenUp 报告每周处理约 10,000 起案件。考虑到前 100 大 PI 律所占美国人身伤害和解量的份额不成比例地高,高端市场渗透尤其重要;EvenUp 在高端客户中的渗透率降低了被竞争替代的风险。公司的规模指标也形成护城河:Piai™ 模型随每个新增案件继续改善,EvenUp 的 300+ 人法律运营团队提供了纯软件竞争者缺少的人工复核兜底。2025 年入选 Forbes Cloud 100,为 EvenUp 作为头部云软件公司的位置提供了第三方验证。公司的按案件计价模式把激励直接与律所吞吐量绑定,相比固定费用企业 SaaS 替代方案,摩擦更小。[CO003, CO004, CO008, CO009, CO013, CO011]
| 指标 | 数值 | 截至 | 置信度 |
|---|---|---|---|
| 累计融资 | $385M | Oct 2025 | 高 |
| 最新估值 | >$2B | Oct 2025 | 高 |
| 活跃 PI 律所客户 | 2,000+ | May 2026 | 高 |
| 前 100 大 PI 律所渗透率 | ~30% | May 2026 | 中 |
| 已解决案件(累计) | 200,000+ | Nov 2025 | 高 |
| 已处理损害赔偿总额 | $14B+ | May 2026 | 中 |
| 每周案件量 | ~10,000 | May 2026 | 中 |
| 成立年份 | 2019 | 历史 | 高 |
| 总部 | 加州旧金山 | 当前 | 高 |
| 已披露收入 / ARR | 2026 | 不可得 |
数值反映最近确认的数据点;带 “~” 前缀的行是二级来源中的近似数字。已处理损害赔偿和已解决案件为生命周期累计总数。收入 / ARR 未公开披露。
[CO001, CO003, CO004, CO005, CO006, CO008]1.5 融资历史与投资方格局
据 Bloomberg Law,截至 2025 年 10 月,EvenUp 已披露累计融资 $385 million。最近一轮是 2025 年 10 月的 $150 million Series E,由 Bessemer Venture Partners 领投,将公司估值推高到 $2 billion 以上。Series E 的共同投资方包括 REV Ventures(RELX/LexisNexis 的风险投资部门)、B Capital Group 和 Bain Capital Ventures。此前的 Series D 于 2024 年 6 月公开报道,随后在 2024 年 10 月通过公司渠道公告;该轮融资 $135 million,估值约 $1 billion,Bain Capital Ventures 被列为领投方,Bessemer 参投。更早轮次包括 Lightspeed Venture Partners、SignalFire(领投种子轮)、NFX 和 DCM Ventures 参投。Bessemer 在后期轮次中重复参与值得注意,因为该机构深耕垂直 SaaS 和法律 AI。RELX/LexisNexis 通过 REV 参投,提示潜在的战略数据和分发协同。公司未披露 ARR、收入增长或烧钱速度,财务尽调因此只能依赖融资轮披露。$385M 的累计融资对一家 Series E 法律科技公司来说异常庞大,可能意味着 PLAAS 外包模式资本强度较高,也可能是在为潜在 IPO 路径显著延长现金跑道。[CO005, CO006, CO007, CO018, CO028, CO029]
| 投资者 / 利益相关方 | 投资角色 | 轮次 | 战略意义 |
|---|---|---|---|
| Bessemer Venture Partners | 主要参与方 / 后期领投方 | Series D 轮(2024)、Series E 轮领投(2025) | 重复押注的后期投资者,拥有深厚垂直 SaaS 经验;最强信心信号来自其领投 Series E |
| Bain Capital Ventures (BCV) | 领投方 / 重复投资者 | Series D 轮领投(2024)、Series E 轮 | 成长期专家;领投推动其成为独角兽的 Series D,并继续参与 Series E |
| Lightspeed Venture Partners | 共同投资者 | Series D 轮之前 | 一线 VC;EvenUp AI Drafts 博客(May 2025)已确认 |
| SignalFire | 种子轮领投方 | 种子轮 | 首个机构投资者;有从种子到独角兽的记录;投资组合页面已确认 |
| NFX | 共同投资者 | Series D 轮之前 | 专注网络效应的基金;EvenUp 博客已确认 |
| DCM Ventures | 共同投资者 | Series D 轮之前 | 美亚跨境基金;EvenUp 博客已确认 |
| REV Ventures (RELX / LexisNexis) | 战略共同投资者 | Series E 轮(2025) | 法律数据 / 研究巨头进入 AI 法律科技;意味着潜在数据和分发协同 |
| B Capital Group | 共同投资者 | Series E 轮(2025) | 全球成长型基金;LawnNext Medical Management 文章(Dec 2025)已确认 |
Series D 之前投资者的轮次参与由 EvenUp 公开博客(May 2025)推断;Lightspeed、NFX、DCM 的确切轮次分配未被独立确认。董事会席位和持股比例未披露。
[CO005, CO006, CO007, CO018, CO028, CO029]| 日期 | 事件 | 类型 | 金额 / 状态 | 关键参与方 | 含义 |
|---|---|---|---|---|---|
| 2019-Q4 | 公司在加州旧金山成立 | 创立 | 创始人:Rami Karabibar、Raymond Mieszaniec、Saam Mashhad | 建立首个 AI 原生 PI 法律运营平台 | |
| 2020 | 种子轮完成;SignalFire 领投 | 融资 | 未披露 | SignalFire(领投) | 用初始资本搭建 MVP,并招聘早期法律运营团队 |
| 2022 | 核心 AI 索赔函产品上线;获得早期 PI 律所客户 | 产品 | EvenUp | 在 PI 诉前工作流中验证产品市场契合 | |
| 2024-06 / 2024-10 | Series D 融资先公开报道,随后正式宣布 | 融资 | $135M,估值约 $1B | Bain Capital Ventures(领投)、Bessemer Venture Partners 和其他投资者 | 触及独角兽估值门槛;进入规模加速阶段 |
| 2025-01 | Express Demands AI 文档产品发布 | 产品 | EvenUp | AI Documents 系列首个产品;大规模自动生成索赔函 | |
| 2025-05 | AI Drafts Suite 发布,采用按案定价模式 | 产品 | 发布时 1,500+ 家 PI 律所活跃 | EvenUp | 按案模式取代订阅;发布时已处理 $7B 损害赔偿 |
| 2025-10 | Series E 融资完成;入选 Forbes Cloud 100 | 融资 | $150M,估值 >$2B | Bessemer(领投)、REV/RELX/LexisNexis、B Capital、Bain Capital Ventures | 累计融资达到 $385M;战略法律数据投资者进入股权结构表 |
| 2025-11 | 累计解决 200,000+ 个 PI 案件 | 规模 | 200,000+ 个案件;2,000+ 家律所 | EvenUp | 吞吐量里程碑确认运营可扩展性 |
| 2025-12 | Medical Management 产品发布 | 产品 | EvenUp | 治疗缺口识别把平台扩展到完整案件生命周期 | |
| 2026-05 | PLAAS(Pre-Litigation as a Service)发布 | 产品 | 已预售 $10M 订阅 | EvenUp | 商业模式从软件扩展到外包法律运营 |
2022 年产品里程碑由融资轨迹和早期客户引用推断;确切日期未确认。早期轮次细节(种子轮金额、是否有 Series A/B)未公开披露。里程碑表遵循创立、融资、产品和规模事件类型要求。
[CO001, CO005, CO006, CO007, CO008, CO009]按时间梳理 EvenUp 从 2019 年创立到 2026 年 5 月 PLAAS 发布期间的关键融资、产品和规模里程碑。
2022 年「AI Demand Letters 上线」为估计日期,由融资轨迹和早期客户引用推导;其他日期均来自一手公告。
[CO001, CO006, CO007, CO014, CO015, CO020]1.6 图表
02市场分析
2.1 市场边界与范围
EvenUp 的可服务市场,是面向原告方人身伤害法律运营的软件与 AI;它属于更广义的美国法律科技垂直赛道,由三点界定:(1)买方是按胜诉酬金收费的原告方 PI 律所,而不是按小时计费的公司法务或辩护业务;(2)核心工作流是人身伤害索赔中的索赔函准备、医疗记录分析、案件估值、intake 分诊和治疗跟踪;(3)经济回报用和解金额提升和律师助理工时节省衡量,而不是按计费小时效率衡量。人身伤害法涵盖机动车事故、工伤、医疗事故、产品责任和有毒侵权等案件——受伤原告都按赔付比例聘请律师。 该边界明确排除:(a)保险公司和辩护律师使用的辩护方法律 AI;(b)合同生命周期管理和 eDiscovery 等公司法务部门工具;(c)销售给保险承运人的 InsurTech 理赔处理平台;以及(d)没有针对 PI 案件类型校准的通用 AI 编码助手或广义研究工具。相邻市场包括 Filevine 和 Clio 等法律案件管理系统(CMS)、医疗记录调取服务和诉讼融资——EvenUp 会与它们对接,但并不直接竞争。 EvenUp 目前替代的现状方案包括:人工律师助理起草索赔函(通常每封 5–20 小时)、通用文字处理模板、只返回结构化摘要的医疗时间线供应商,以及缺少 PI 专属训练数据和 ABA Rules 与 HIPAA 要求的保密控制的现成大语言模型(Claude、ChatGPT)。 [CM001, CM002, CM003, CM004, CM019]
| 细分 / 类别 | 纳入支出 | 排除支出 | 买方 / 付款方 | 与 EvenUp 的相关性 |
|---|---|---|---|---|
| PI 原告法律运营 AI | 面向 PI 原告律所的索赔准备、案件估值、医疗时间线、进件 AI、沟通智能体和工作流自动化 | 被告侧法律 AI;保险理赔平台 | PI 律所所有者(买方)、案件经理、索赔函撰写人(用户) | 核心市场——EvenUp 的主要收入细分 |
| 更广义法律 AI / 法律科技 | 电子取证、合同审查、企业法律研究、法律聊天机器人 | PI 特定案件准备和和解工作流 | BigLaw 律师、企业总法律顾问(GC)、政府法律顾问 | 相邻市场——仅用于上限 TAM 参考;买方和 ROI 模型不同 |
| 保险科技 / 理赔处理 | 保险公司理赔自动化、AI 准备金设定、欺诈检测 | 原告侧索赔准备和案件策略 | 保险公司理赔部门 | 相邻市场——EvenUp 间接受益于保险公司,但不向其销售 |
| 案件管理系统(CMS) | Filevine、MyCase、Clio——案件 / 联系人 / 账单管理 | AI 驱动的文件生成和医疗分析 | PI 律所管理员和律师 | 集成伙伴——EvenUp 架在 CMS 之上,不是替代品 |
| 现状替代方案 | 律师助理手工起草、通用 LLM 提示、病历供应商摘要 | 不转化就不是 EvenUp 可触达支出 | 同一批 PI 律所和案件经理 | 直接替代目标——EvenUp 替代每封索赔函所需数小时手工作业 |
边界定义来自 EvenUp 产品描述、ABA 职业行为规则和分析师市场报告范围。单个细分的支出估计不可得。
[CM001, CM002, CM003, CM004]2.2 市场规模测算视角
没有分析机构发布过独立的 PI 原告法律科技 AI 市场规模估计。广义法律 AI 市场可提供两个自上而下视角,且差异明显:Grand View Research 估计 2024 年全球法律 AI 市场为 $1.45 billion,并预计以 17.3% CAGR 增至 2030 年 $3.90 billion,其口径锚定 eDiscovery、法律研究和合同审查。MarketsandMarkets 报告 2025 年法律 AI 软件市场为 $3.11 billion,并以 28.3% CAGR 增至 2030 年 $10.82 billion——口径明显更宽,纳入生成式 AI 智能体和行业专属起草工具。两项估计之间的差距反映定义差异,而非实证分歧;两者都应被视为上限参考点,而不是 PI 专属市场规模。 美国 PI 子市场的自下而上估计,可以从规模指标开始:BLS 报告美国 2024 年有 864,800 名执业律师,其中一部分专做人身伤害原告业务。EvenUp 自身数据暗示活跃 PI 原告律师约 30,000–60,000 名(基于“数千家律所”和“服务 2,000+ 家律所”)。若粗略假设 AI-first 平台可触达 30,000 家 PI 律所,ARPU 约 $5,000–$20,000 / 年,则当前价格点下的美国 PI AI SAM 为 $150M–$600M。EvenUp 2026 年 PLAAS 发布并产生 $10M+ 早期订阅,提示托管服务产品在该区间高端具备定价牵引力。上述估计不确定性很高,应视为示意而非定论。缺少已发布的 PI 专属市场规模本身就是重大证据缺口,本报告将其保留为尽调问题,而不是已解决数字。 [CM005, CM006, CM007, CM008, CM009, CM010]
| 发布方 | 年份 | 地域 | 市场规模 | CAGR | 方法 / 范围 | 置信度 | PI 测算局限 |
|---|---|---|---|---|---|---|---|
| Grand View Research | 2024 | 全球 | $1.45B (2024) | 17.3% (2025–2030) | 电子取证、法律研究、合同审查、合规 AI | 中 | 未单独拆出 PI 原告子垂直;美国占比约 46% |
| MarketsandMarkets | 2025 | 全球 | $3.11B (2025) | 28.3% (2025–2030) | 法律 AI 软件,包括生成式智能体、起草工具 | 中 | 范围比 GVR 更宽;包含尚未进入市场的 GenAI 智能体 |
| IBISWorld(PI 律师) | 2025 | 美国 | 未公开披露(行业分散) | N/A | 人身伤害律师子行业收入 | 低 | 律所自身收入,不是法律科技软件支出 |
| EvenUp 自下而上估算(估计) | 2026 | 美国 | $150M–$600M SAM(估计) | 估计近期 20–35% | 30,000–60,000 家可触达 PI 律所 × $5K–$20K ARPU | 低 | 自有估算;PI 律所数量和 ARPU 假设未经验证 |
| EvenUp 牵引力指标 | 2026 | 美国 | 每周 10,000+ 个案件;$14B+ 损害赔偿;$10M+ PLAAS 早期订阅 | N/A | 公司披露的平台活动指标 | 中低 | 公司声称;未经独立审计;可作为 SAM 下限参考 |
所有市场规模值都针对软件 / AI 服务,而不是律所收入。没有独立分析师发布 PI 特定市场规模;自下而上估计是基于 EvenUp 公开数据和 BLS 律师统计的分析师近似值。
[CM005, CM006, CM007, CM008, CM011, CM013]从 TAM(全球法律 AI)收窄到 SAM(美国 PI AI 市场)和 SOM(EvenUp 可触达机会),附 2025/2026 年近似值。
SAM 和 SOM 是自下而上的示意性估算;没有独立分析师专门跟踪 PI AI。TAM 来自 MarketsandMarkets 2025 报告(已发布范围中最宽)。SOM 由公开数据推断,并非公司披露 ARR。
[CM007, CM036, CM039]基于律所数量和 ARPU 假设,给出美国 PI 原告法律 AI 可服务市场的低、中、高情景。
所有数字均为 USD 百万美元。律所数量和 ARPU 区间来自 EvenUp 披露的牵引力数据、BLS 律师统计和 IBISWorld 行业分散度说明。没有独立 PI AI 市场规模测算;区间反映结构性假设,不是模型预测。
[CM005, CM007, CM011, CM035]2.3 买方、用户与付款方细分
PI 律所是采购单位,但购买权和日常使用分散在不同角色中。律所所有者和管理合伙人是经济买方,负责签约并控制软件预算;对胜诉酬金律所而言,决策算式很直接:这个工具能否改善和解结果,或把单案耗时降到足以覆盖订阅成本?案件经理和律师助理是主要日常用户,处理 intake 文件、协调医疗记录并发送客户沟通——这是 EvenUp 自动化的最高频工作流。索赔函撰写者(专职律师和资深律师助理)用 AI 起草工具准备和复核索赔包,这是 PI 案件中价值最高的文档类型。Intake 团队和法律运营经理用 AI 做案件分诊,识别哪些新进文件价值高或接近诉讼时效风险。 EvenUp 扩展到 Pre-Litigation as a Service(PLAAS)后,增加了一种新买方关系:律所所有者想要外包案件管理伙伴,而不是软件许可证,付款模式也从按席位 SaaS 转为按案件托管服务合同。PLAAS 模式显著扩大了可触达市场,覆盖缺少内部能力独立落地软件工具的律所。 PI 律所绝大多数很小:大部分执业团队少于 10 名律师。大型全国律所(例如 Lerner & Rowe、Morgan & Morgan)在律所数量中占比不高,却贡献了不成比例的年度案件量,也是 EvenUp 最显眼的标杆客户。截至 2026 年 5 月,EvenUp 已渗透前 100 大 PI 律所的 30%,说明其已成功上探高端市场;但大量小型长尾律所仍构成未开发 SAM 的主体。 [CM016, CM017, CM018, CM019, CM020, CM021]
| 细分 | 律所内角色 | 主要用户 | 预算负责人 | 核心工作流 | 采用触发点 | EvenUp 产品适配 |
|---|---|---|---|---|---|---|
| 律所所有者 / 单独执业律师 | 决策者;通常亲自处理案件 | 是(小型律所) | 是 | 和解策略、最终索赔函审阅 | 单案 ROI;和解改善数据 | Companion AI、索赔函起草、PLAAS |
| 管理合伙人 | 监督助理律师工作;掌控预算 | 部分 | 是 | 案件量监督、绩效分析 | 律所层面的效率提升和收入增长 | Firmwide Knowledge Base、分析仪表盘 |
| 案件经理 / 律师助理 | 日常案卷管理;工作流量最高 | 是(主要用户) | 否 | 医疗记录、客户跟进、文件归档 | 节省工时;治疗缺口可见性 | Medical Management、Communication Agents、进件 AI |
| 索赔函撰写人(专职律师 / 高级律师助理) | 起草并审阅索赔材料包 | 是 | 否 | 索赔函准备、医疗时间线审阅 | 草稿质量、速度、准确性 | 产品:AI Drafts、Express Demands、Mirror Mode |
| 进件 / 法律运营团队 | 筛选新案;分派给律师 | 是 | 否 | 案件分流、保险覆盖核验、诉讼时效检查 | 处理量上限;高价值案件识别 | Intake AI、Companion AI、PLAAS 索赔设置 |
在小型律所(1–10 名律师)中,细分角色会重叠,律所所有者往往包办所有职能。大型全国性律所有清晰的人员层级。 EvenUp 产品匹配度基于公开描述的能力判断。
[CM016, CM017, CM018, CM019, CM020, CM021]PI 律所 AI 采用旅程中,决策者、主要用户和结果付款方之间的关系。
该流程图仅示意角色层级;具体组织结构会随律所规模变化。大型律所有独立分工层级,独立执业者则把所有角色合并到一人。
[CM021, CM022]2.4 增长驱动因素
最具结构性的驱动因素,是胜诉酬金模式本身:PI 律师只有案件解决后才收费,并按和解金额抽成,因此每省下一小时律师或律师助理的行政时间,都会直接改善律所盈利能力。按小时计费律所采用 AI 后,节省时间会部分减少收入;PI 律所则把 AI 效率收益全部转化为利润率,或转化为承接更多案件的能力。EvenUp 客户证言持续印证这种经济性:一家 Houston 律所报告和解改善 300%,另一家报告节省 3,500 小时,索赔效率提升 34%。 第二个驱动因素是医疗记录量:据 EvenUp 平台分析,43% 的 PI 案件出现超过 30 天、会削弱索赔价值的治疗缺口,16.8% 在前三个月内形成 30 天缺口。人工管理这些缺口很费人力;AI 辅助监控能直接转化为被保住的案件价值。保险公司在理赔处理上也变得更成熟,要求更完备的索赔材料来准确设定准备金,从而创造了对更高质量、数据支撑索赔包的需求。 法律 AI 投资环境也支撑市场增长。EvenUp 2025 年 10 月 Series E($150M,$2B+ 估值,由 Bessemer 和 RELX/REV 领投)以及更广泛的法律科技投资趋势,都确认资本仍可用于资助 go-to-market 扩张。2020 年以来法律支持岗位劳动力短缺加剧,也加快了自动化需求。 [CM023, CM024, CM025, CM026, CM027, CM028]
| 因素 | 方向 | 时间节奏 | 对 EvenUp 的影响 | 尽调问题 |
|---|---|---|---|---|
| 风险代理收费模式 ROI | 顺风 | 即时 | 每省 1 小时都直接变成利润率;相比按小时收费法律服务,这是 PI 领域最强的结构性采用驱动因素 | 验证 AI 每案节省更多工时后,ARPU 弹性如何变化 |
| 法律支持岗位劳动力短缺 | 顺风 | 当前且加剧 | PI 律所在人员压力下会加快采用自动化 | 跟踪 PI 领域律师助理招聘率和薪资通胀 |
| 医疗记录体量与复杂度 | 顺风 | 多年趋势 | 每案记录越多,AI 杠杆越大;43% 的治疗缺口率构成价值论证 | 用独立医疗账单来源核验治疗缺口数据 |
| 保险公司准备金设定更精细 | 顺风 | 当前 | 保险公司需要数据支撑的索赔函来设定准备金;这会改善 EvenUp 产品接受度 | 评估保险公司侧工具(竞争产品)是否削弱 EvenUp 的索赔函质量优势 |
| ABA 胜任能力与保密规则 | 逆风 | 长期 | 律师必须监督 AI;合规开销可能拖慢 SMB 采用 | 跟踪主要 PI 州份州律师协会 AI 指引是否更新出更严格立场 |
| AI 幻觉与执业过失风险 | 逆风 | 当前 | 索赔函引用错误会让律师暴露在风险下;行级引用只能部分缓释 | 获取 EvenUp 索赔函质量相对人工审阅的独立准确率基准 |
| 小型律所预算与集成约束 | 逆风 | 近期 | 多数 PI 律所技术预算有限;与 CMS 集成的工作量会形成采用摩擦 | 跟踪 10 名以下律师客户群的流失率和 NPS |
方向判断截至 2026 年,基于公开来源。“即时”表示已经影响采购决策的因素;“多年趋势”表示结构性趋势。 逆风因素不会阻断增长,但会限制速度和市场渗透率上限。
[CM023, CM024, CM026, CM027, CM028, CM029]呈现 PI 律所要全面采用 AI 法律运营平台必须经过的阶段,以及每一阶段的关键摩擦点。
阶段数量仅为示意。EvenUp 的 2,000+ 客户数来自公司截至 2025 年中披露。美国 PI 律所总量按 BLS 和 IBISWorld 估计;目前没有面向 PI 律所的官方普查。
[CM029, CM031]2.5 采用约束与反向证据
PI 法律中的法律 AI 采用面临一组独特的结构性约束,并非所有软件类别都同样适用。ABA Model Rule 1.1(Competence)要求律师对其在客户代理中使用的技术保持足够理解,ABA 关于技术胜任力的注释(Comment 8)把该要求延伸到 AI 工具——也就是说,律师不能在缺少充分监督的情况下把判断简单委托给 AI 平台。ABA Model Rule 1.6(Confidentiality)禁止未经同意披露客户信息,要求任何处理案件数据,尤其是医疗记录的 AI 平台,在严格的数据隔离和合同保密保证下运营。ABA Rule 5.5 还禁止协助未经授权执业,引发关于 AI 生成且未由持牌律师审查批准的法律策略的疑问。EvenUp 通过取得 SOC 2 Type 2 再认证和 HIPAA 证明来回应这些合规义务,但维持合规是一项持续运营成本。 AI 幻觉是法律场景中具体且已有充分记录的风险。索赔函中若嵌入看似自信但错误的引用或伪造判例,可能让律师面临执业过失索赔、纪律处分或制裁。EvenUp 的逐行引用模型试图通过把每个主张绑定到源文件来缓解风险,但风险没有消失,也无法完全靠自动化消掉。Above the Law 对 EvenUp 的评论明确指出,算法偏见、数据隐私以及 AI 生成索赔函可能用标准化文件“堵住”法院系统的风险仍然存在。一个可能的反向情景是:保险理算员开始压低可识别 AI 生成模板的索赔函估值,随着时间推移削弱 EvenUp 提升和解金额的价值主张。 小律所预算约束是最后一个结构性摩擦:大多数 PI 律所少于 10 名律师,技术预算有限。与既有案件管理系统(Filevine、MyCase、Clio)集成需要实施投入和员工培训,可能劝退最小型律所。 [CM029, CM030, CM031, CM032, CM033, CM034]
2.6 图表
03竞争格局
3.1 竞争格局概览
EvenUp 所处的竞争环境横跨五类玩家,每一类都带来不同类型的替代威胁。PI 法律运营 AI 市场仍处早期,并在快速扩张:截至 2026 年,专为 PI 构建的 AI 平台只有少数几家,但更广义的法律 AI 赛道拥挤且融资充足,提高了横向或相邻玩家延伸进 EvenUp 垂直市场的风险。 第一类——直接 PI 专业 AI 同行——是最接近的竞争威胁,目前主要是 Supio。该平台总部位于 Seattle / San Francisco、成立于 2021 年,和 EvenUp 一样聚焦原告方 PI 律所,覆盖索赔函起草、医疗时间线、从 intake 到和解的工作流以及群体侵权。截至 2026 年 5 月,尚未发现其他融资独立、规模相近的 PI 专属 AI 公司;Darrow AI 做的是诉讼上游的法律风险敞口管理,并不是直接的索赔函工作流竞争者。 第二类——工作流平台既有厂商——包括 PI 律所已经使用的案件管理系统(CMS):Filevine(企业 / 中端市场)、Litify(原生基于 Salesforce,企业级 PI)、Clio(SMB 和中端市场)、CasePeer(PI 专业 SMB CMS)和 SmartAdvocate(PI / 工伤专业,中端市场)。上述公司都在 2024–2026 年推出或扩展 AI 功能,包括 AI medchrons、索赔函起草助手和智能体 intake 工作流。它们的结构性优势是深度数据访问和日常工作流所有权;劣势在于 AI 只是功能追加,而不是核心产品,且没有一家使用 EvenUp 这种规模的 PI 专属数据集训练。 第三类——横向法律 AI 平台——包括 Harvey(BigLaw / 公司法律顾问)、CoCounsel(Thomson Reuters,诉讼和研究)以及带 Protégé 的 Lexis+(LexisNexis,研究和起草)。它们资金充足、技术成熟,但并未针对 PI 原告工作流、胜诉酬金经济性或医疗记录分析校准。 第四类——通用 LLM——代表 DIY 替代方案:律师用 ChatGPT、Claude 或类似工具起草索赔函,但没有 PI 专属训练。EvenUp 在产品定位中明确回应通用 LLM 威胁:通用 LLM 缺少 PI 案件数据、HIPAA 合规数据隔离,以及逐行引用可追溯性。 第五类——人工现状替代方案——仍是被替代收入最大的“竞争者”:内部索赔函撰写者、律师助理和离岸法律支持供应商,都是 EvenUp 替换的工作流;人工起草每封索赔函需要 5–20 小时。 [CP001, CP002, CP003, CP004, CP005]
| 竞争对手 | 类别 | 规模 / 融资 | 目标客群 | 核心差异化 | 相对 EvenUp 的关键短板 |
|---|---|---|---|---|---|
| EvenUp | PI 垂直 AI(规模领先者) | $385M 融资;$2B+ 估值;2,000+ 家律所;2025 年 10 月 Series E | 美国各规模 PI 原告律所 | 250K+ 份判决数据集;全生命周期智能体;Top 100 PI 律所渗透率 30% | 平台依赖 CMS 集成;索赔函商品化风险 |
| Supio | PI 垂直 AI(挑战者) | 2021 年创立;未披露公开融资;数百家律所 | 美国 PI 原告与大规模侵权律所 | 原生智能体平台;Westlaw Advantage 集成(2026 年 4 月);不限量定价模式 | 规模和数据语料小于 EvenUp;品牌认知更晚近 |
| Filevine | 工作流平台既有玩家(企业级) | 私营;覆盖广泛业务领域;企业级 | 中端与企业级律所;所有业务领域,包括 PI | LOIS 智能体 AI;医疗时间线生成(2025 年 4 月);深入访问 CMS 数据 | AI 未用 PI 判决语料训练;仅定制定价;通用型 |
| Litify | 工作流平台既有玩家(企业级 PI) | Salesforce 原生;企业级;未披露公开融资 | 企业级 PI 与原告律所 | ACE 智能体进件 / 索赔函;Salesforce 集成;和解情景建模 | SMB 实施成本高;PI AI 只是附加功能,不是核心产品 |
| Clio | 工作流平台既有玩家(SMB / 中端) | ~$900M 融资;150,000+ 名法律专业人士(公司声称) | SMB 律所、所有业务领域;CasePeer 面向 PI SMB | Clio Work AI 于 2025 年发布;250+ 个集成;SMB 分销最广 | 跨业务领域 AI;未用 PI 专属判决 / 和解数据训练 |
| CasePeer | 工作流平台既有玩家(PI SMB) | 被 Clio 收购;PI 专业 CMS;聚焦 SMB | 需要交钥匙 CMS 的小型 / 中型 PI 律所 | PI 专用 CMS;易用;可接入 Clio AI 路线图 | 独立 AI 深度有限;AI 跟随 Clio 路线图,而不是 PI 专业训练 |
| SmartAdvocate | 工作流平台既有玩家(PI / 工伤赔偿中端) | 私营;专攻 PI 与工伤赔偿 | 中端 PI 与工伤赔偿律所 | 内置 SmartIntelligence AI;目前与 EvenUp 集成 | AI 范围窄于 EvenUp(摘要 vs. 全生命周期) |
| Harvey | 横向法律 AI | $300M+ 融资(估计);2022 年创立;在 BigLaw 快速扩张 | AmLaw 200、BigLaw、企业法务 | 按事项定制 AI 智能体;BigLaw 研究与起草能力强 | 未针对 PI 原告工作流或医疗记录分析校准 |
| CoCounsel (Thomson Reuters) | 横向法律 AI | ~$650M 被 TR 收购(2023 年);Westlaw / Practical Law 集成 | 企业级诉讼与企业法务 | Westlaw 与 Practical Law 集成;TR 品牌信任;Supio 合作关系 | 没有 PI 专属训练;TR / Supio 合作可能限制 EvenUp 访问 TR |
| Lexis+ 与 Protégé | 横向法律 AI | LexisNexis(RELX 子公司);大型企业级 | 律所助理律师、诉讼律师、交易律师 | Shepard's 引证核验;DMS 集成;Lexis 数据库访问 | 通用型;没有 PI 判决 / 和解语料;未针对风险代理收费 ROI 校准 |
| 通用 LLM(ChatGPT、Claude、Gemini) | DIY 替代品 | OpenAI、Anthropic、Google——融资 / 资金规模达数十亿美元 | 任何具备提示工程能力的律师 | 持续改进;单案授权成本为零 | 没有 HIPAA 隔离;没有行级引证可追溯性;没有 PI 数据或和解基准 |
| 人工律师助理 / 内部索赔函撰写人 | 现状替代品 | 内部成本中心;标准劳动力市场费率 | 所有 PI 律所(许多律所的当前标准) | 律师监督;定制起草;既有工作流 | 每份索赔函 5–20 小时;量大后错误率更高;扩展性受人头限制 |
私营竞争对手(Filevine、Litify、CasePeer、SmartAdvocate、Supio)的融资数字除注明外未公开披露。Harvey 融资基于截至 2025 年中的媒体估计。EvenUp 规模数据截至 2026 年 5 月,来源为公司披露、Bloomberg Law 与 LawNext 报道。
[CP001, CP006, CP007, CP014, CP015, CP016]3.2 直接 PI AI 竞争者
Supio 是 EvenUp 最直接、资源最充足的 PI AI 竞争者。Supio 由 Jerry Zhou(CEO)和 Kyle Lam(CTO)于 2021 年在 Seattle 创立,并于 2026 年初开设 San Francisco 办公室;它自称是“唯一为原告法律和群体侵权案件构建的智能体法律 AI 平台”。其 2026 年 5 月发布的 Supio Agent 覆盖案件级和律所级智能体工作流,2026 年 4 月又与 Thomson Reuters 合作接入 Westlaw Advantage,上述动作把它定位成有能力的全谱系对手。Supio 提供医疗时间线、索赔准备、intake 到和解工作流、面向诉讼的 Case Bench,以及律所级分析。其定价模式有两档(面向高案量律所的 Case Subscription;追求最低单案成本的 Unlimited Firm Access),没有锁定合同或平台费,直接对标 EvenUp 于 2025 年 5 月宣布的按案件计价转向。 截至 2026 年 5 月,Supio 官网称“已有数百家人身伤害律所加入”,明显少于 EvenUp 的 2,000+ 家,说明按客户数看 Supio 大约小 5–10 倍。Supio 尚未公开披露风险融资;其产品节奏(Supio Agent、Supio Intake、Thomson Reuters 集成、2026 年 2 月开设 SF 办公室)显示其仍在投入 go-to-market 和产品。Supio 的差异化定位在于平台原生构建于智能体工作流,而不是把 AI 改装到文档生成上。 Supio 的 Thomson Reuters / Westlaw Advantage 集成(2026 年 4 月宣布)是一项战略差异化:把经过验证的法律引用能力直接嵌入平台后,Supio 回应了 EvenUp 尚未通过外部数据库集成解决的法律引用 AI 幻觉风险。EvenUp 提供到源文件的逐行引用(对医疗事实而言 HIPAA 合规且高度准确),但没有集成经过验证的外部法律引用数据库。随着法院在 2026 年施加更严格的 AI 披露要求,法律引用数据库缺口可能会影响合规敏感的大型律所。 除 Supio 外,尚未公开发现其他规模相当、融资独立的 PI 专业 AI 公司。Darrow AI 聚焦诉讼上游的法律风险敞口管理——在损害发生前发现监管风险和诉讼信号——不是直接的 PI 原告索赔函工作流竞争者。因此,有效的直接同行竞争集合就是 EvenUp(规模领导者)对 Supio(智能体原生挑战者)。 [CP006, CP007, CP008, CP009, CP010, CP011]
| 采购标准 | EvenUp | Supio | Filevine (LOIS) | Litify (ACE) | Harvey | Lexis+ Protégé 产品 |
|---|---|---|---|---|---|---|
| PI 专属索赔函起草 | 是——AI Drafts、Express Demands、Mirror Mode 按律所定制语气 | 是——集成索赔函与案件构建工具 | 是——FilevineAI 医疗时间线 / 索赔函(2025 年 4 月发布) | 是——ACE 源链接索赔材料包 | 部分——通用诉讼起草,未针对 PI 调优 | 部分——通用诉讼起草,不是 PI 专属 |
| 医疗时间线生成 | 是——核心能力;ICD 编码伤情、治疗缺口检测 | 是——医疗时间线内置到案件工作流 | 是——FilevineAI 医疗时间线(2025 年 4 月) | 是——ACE 即时医疗时间线(源链接) | None | None |
| 和解估值 / 判决基准 | 是——250,000+ 份判决 / 和解数据集;量化损害赔偿 | 未知——未公开披露 | 未知——PI 垂直未公开披露 | 部分——基于历史数据预测案件价值 | None | None |
| 主动式工作流智能体(从进件到证据开示) | 是——全生命周期:进件、治疗跟踪、索赔函时点、谈判、证据开示 | 是——Supio Agent(案件级与律所级,2026 年 5 月) | 部分——智能体医疗时间线 / 证言笔录;不是完整 PI 生命周期 | 部分——ACE 进件智能体、索赔函;不是完整生命周期 | 部分——按事项配置法律智能体(2026 年);不是 PI 专属 | None |
| HIPAA 证明 | 是 | 是(符合 HIPAA 与 GDPR) | 是(企业级合规) | 是(Salesforce Trust 模型) | 未知——未披露 | 是(RELX 企业级合规) |
| SOC2 认证 | 是(经 SOC2 审计) | 是(SOC2 Type 2) | 是 | 是 | Unknown | 是 |
| 法律引证核验(Westlaw / Shepard's) | 部分——行级来源引用;无外部法律数据库 | 是——Westlaw Advantage 直接集成(2026 年 4 月) | 未知——未公开披露 | 未知——未公开披露 | 部分——通用法律研究 | 是——Shepard's 引证核验 |
| CMS 集成(Filevine、Litify、CasePeer) | 是——已确认 SmartAdvocate、Litify、CasePeer | 是——已披露 CMS 集成 | N/A(本身就是 CMS) | N/A(本身就是 CMS) | None | None |
| 大规模侵权支持 | 未知——未突出披露 | 是——明确列为用例 | 部分——大规模侵权在路线图中 | 部分——企业级 PI 包含大规模侵权 | None | None |
| 单案透明定价 | 是——单案定价模式(2025 年 5 月) | 是——案件订阅或不限量;不锁定 | 无——仅定制定价 | 无——仅定制定价 | 无——企业授权 | 无——企业 / 订阅授权 |
标为“未知”的矩阵单元表示研究时点(2026 年 5 月)缺少公开产品文档。“部分”表示披露能力只能不完整地满足该标准。 本矩阵不反映输出质量对比;这需要独立基准测试,而公开来源没有。
[CP007, CP008, CP010, CP014, CP015, CP016]3.3 工作流平台既有厂商
五大 PI 导向案件管理系统——Filevine、Litify、Clio、CasePeer 和 SmartAdvocate——同时是 EvenUp 的集成伙伴,也可能成为平台侵入型竞争者。每家公司都在 2024–2026 年推出 AI 能力,切入 EvenUp 核心工作流的若干环节。 Filevine 是该组中最大的企业法律 AI 平台。其 LOIS AI 层提供医疗记录分析(FilevineAI for medchrons,2025 年 4 月)、庭审记录和证言管理(2025 年 8 月)、截止日期管理(2025 年 10 月)、文档生成和实时案件分析。Filevine 2026 年以“2026 Legal AI Trust Index”平台定位,主张 AI 信任来自律所自己的数据和工作流——该定位直接反击了 EvenUp 的专属数据集论点。Filevine 定价完全按团队定制,显示其企业销售动作。 Litify(原生基于 Salesforce)面向企业级 PI 律所,其 Plaintiff Practice Management 模块叠加“Litify ACE”(Agentic Case Expert):AI intake 智能体、源链接医疗时间线、索赔包起草、案件价值预测和和解情景建模。Litify 的 Salesforce 血统带来强企业集成和分析能力,但也让 SMB 律所承担更高实施负担。 Clio 在覆盖所有执业领域的 SMB 法律执业管理市场占据主导,也包括 PI。Clio 2025 / 2026 年产品扩展包括 Clio Work(AI 案件分析、研究集成、基于案件事实和法律的起草)、Clio Draft(文档自动化)以及收购 CasePeer(PI 专业 CMS)。Clio Work 以“唯一理解你的案件、案件背景和法律的 AI”定位,与 EvenUp 的平台愿景直接重叠,尽管 Clio 的训练跨执业领域,而非 PI 专属。Clio 的 250+ 集成生态和广泛 SMB 采用,为其自有 AI 功能提供天然分发。 CasePeer 是 PI 专业 CMS(Clio 子公司),面向 SMB 提供交钥匙价值主张。其 AI 能力跟随 Clio 产品路线图,而不是基于 PI 专属训练。SmartAdvocate 提供 SmartIntelligence(内置 AI,用于案件摘要、医疗记录分析、邮件润色、翻译和转录),并且目前根据 EvenUp 自身博客与 EvenUp 集成,同时也在增加自有原生 AI 能力。 核心战略张力在于:EvenUp 与 Filevine、Litify、CasePeer 和 SmartAdvocate 集成,借助 CMS 数据访问驱动其 AI 工作流。当这些 CMS 平台增加自有 AI 功能、复制 EvenUp 的核心能力时,它们维持开放集成的动机会下降。由此产生的平台依赖风险,又被每家 CMS 供应商用更低增量成本利用同一批案件数据的能力进一步放大。 [CP014, CP015, CP016, CP017, CP018, CP019]
| 公司 | 模式 | 单位 / 基础 | 已披露区间 | 主要包含项 | 含义 |
|---|---|---|---|---|---|
| EvenUp | 单案 SaaS | 按处理案件 | 未公开;单案模式于 2025 年 5 月发布 | 完整 Claims Intelligence Platform:索赔函起草、医疗时间线、工作流智能体、分析 | ROI 与案件量透明挂钩;ARPU 估计为每律所每年 $5K–$20K(分析师估计) |
| Supio | 案件订阅或律所不限量访问 | 按案件量或不限量 | 未公开列示;无锁定合同,无平台费 | CaseAware AI、Case Engine(诉前)、Case Bench(诉讼)、CMS 集成、上线培训 | 单案经济性与 EvenUp 可比;不限量方案瞄准高案量律所 |
| Filevine | 定制企业级 | 按用户 / 事项;定制 | 未公开;企业级谈判 | LOIS AI(智能体)、医疗时间线、证言笔录工具、文档生成、事项分析 | 入门成本高;利好有规模的律所;集成深度有优势 |
| Litify | 定制企业级(Salesforce 原生) | 按用户 / 事项;定制 | 未公开;企业级谈判 | ACE 进件 / 索赔函、和解建模、Salesforce 分析、事项管理 | 实施和授权成本高;有 Salesforce 开销;适合超大型 PI 业务 |
| Clio (Work + Manage + Draft) | 分层 SaaS | 按用户 / 月 | Manage ~$49–$149/用户/月(公开层级);Work AI 未列价 | 事项管理、Clio Work AI、Clio Draft 文档、250+ 个集成 | SMB PI 用得起;Work AI 是附加项;对单人 / 小型律所,PI 总成本低于 EvenUp |
| 人工律师助理(内部) | 人工成本 | 按小时或薪资 | $50K–$80K/年全成本;离岸约 $15–$40/小时等价 | 人工判断;定制草稿;既有工作流;律师监督 | 被 EvenUp 单案经济性替代;每份索赔函 5–20 小时,按 $50–$100/小时计为 $250–$2,000/份 |
EvenUp 和 Supio 不公开单案价格;价格在销售对话中披露。Clio 公开层级价格来自其官网,但 Clio Work AI 定价需要销售沟通。EvenUp ARPU 是分析师估计,基于律所数量、融资规模和披露客户数;不是公司披露数字。
[CP007, CP009, CP020, CP030, CP031, CP032]3.4 横向法律 AI 平台
三个资金充足的横向法律 AI 平台,如果扩展到 PI 原告垂直赛道,可能成为未来竞争者:Harvey、CoCounsel(Thomson Reuters)以及带 Protégé 的 Lexis+(LexisNexis)。 Harvey AI(2022 年成立于 San Francisco)估计到 2025 年已融资 $300M+,面向 AmLaw 200 律所、BigLaw 业务组和公司法律顾问,提供法律研究、合同起草、诉讼文件准备和 M&A 尽调。Harvey 2026 年产品方向包括“为每个案件打造、按你需求定制的法律智能体”——可按案件类型定制的智能体工作流。Harvey 未公开面向 PI 原告律所营销,也缺少 PI 专属训练语料;其战略重心在高收入 BigLaw 买方,使近期进入 PI 的经济吸引力有限。 Thomson Reuters 的 CoCounsel(2023 年以约 $650M 从 Casetext 收购)定位为基于 Westlaw 和 Practical Law 数据库的企业 AI 助手,处理法律研究、合同分析、证言准备,以及诉讼和公司业务的文档审查。值得注意的是,Thomson Reuters 同时与 Supio 合作(Westlaw Advantage 集成)并在更广泛法律 AI 领域竞争——合作关系可能限制 EvenUp 未来接入 Thomson Reuters 研究基础设施。 带 Protégé 的 Lexis+(原 Lexis+ AI,2026 年改名)是 LexisNexis 的一体化法律 AI 产品,把 Protégé AI 助手、Shepard's 引用验证、iManage / NetDocuments DMS 集成和移动访问结合起来。其明确能力包括从上传材料中起草证据开示文件、证言问题和诉讼文件——在概念上与 EvenUp 的文档套件重叠。不过,它没有用 PI 专属医疗记录和判决数据训练,在需要 ICD 编码伤情分析和和解价值基准比较的 PI 索赔工作流中竞争力有限。 横向法律 AI 威胁主要是长期商品化风险:如果模型能力持续提高,基础 AI 供应商又能大规模训练法律数据,今天需要 PI 专属训练才能形成的差异化可能会被侵蚀。EvenUp 的回应——平台深度、主动智能体和 250,000+ PI 判决与和解的自有数据集——具备防御性,但如果不能持续积累数据,就并非永久优势。 [CP023, CP024, CP025, CP026, CP027, CP028]
3.5 现状与人工替代方案
EvenUp 替代的最大竞争力量是现状:人工律师助理和索赔函撰写工作流,再加上通用文档模板,以及近来没有 PI 专属校准的通用 LLM。 在 EvenUp 出现前,一封典型 PI 案件索赔函需要律师助理或专职律师投入 5–20 小时:审阅医疗记录、构建时间线、起草叙事、纳入 ICD 代码,并与监督律师反复修改。按律师助理 $50–$100 / 小时费率计算,仅每封索赔函的人力成本就代表单案 $250–$2,000 的内部成本。EvenUp 的按案件计价模式,就是要定在低于或接近这一替代价值的位置。EvenUp 客户证言支持 ROI 主张:Jeffcoat Injury Lawyers 使用 Express Demands 后生成 3 倍更多索赔函,案件和解快 30 天;ELG Injury Lawyers 实现 400%+ 收入增长;John K. Zaid & Associates 报告和解改善 300%。 离岸法律支持供应商代表混合替代方案:这些公司以等同 $15–$40 / 小时的成本提供文档处理和基础时间线服务。它们缺少 AI 驱动的和解基准比较、ICD 编码、治疗缺口检测和案件财务分析。EvenUp 的按案件计价瞄准同一经济位置,但提供更高质量和更快周转。 通用 LLM(ChatGPT、Claude、Gemini)是越来越有分量的 DIY 替代方案,尤其对熟悉 prompt engineering 的律师而言。EvenUp 博客明确回应通用 LLM 竞争,列出通用 LLM 不足的五个原因:缺少 PI 专属训练数据,没有 HIPAA 合规数据隔离,没有可追溯到源文件的逐行引用,没有用于和解估值的 PI 判决基准数据,也没有主动工作流智能体。EvenUp 的 Smart Workflows 和主动智能体提供了仅靠 prompt 型 LLM 使用无法复制的功能,除非投入大量定制工程——尤其是驱动 EvenUp 索赔时点提醒的 42% 统计数据(索赔函在最后一次治疗后 >100 天才发送)。 反向情景是:不断改善的通用 AI 工具消除了文档生成中的大部分生产力差距,让 EvenUp 的竞争重点从输出速度转向合规、数据集质量和工作流集成。 [CP030, CP031, CP032, CP033, CP034, CP035]
3.6 护城河耐久性与反向竞争证据
EvenUp 的竞争护城河建立在四个相互扣合的支柱上,每个支柱都有不同的耐久性和威胁向量。 支柱 1——自有 PI 数据集:EvenUp 的 Piai 系统用 250,000+ 起判决和和解、数百万次医疗就诊,以及来自 2,000+ 家律所的数据训练。该数据集创造了新进入者无法立即复制的训练优势,估计需要三到五年才能在规模上追平。数据集护城河的长期威胁,来自可能收购或授权同等数据的 LLM 供应商,也来自 Supio 的 Thomson Reuters 集成所提供的法律引用验证层。尚无竞争者披露可比的 PI 专属判决 / 和解语料。 支柱 2——网络效应与品牌:EvenUp 渗透前 100 大 PI 律所的 30%(2026 年 5 月),形成了标杆客户护城河。随着 EvenUp 处理更多案件,其和解基准数据会继续改善,产生数据网络效应。该动态由以下增长体现:律所客户从约 1,500 家(2025 年中)增至 2,000+ 家(2026 年 5 月),前 100 大律所渗透率从 20%(2025 年 10 月 Series E)增至 30%(2026 年 5 月)。反向情景是:Supio 和 CMS 既有厂商也在签下同一批顶级 PI 律所,可能通过多归属削弱 EvenUp 的数据独占优势。 支柱 3——工作流集成深度:EvenUp 的主动智能体模型会持续扫描案件数据、触发索赔时点提醒、管理治疗跟进、标记文档缺口;一旦律所围绕 EvenUp 智能体标准化案件流程,切换成本就会形成。不过,CMS 既有厂商掌握所有案件数据产生的数据层;如果 Filevine、Litify 或 Clio 达到相当的 AI 深度,单独购买 EvenUp 订阅的增量价值会被侵蚀。 支柱 4——合规与信任:EvenUp 的 SOC2 认证和 HIPAA 证明回应了 AI 采用中的主要法律伦理障碍。不过,Supio 也取得 SOC2 Type 2 和 HIPAA / GDPR 合规,所有主要 CMS 平台也维持等同认证。合规趋同后,合规护城河会降为准入门槛。 主要反向竞争证据包括:(1)CMS 平台侵入——SmartAdvocate 与 EvenUp 的集成(EvenUp 博客确认)显示了竞合动态;随着 SmartAdvocate 内置 AI 改善,对单独 EvenUp 层的需求会下降;(2)索赔函商品化——保险理算员大规模遇到 EvenUp 生成索赔函后,可能形成系统性回应规则,削弱和解改善优势,EvenUp 也通过 Mirror Mode 承认这一风险;以及(3)Supio 的 Westlaw 集成提供了 EvenUp 平台缺少的法律引用验证。 净评估:EvenUp 是规模领导者,拥有耐久的数据集护城河,可为其相对直接同行提供三到五年的跑道。它最严重的中期威胁不是 Supio 或横向 AI 玩家,而是 Filevine、Litify 和 Clio 的有机 AI 演进——这些平台已经掌握 EvenUp 的分发渠道,并能用更低增量成本利用同一批案件数据。 [CP036, CP037, CP038, CP039, CP040, CP041]
| 护城河主张 | 主要威胁 | 严重性 | 时间范围 | 缓释措施 / 尽调问题 |
|---|---|---|---|---|
| 专有 PI 数据集(250K+ 份判决 / 和解) | LLM 提供商或竞争对手获取同等 PI 数据;CMS 既有玩家凭借工作流所有权积累可比数据 | 高 | 3–5 年 | 跟踪 Supio、Filevine、Litify 的数据集披露;确认 EvenUp 数据许可条款阻止竞争对手访问 |
| Top 100 PI 律所 30% 渗透率与品牌领先 | CMS 既有玩家用打包定价向同一批律所扩展 AI;Supio 签下 Top 100 客户 | 中 | 1–3 年 | 监测 EvenUp 在 Top 100 客户中的流失;跟踪企业销售周期中的竞争输赢 |
| 全生命周期主动式智能体(从进件到证据开示) | CMS 平台原生加入可比的智能体深度;Supio Agent(2026 年 5 月)覆盖同一生命周期 | 中 | 1–2 年 | 在具体工作流步骤上,对比 EvenUp 智能体深度与 Supio Agent、Filevine LOIS |
| SOC2 / HIPAA 信任与合规姿态 | Supio 具备 SOC2 Type 2 与 HIPAA / GDPR 同等能力;合规是标配,不是差异化 | 低 | 当前 | 跟踪新合规要求(州律师协会 AI 指引、HIPAA 针对 AI 的规则) |
| CMS 集成访问(Filevine、Litify、CasePeer、SmartAdvocate) | 随着 CMS 平台自身 AI 成熟,它们可能限制 EvenUp API 访问;数据访问护城河取决于第三方许可 | 高 | 2–4 年 | 确认集成协议的合同性质;评估排他性保护;跟踪 CMS AI 路线图 |
| 和解改善 ROI 叙事(300% 改善主张) | 保险理赔员在规模化后学会折价 EvenUp 风格索赔函;索赔函格式商品化压低和解增量 | 中 | 2–3 年 | 索取匿名化队列结算改善数据;评估 Mirror Mode 在规模化场景下的效果 |
| Supio Westlaw 集成——EvenUp 的法律引文缺口 | Supio 的 Westlaw Advantage 集成提供 EvenUp 缺少的已核验法律引文;2026 年法院开始施加更严格的 AI 披露要求 | 中 | 当前 | 评估 EvenUp 产品路线图是否会接入外部法律数据库;判断 PI 索赔需求是否需要频繁核验法律引文 |
严重程度评级反映威胁一旦落地可能对 EvenUp 竞争位置造成的影响。时间窗口是基于截至 May 2026 公开可见的竞品产品路线图和市场信号作出的方向性估计。内部路线图和合同安排可能大幅改变这些判断。
[CP036, CP037, CP038, CP039, CP040, CP041]3.7 图表
04财务情况
4.1 收入模式与变现架构
从创立到 2026 年 5 月,EvenUp 的商业化策略已经明显改道,最终形成两条并行收入线:面向 AI 驱动 Claims Intelligence Platform 的按案收费 SaaS 模式,以及面向 PLAAS 的订阅式托管服务模式。 2025 年 5 月以前,EvenUp 采用分层功能授权销售平台访问权——需求函生成、医疗时间线等模块分别定价。2025 年 5 月,EvenUp 推出重构后的 Claims Intelligence Platform,主打「每案一个清晰、可预测的成本」,把全套产品(AI Drafts、Smart Workflows、Medical Bill Summary、Express Demands)整合成统一的按案计费单元。这一转向让 EvenUp 的定价贴合人身伤害(PI)律所客户的胜诉收费经济模型:律所按处理案件付费,而不是按用户席位或功能访问付费,因此可以直接衡量 ROI。单案价格没有公开披露,需要通过销售沟通确定;分析师基于律所数量、累计融资额和披露的客户数量轨迹,估计每家律所平均收入(ARPU)在每年 $5,000–$20,000 区间,但这不是公司披露数据。 2026 年 5 月,EvenUp 推出 PLAAS(Pre-Litigation as a Service),这是创立以来最重要的收入模型扩张。PLAAS 把 EvenUp 的 AI 技术与一支 300+ 人的美国法律运营团队结合起来,覆盖完整诉前案件管理生命周期:索赔建档、护理协调、医疗记录调取、需求准备、和解谈判,以及可选的留置权处理。不同于按案计费的 SaaS 产品,PLAAS 以订阅方式销售,覆盖律所在诉前阶段的全部案件。LawNext 对 2026 年 5 月发布的报道显示,早期测试带来 $10M+ 的 PLAAS 订阅销售额;这是所有公开来源中唯一具体的收入数字。 第三条收入线 AI Drafts Suite 把专业文档生成能力(AI Drafts 证据开示文件、Medical Bill Summary、Smart Workflows 自动化)打包进整合平台。EvenUp 的公司介绍页和新闻稿显示,PLAAS 可选包含留置权处理服务,意味着未来可能在和解阶段再叠加一层增购变现。 截至 2026 年 5 月,收入完全来自美国市场;EvenUp 没有披露国际扩张计划,所有公开来源也都把法律运营团队描述为美国团队。 [CI001, CI002, CI003, CI004, CI005, CI006]
| 产品 / 服务 | 类型 | 定价模型 | 发布时间 | 核心功能 | 收入阶段 |
|---|---|---|---|---|---|
| Claims Intelligence Platform 平台 | AI SaaS | 按案件计费(未公开列示) | May 2025(统一按案件计费) | 索赔函起草、MedChrons、和解基准对比、Smart Workflows、分析、Express Demands、Mirror Mode | 核心平台;主要 ARR 驱动项 |
| AI Drafts Suite | SaaS 模块 | 包含在 Claims Intelligence 按案件计价中 | May 2025 | AI 生成索赔函、证据开示文件和结构化文档工作流 | 打包销售;不单独定价 |
| Medical Bill Summary | SaaS 模块 | 包含在 Claims Intelligence 按案件计价中 | May 2025 | AI 生成医疗费用摘要,标记治疗缺口 | 打包销售;不单独定价 |
| PLAAS — 诉前即服务(Pre-Litigation as a Service) | 托管服务订阅 | 全案件订阅(未公开列示) | May 2026(早期测试) | AI + 300+ 名美国法律运营人员:索赔建档、护理协调、病历调取、索赔函准备、和解谈判,可选留置权处理 | 早期测试已售出 $10M+ 订阅 |
| Medical Management | 托管服务附加项 | 未公开披露 | December 2025 | 主动识别治疗缺口、随访智能体、护理协调,处理压低案件价值的治疗不足问题 | 新产品;收入早期 |
| Express Demands | SaaS 功能 | 包含在 Claims Intelligence 按案件计价中 | January 2025 | 面向较简单 PI 案件的快速索赔函;首个 AI Documents 版本 | 打包功能 |
May 2025 引入按案件计费,用来统一此前按功能分层授权的模式。具体按案件价格未公开披露,会随律所规模和案量变化。$5,000–$20,000/firm/year 的 ARPU 分析师估计基于律所数量、累计融资和增长率,并非公司披露数据。除 LawNext(May 2026)报道的早期测试 $10M+ 订阅额外,PLAAS 订阅定价和 PLAAS ARR 均未公开披露。
[CI001, CI002, CI003, CI004, CI006]| 融资轮次 | 交割 / 公布日期 | 金额(USD) | 领投方 | 参投方 | 投后估值 |
|---|---|---|---|---|---|
| 种子轮 | ~2022(估计) | 未披露 | SignalFire | 未公开确认 | 未披露 |
| Series A/B 轮 | 2022–2023(估计) | 未披露;D 轮前累计融资隐含约 $100M | Bessemer Venture Partners 等 | 未完全公开确认 | 未披露 |
| Series C 轮 | ~2023(估计) | 未披露;属于 D 轮前约 $100M 的一部分 | 未公开确认 | 未公开确认 | 未披露 |
| Series D 轮 | October 8, 2024 | $135M | Bain Capital Ventures | 投资方:Premji Invest、Lightspeed VP、Bessemer VP、SignalFire、B Capital Group | >$1 billion(独角兽) |
| Series E 轮 | October 1, 2025 | $150M | Bessemer Venture Partners | REV(RELX/LexisNexis 风险投资部门)、B Capital Group、Bain Capital Ventures | >$2 billion |
种子轮至 Series C 轮细节未公开确认。D 轮前累计融资为推算值(~$100M),因为 Series D 时披露的累计融资为 $235M($135M Series D + 约 $100M 此前轮次,依据 EvenUp 的 Series D 博文称累计融资 $235M)。Series E 后累计融资达到 $385M。ServiceTitan 在 SEC EDGAR 的 DRS/A 文件(October 2024)将 EvenUp, Inc. 标识为一家私有软件公司;EDGAR 中未发现 EvenUp 的 Form D 文件。
[CI008, CI009, CI010, CI011, CI012, CI013]4.2 融资历程与资本结构
EvenUp 已披露的风险投资融资总额为 $385M,至少横跨五轮(种子轮至 Series E),但公开确认轮次细节的只有 Series D 和 Series E。截至 2026 年 5 月,公司仍为私营公司;在 SEC EDGAR 中用任何 EvenUp 名称变体都未找到 Form D 备案,这与无需提交 Form D 的 Regulation D 豁免相符,也可能是文件以尚未检索到的实体名称误提交。 Series D 于 2024 年 10 月 8 日宣布,金额 $135M,由 Bain Capital Ventures 领投。参与方包括 Premji Invest、Lightspeed Venture Partners、Bessemer Venture Partners、SignalFire 和 B Capital Group。投后估值超过 $1B,EvenUp 因此成为独角兽。当时累计融资约 $235M,意味着从种子轮到 Series C 合计约融资 $100M。EvenUp 博客称,Series D 前十二个月员工数翻倍,收入同比增长超过 100%。 Series E 于 2025 年 10 月 1 日宣布,金额 $150M,由 Bessemer Venture Partners 领投。新投资方包括 RELX/LexisNexis 的风险投资部门 REV,B Capital Group 和 Bain Capital 作为既有投资方继续参与。投后估值超过 $2B。REV 代表 RELX/LexisNexis 入股具有战略意义:这意味着一家重要法律研究和工作流竞争者(LexisNexis)的母公司持有 EvenUp 少数股权,可能反映合作选择权,也可能带来竞争情报;这一双刃结构需要对任何共同投资或数据共享条款做尽调。 Bloomberg Law、Reuters 和 AboveTheLaw 对 Series E 的报道口径一致:估值 $2B+,股权融资 $150M。EvenUp 的 GlobeNewswire 新闻稿以及 SEC EDGAR 上 ServiceTitan 的 DRS/A 文件也相互印证;后者确认 EvenUp 的法律实体为 EvenUp, Inc.,是一家特拉华注册的私营软件公司,并自 2023 年 4 月起拥有董事会代表。 没有任何债务融资、可转债或老股交易被公开披露。已披露股权融资总额为 $385M。任何时期的账面现金、现金跑道或烧钱速度也都没有披露。 [CI008, CI009, CI010, CI011, CI012, CI013]
| 成本项 | 类别 | 规模指标 | 披露程度 | 备注 |
|---|---|---|---|---|
| 美国法律运营团队(PLAAS + QA) | COGS / 托管服务人工 | 300+ 名美国员工 | 公司披露员工数 | 估计 $45M–$75M/yr,全成本按 $100K–$150K/人计;PLAAS 交付成本核心 |
| 工程与产品开发 | 研发 | Series D 时 100+ 名工程 / 产品人员 | Series D 披露(Oct 2024) | 包括 Piai 模型训练、产品开发、平台基础设施;GPU 算力成本较高 |
| AI 基础设施(Piai 模型算力) | COGS / 基础设施 | 截至 Oct 2025 已处理 200,000+ 起案件 | 由案件量推算;未披露 | 估计规模化推理 + 训练为 $5M–$20M/yr;典型法律 AI LLM 工作负载 |
| 合规与信任(SOC2、HIPAA、BAA) | G&A / COGS | SOC2 Type II 认证;HIPAA 证明 | 公开材料确认 | 持续审计成本、PHI 安全基础设施、法律合规团队;未量化 |
| 销售与市场 | OpEx | 未披露 | 未披露 | 包括面向 2,000+ 家律所的直销、合作伙伴渠道(CMS 集成)、活动和品牌 |
除已披露员工数外,所有成本数字都是分析师估计;公司未发布财务报表。PLAAS 把法律运营团队从成本中心变成创收单元。如果 PLAAS 定价覆盖法律运营成本,混合毛利率可能接近 SaaS 水平(70–80%+);否则混合毛利率会显著低于纯软件同业。
[CI016, CI017, CI018, CI019, CI020]4.3 成本结构与单位经济
EvenUp 的成本画像与纯软件 SaaS 公司显著不同,关键原因是其 300+ 人美国法律运营团队。这支团队支撑 PLAAS,也支撑持续的数据集与质保运营。这种劳动密集环节带来结构性毛利率压力;在以 $2B+ 估值融资的 AI 初创公司中并不常见。 Series D 时(2024 年 10 月),EvenUp 披露过去十二个月员工数翻倍,并拥有约 100+ 名工程和产品人员。300+ 人法律运营团队明确位于美国;按美国法律运营岗位的标准薪酬水平(每名员工全口径 $100K–$150K)估算,这意味着每年全口径人工成本约 $45M–$75M,且尚未计入工程、销售市场和基础设施成本。 2026 年 5 月推出 PLAAS 后,法律运营团队从成本中心(质保、数据集整理、支持)转为创收的托管服务团队。如果 PLAAS 收入足以覆盖法律运营团队的全口径成本,底层 SaaS 平台的单位经济可能接近典型 SaaS 毛利率(70–80%+)。但如果 PLAAS 采用速度慢,或定价无法覆盖法律运营成本,EvenUp 的混合毛利率会继续显著低于纯软件水平。 AI 基础设施是第二条重要成本线:Piai 模型在 200,000+ 个案件上训练和推理,需要 GPU 算力。EvenUp 没有披露云服务商或算力支出。参考可比法律 AI 公司,200K+ 案件 / 年的推理成本可能在 $5M–$20M 之间,取决于模型架构和提示词复杂度。 合规与信任投入构成第三个成本维度:SOC2 Type II 审计、HIPAA Business Associate Agreement(BAA)合规计划,以及处理 PHI(Protected Health Information,受保护健康信息)所需的网络安全基础设施,相比非医疗 SaaS 都会增加实质性开销。 任何时期的收入、毛利率、EBITDA 或烧钱数据都没有公开披露。累计融资 $385M、后期估值 $2B+ 表明,投资人在为可观的持续亏损提供资金,同时押注公司能扩张到可防守的长期利润率结构。 [CI016, CI017, CI018, CI019, CI020, CI021]
| 指标 | 数值 | 日期 | 来源 | 验证状态 |
|---|---|---|---|---|
| 收入增长(YoY) | >100% | Oct 2024(Series D) | EvenUp 博客(Series D 公告) | 公司声称;未经独立审计 |
| 每周处理的索赔文件 | 1,000+(Series D);~10,000(May 2026) | Oct 2024 / May 2026 | EvenUp 博客;EvenUp 主页 | 公司声称 |
| PI 律所客户 | 1,000+(Series D)→ 2,000+(May 2026) | Oct 2024 / May 2026 | EvenUp 博客;主页 | 公司声称 |
| Top-100 PI 律所渗透率 | ~30% | May 2026 | Bloomberg Law、EvenUp | 第三方报道;与公司口径一致 |
| 累计处理案件 | 200,000+(Oct 2025) | Oct 2025 | Unite.ai 访谈(COO Raymond Mieszaniec) | 公司通过访谈披露 |
| 为客户追回的赔偿 | $14B+ | May 2026 | EvenUp 主页 | 公司声称 |
| PLAAS 早期测试订阅销售额 | $10M+ | May 2026 | LawNext 报道(发布文章) | 第三方报道 |
| PLAAS 结果:达到保单限额 | 可用第三方限额的 95% | May 2026 | LawNext(发布文章) | 第三方报道;数据来自公司 |
| PLAAS 结果:病历调取速度 | 比行业常态快 66 天 | May 2026 | LawNext / EvenUp | 公司声称;仅限早期测试队列 |
| PLAAS 结果:索赔函准备速度 | 比行业常态快 47 天 | May 2026 | LawNext / EvenUp | 公司声称;仅限早期测试队列 |
| PLAAS 结果:单案成本节省 | 每案 $1,000 | May 2026 | LawNext / EvenUp | 公司声称;仅限早期测试队列 |
| Sweet James:年度结果 | $500M+,YoY 增长 70% | Oct 2025 | Unite.ai 访谈 | 公司通过访谈披露;未经审计 |
| Jeffcoat:索赔函数量提升 | 索赔函数量 3×;和解快 30 天 | May 2025 | EvenUp 博客(AI Drafts Suite 发布) | 客户证言;公司控制来源 |
| Dwuan Hammond:收入增长 | 营收增长 300% | May 2025 | EvenUp 博客 | 客户证言;公司控制来源 |
| 多产品用户占比 | 20% | Oct 2024 | EvenUp 博客(Series D) | 公司声称 |
| 员工增长(Series D 前 12 个月) | 翻倍 | Oct 2024 | EvenUp 博客(Series D) | 公司声称 |
所有牵引指标均由公司披露,或由第三方基于公司数据报道。公司未发布独立审计、队列级分布或控制组对比。PLAAS 早期测试指标只覆盖一部分早期采用律所,未必能外推到平均客户。收入增长率和 PLAAS 订阅金额是截至 May 2026 可得、独立印证最强的财务指标。
[CI022, CI023, CI024, CI025, CI026, CI027]4.4 收入牵引力与增长指标
EvenUp 没有披露 ARR 或收入数字,但一组运营和融资指标间接勾勒出财务轨迹。这些信号持续偏正面,但必须加上限定:它们要么来自公司口径,要么来自公司控制的营销材料,而不是独立审计数据。 Series D 时(2024 年 10 月),EvenUp 披露:收入同比增长超过 100%;每周处理超过 1,000 份需求文件;20% 客户使用多个产品;员工数在十二个月内翻倍。到 Series E(2025 年 10 月),Bloomberg Law 和 AboveTheLaw 报道的律所数量从约 1,500+(Series D 后基线)提升到 2,000+ 以上(2026 年 5 月)。COO Raymond Mieszaniec 在 Unite.ai 的访谈(2025 年 10 月)中披露:为客户的当事人累计追回 $10B+ 损害赔偿,处理 200,000+ 个案件,并在 2026 年 5 月达到超过 $14B 的损害赔偿额。 2026 年 5 月 PLAAS 发布带来了所有公开来源中唯一具体的收入数字:据 LawNext 报道,早期测试阶段订阅销售额超过 $10M。LawNext 和 EvenUp 自有材料还披露了早期结果:取得可用第三方保单限额的 95%;记录调取比行业常态快 66 天;需求包准备快 47 天;PI 律所每案节省 $1,000 成本。 客户 ROI 披露仅来自公司:Sweet James PI 律所报告年结果超过 $500M,收入同比增长 70%,并称 EvenUp 有部分贡献;Jeffcoat Injury Lawyers 报告需求函产出提升 3×,案件和解快 30 天;律所客户 Dwuan Hammond 报告营收增长 300%。Clio 的 2025 Legal Trends Report 虽然不是 EvenUp 专属,但记录了法律 AI 采用浪潮,为 EvenUp 的客户增长提供市场背景。 EvenUp 2026 年 5 月的网站首页称,公司服务 2,000+ 家 PI 律所,每周处理约 10,000 个案件,高于 Series D 时每周 1,000+ 个案件。如果 2025–2026 年仍保持类似节奏,这些增长指标与 Series D 时 100%+ YoY 收入增长的说法一致,但实际收入轨迹无法独立验证。 [CI022, CI023, CI024, CI025, CI026, CI027]
| 投资者 | 轮次 | 类型 | 战略意义 |
|---|---|---|---|
| Bain Capital Ventures | Series D(领投)、Series E | 一线风险投资机构 | 成长阶段专长;Series D 领投方;继续参投 E 轮,显示其看好扩张阶段 |
| Bessemer Venture Partners | Series D、Series E(领投) | 一线风险投资机构;法律 AI 专家 | 领投 Series E;Bessemer 对法律 AI 有深入投资逻辑(发布 “Roads to $100M ARR in Legal AI”);为 EvenUp 的资本效率叙事背书 |
| SignalFire | 种子轮、Series D | 聚焦数据科学的早期风险投资机构 | 种子轮投资者;早期判断坚定;技术与数据科学平台可提供 AI 基础设施支持 |
| B Capital Group | Series D、Series E | 多阶段成长型风险投资机构 | 连续参投多轮,显示 LP 信心延续,也没有降轮担忧 |
| Premji Invest | Series D | 主权 / 家族办公室(Wipro) | 提供近似非稀释的长期耐心资本,并带来企业 IT 关系 |
| Lightspeed Venture Partners | Series D | 一线风险投资机构 | 仅参投 Series D;未参与 Series E 值得注意,公开资料未解释 |
| REV(RELX / LexisNexis 风险投资部门) | Series E | 企业风险投资(战略) | 战略上最复杂的投资者:RELX 拥有 LexisNexis(直接法律研究竞争对手);REV 持股可能指向收购选择权、合作关系,或竞争情报兴趣 |
公开资料无法解释 Lightspeed Venture Partners 缺席 Series E,可能来自组合约束、回报优化,或对 EvenUp 轨迹判断不同。考虑到 LexisNexis 产品在法律 AI 市场构成竞争冲突,REV 参投需要专项尽调投资者权利(董事会观察员、信息权、ROFR)。
[CI009, CI010, CI011, CI012, CI013, CI014]4.5 财务披露缺口与尽调风险
以 EvenUp 所处融资阶段和估值来看,其财务画像明显不透明。公司五年经营历史中,没有任何公开来源披露 ARR、收入、毛利率、经营亏损、烧钱速度、现金跑道或 EBITDA。这种不透明在私营风险投资支持公司中结构上常见,尤其是未来可能 IPO、希望掌控初始财务叙事的公司;但在任何机构投资前,它会制造必须处理的具体尽调风险。 唯一具体收入数据(PLAAS 订阅 $10M+)来自早期测试中的新产品,而不是核心平台 ARR,因此不能可靠代表 EvenUp 的整体经常性收入基础。所有客户 ROI 数字(300% 收入增长、Sweet James $500M+ 结果)都来自 EvenUp 自有营销材料,没有独立佐证,也没有披露方法论。 RELX/LexisNexis 的风险投资部门 REV 参与 Series E,带来战略利益冲突隐忧:RELX 是 LexisNexis 的母公司,而 LexisNexis 在法律 AI 工作流上与 EvenUp 直接或近似竞争。尽调应确认 REV 入股是否包含任何数据访问条款、收购优先要约权,或董事会信息权;这些安排可能让 RELX/LexisNexis 获得竞争情报。 AboveTheLaw 对 Series E 的报道提出了 AI 实践中的标准担忧:需求生成中的算法偏见、流程中 PHI 的数据隐私,以及 EvenUp 由 AI 驱动的索赔金额膨胀可能「进一步堵塞法院系统」——系统性抬高初始需求数字,让保险理赔员学会打折对待这些数字,并在规模化后侵蚀 ROI 优势。这些风险真实存在,也是任何高容量法律 AI 平台的结构性风险;其财务影响(失望客户流失、监管介入、保险公司反制技术)尚未量化。 PLAAS 究竟是真正的新产品类别,还是把既有 AI 套上组织化服务外壳重新包装,仍是开放问题;LawNext 2026 年 5 月报道指出,PLAAS 的长期价值取决于律所整合深度和结果规模,仅凭早期测试数据无法评估。 [CI031, CI032, CI033, CI034, CI035, CI036]
| 缺失数据点 | 重要性 | 可用代理或部分信号 | 尽调路径 |
|---|---|---|---|
| 年经常性收入(ARR) | 核心 SaaS 健康指标;缺少 ARR,估值倍数就没有锚 | PLAAS 早期测试 $10M+;2,000+ 家律所;Series D 时 YoY 增长 100%+ | 索取当前 ARR 和过去 12 个月增长率;需要 NDA |
| 毛利率(SaaS + PLAAS 混合) | 300+ 名法律运营人员带来结构性 COGS;毛利率未知 | 法律运营员工数意味着人工 COGS 至少 $45M–$75M | 按收入流索取毛利率;PLAAS 与 SaaS 分开 |
| 烧钱速度与现金跑道 | 已融资 $385M 后,现金余额和月度烧钱速度决定现金跑道 | 无可用代理;已知最近一轮为 Oct 2025($150M) | 索取月度烧钱、现金余额,以及达到盈亏平衡前的预计现金跑道 |
| PLAAS 单位经济 | PLAAS 是最新收入流;按案件或按订阅的经济性未知 | 早期订阅 $10M+;披露每案节省 $1,000 | 索取 PLAAS 定价结构、毛利率和客户数 |
| 净收入留存率(NRR) | 关键 SaaS 健康指标;反映 2,000+ 家律所中的扩张与流失动态 | Series D 时多产品占比 20%;索赔函量正增长 | 按队列索取 NRR;同时索取 logo 流失和扩张收入数据 |
| 客户集中度 | Top 10 律所是否贡献 >20% ARR 未知 | Sweet James(年度结果 $500M+)显示存在大客户;占比未知 | 索取 Top-10 客户 ARR 集中度;客户收入 HHI |
| REV/LexisNexis 投资者权利 | 企业 VC 的母公司构成竞争;权利条款可能影响竞争独立性 | 公开 Series E 公告未披露投资者权利细节 | 索取完整 Series E 条款清单;确认 REV 的董事会观察权和信息权 |
该清单列出公开资料无法回答的重大财务尽调问题。ARR、毛利率和烧钱速度不厘清,任何 EvenUp 估值分析都必然带有推断性。$2B+ 的 Series E 估值在不同收入倍数(10–50×)下隐含约 $40M–$200M 的 ARR 区间,但这完全是推算。
[CI031, CI032, CI033, CI034, CI035, CI036]4.6 附录
05产品与技术
5.1 Claims Intelligence Platform:产品套件与模块图谱
EvenUp 的 Claims Intelligence Platform™ 是 Piai™ 能力触达 PI 律所的商业入口。截至 2026 年 5 月,平台覆盖十二个不同模块,横跨从接案到和解谈判的完整诉前案件生命周期。 Express Demands™ 于 2025 年 1 月推出,是 EvenUp 第一个 AI Draft 模板,可直接基于案件文件内容生成专业格式的需求函。更完整的 AI Drafts Suite™ 于 2025 年 5 月发布,把需求生成延伸到起诉状、医疗摘要、谈判表和对询问书的回复;所有产出都锚定案件文件数据,而不是开放域 LLM 生成。Smart Workflows 与 AI Drafts™ 配套,提供数据驱动的案件生命周期自动化:它监控案件事件,基于 EvenUp 自有数据集计算需求准备度(该数据集显示,42% 的需求函在最后一次治疗后超过 100 天才发出),并在条件满足时触发员工提醒。Medical Bill Summary 在 Case Financials 中自动跟踪费用,并按保单限额验证开支。 MedChrons™ 把原始医疗记录组织成交互式、经专业审阅的时间线,律师可以浏览和追问。Medical Management 于 2025 年 12 月推出,把时间线扩展为实时治疗跟踪系统:它暴露治疗空档(EvenUp 数据显示,43% 的 PI 案件会出现 30 天治疗空档),并驱动 AI Communication Agents™ 通过语音和 SMS 自动回访客户。 AI Playbooks™(2025 年 7 月推出)会自动分析新上传案件文件,在无需律师提示的情况下暴露责任标记、创伤性脑损伤(TBI)指标、商业被告和保险覆盖问题。Case Companion™ 是一个 AI 助手,可基于原始案件文件回答问题,并提供行级引用。Settlement Repository™ 摄取律所历史和解数据,并用 Piai™ 对伤害、治疗和保单进行匹配,把结果与可比案件做基准对照。Executive Analytics™ 提供律所级 KPI 仪表盘,可按时间、办公室、角色和具体员工筛选。 Voice Agent™ 处于早期访问阶段,是覆盖完整案件生命周期、用于客户触达的 24/7 对话式 AI。PLAAS(Pre-Litigation as a Service)于 2026 年 5 月推出,是一项托管服务订阅,把 Piai™ 自动化与美国 EvenUp 案件管理人员结合起来,覆盖完整诉前工作流。 定价在 2025 年 5 月转为按案模式:一个覆盖完整 Claims Intelligence Platform 的全包价格,取代此前按功能分层授权。PLAAS 作为单独订阅定价,覆盖律所诉前阶段的全部案件。 [CE001, CE002, CE003, CE004, CE005, CE006]
| 模块 | 主要用户 | 状态(May 2026) | 核心差异化 | 尽调缺口 |
|---|---|---|---|---|
| Express Demands™ | 律师助理 / 律师 | 正式发布(Jan 2025) | 从原始案件记录生成索赔函,附行级来源引用 | 理赔员接受率与律师起草索赔函的对比未公开披露 |
| AI Drafts Suite™ | 律师 / 律师助理 | 正式发布(May 2025) | 基于案件事实,覆盖完整文档生命周期起草(起诉状、质询书、谈判表) | 相对人工草稿的质量基准未经独立验证 |
| Smart Workflows | 案件经理 / 律师助理 | 正式发布(May 2025) | 用 EvenUp 自有延迟模式数据集,驱动案件生命周期自动化 | 工作流采纳率和规模化后的实际省时未披露 |
| Medical Bill Summary | 案件经理 | 正式发布(May 2025) | 自动跟踪费用;按保单限额校验医疗费用 | 争议医疗账单的准确率未披露 |
| MedChrons™ | 律师 / 案件经理 | 正式发布 | 由专业人员复核的 AI 辅助交互式医疗时间线 | 与人工编制时间线相比的准确率尚无基准测试 |
| Medical Management | 案件经理 | 正式发布(Dec 2025) | 实时治疗时间线;上传后一小时内标记 30 天治疗间断 | 与专科转诊网络的集成深度不明 |
| Communication Agents™ | 案件经理 / 接案 | 正式发布 | 用于开立理赔、确认保险范围、客户回访、病历跟进的 AI 语音 / SMS 智能体 | 面对对抗性理赔员行为时的通话质量尚无公开测试 |
| AI Playbooks™ | 律师 / 案件经理 | 正式发布(Jul 2025) | 自动分析新文件,标记责任、TBI、商业被告和 DUI 线索 | TBI 和高价值案件标记的误报率未披露 |
| Case Companion™ | 律师 / 律师助理 | 正式发布 | 面向原始案件文件的 AI 问答助手,提供行级引用 | 复杂多文档查询的精确率 / 召回率尚无独立评估 |
| Settlement Repository™ | 律师 / 合伙人 | 正式发布 | 将律所自身和解记录与可比 EvenUp 案件对标的专有数据库 | 跨律所共享和解数据,需要律师信任匿名化处理 |
| Executive Analytics™ | 律所管理层 | 正式发布 | 律所 KPI 仪表盘,可按办公室、角色或员工对标同行 | 对标同行数据质量和样本量未披露 |
| Voice Agent™ | 案件经理 | 早期访问 | 7×24 小时会话式 AI 客户触达,并生成结构化通话摘要 | 在地区口音、医学术语、非英语场景下的准确率未记录 |
状态和发布日期来自 EvenUp 官方产品页面和博客文章(May 2026)。尽调缺口反映缺少独立基准测试。所有性能声明均来自公司。
[CE005, CE006, CE007, CE008, CE009, CE010]| 日期 / 阶段 | 功能 / 里程碑 | 状态 | 战略含义 | 来源 |
|---|---|---|---|---|
| Jan 2025 | Express Demands™ 发布——首个 AI Draft 产品 | 正式发布 | 验证 AI 起草索赔函市场;跑通按案收费概念验证 | EvenUp 博客 / PR Newswire |
| May 2025 | AI Drafts Suite™、Smart Workflows、Medical Bill Summary、按案收费 | 正式发布 | 平台整合:用单个按案 SKU 覆盖全生命周期起草;将 EvenUp 定位为平台而非点解决方案 | EvenUp 博客(introducing-ai-drafts-suite) |
| Jul 2025 | AI Playbooks™ 和 Voice Agent™(早期访问) | 正式发布(Playbooks);早期访问(Voice) | 从被动转向主动案件分析;7×24 小时 AI 客户触达延展人力产能 | EvenUp 博客(ai-playbooks-voice-agent) |
| Dec 2025 | Medical Management 和 Communication Agents™ 正式发布 | 正式发布 | EvenUp 从文档生成扩展到案件运营管理;为 PLAAS 服务产品铺路 | LawNext Dec 2025 / EvenUp 博客 |
| May 2026 | PLAAS(Pre-Litigation as a Service)和带 Firmwide Knowledge Base 的新版 Case Companion | 正式发布(PLAAS);正式发布(新版 Companion) | 商业模式从 SaaS 扩展到托管服务;早期订阅销售额 $10M+;成本基数结构性上升 | LawNext May 2026 |
| H2 2026(预计) | Voice Agent 全面正式发布;Communication Agent 覆盖范围扩大(留置权解决、证据开示) | 路线图 / 已宣布 | 如果成功,将补齐沟通密集型法律运营任务自动化;降低 PLAAS 人力依赖 | EvenUp 产品页 / AI Playbooks 博客 |
发布日期来自 EvenUp 官方博客文章和第三方新闻报道。H2 2026 预测基于 EvenUp 在产品页和博客中宣布的路线图方向,并非已确认承诺。
[CE005, CE006, CE010, CE012, CE013, CE016]截至 2026 年 5 月,各平台模块的成熟阶段、PI 数据深度、CMS 集成广度和 LLM 商品化风险。
成熟阶段与集成广度来自 EvenUp 官方产品页(2026 年 5 月)。LLM 商品化风险是作者基于已发布产品描述和竞争格局的判断,不是正式市场评估。
[CE041, CE005, CE006, CE009, CE007, CE010]5.2 Piai™ 架构:从原始记录到可行动智能
Piai™ 是 EvenUp 的自研 AI 系统,不是单一模型,而是一套专门为 PI 法律工作流设计的专用模型协同栈。其架构沿着三个功能层推进,直接对应 EvenUp 的 System of Record → System of Intelligence → System of Action 框架。 Reading Layer 是基础摄取阶段。它从各类 PI 案件材料中标准化并抽取结构化信息:版式各异的 PDF、扫描文档、手写笔记、表格、带 ICD 编码的医疗表格、警方报告和多页医疗图表。Reading Layer 维护一组高度专用模型,每一种抽取实体类型对应一个模型(医疗服务提供方名称、服务日期、ICD 代码、治疗描述、法院标识、车辆元数据),并由每天数千条用户纠错信号持续再训练。对于视觉内容(事故示意图、放射影像),EvenUp 使用基于 PPO 的强化学习训练图像转文本摘要器:系统生成基于图像的问题集,评估纯文本 QA 模型能否根据摘要回答这些问题,并训练摘要器,直到可回答性达到基于像素的准确度。摘要在摄取时连同完整来源信息(页面区域、模型版本、时间戳)只存储一次,从而避免昂贵的重复视觉模型推理。 Writing Layer 把文档生成锚定在 Reading Layer 抽取出的事实中,而不是一般世界知识。因此 AI Drafts™ 产出会引用从案件记录中抽取的精确医疗服务方名称、ICD 代码、治疗次数和金额,而不是幻觉出的看似合理数值。每个产出都包含行级引用,把每个主张链接到原始文档、页面和行,便于人工审阅团队和法院把主张追溯回源材料。 在两个核心层之上,System of Record(SoR)持久化标准化案件实体(索赔人、医疗服务方、就诊、手术 / 程序、账单行项目、车辆),并保留关系图和时间线。System of Intelligence(SoI)规划查询执行:事实查找跑在 SoR 上;语气和论证类查询则对原始文档做检索。System of Action(SoA)支持主动自动化:缺失记录请求、需求函时点提醒、治疗空档通知和客户触达,都会在案件时间线的正确时点由 SoR 触发。 EvenUp 的工程副总裁兼 AI 负责人是 Haixun Wang,他是 ACM Fellow 和 IEEE Fellow,以图系统和知识库基础研究闻名,并带来 Microsoft、Google、Amazon、Meta 和 Instacart 的研究脉络。机器学习工程团队包括专攻 NLP、文档理解和优化的博士。 [CE017, CE018, CE019, CE020, CE021, CE022]
| 层 / 组件 | 角色 | 关键依赖 | 技术风险 |
|---|---|---|---|
| 读取层——OCR 与文档规范化 | 将 PDF、扫描件、手写内容、表格、图像转换为保留结构的规范化文本 | 接案时的文档质量;CMS 上传完整性 | 低质量扫描件或非标准医疗记录格式会拉低下游抽取准确率 |
| 读取层——实体抽取模型 | 按实体类型使用专门模型抽取案件实体(服务方名称、ICD 代码、日期、治疗、金额) | 每日用户纠错数据流(每天数千次编辑) | 如果纠错信号量下降,实体模型会漂移;单个实体模型失败会传导到所有引用该实体的下游文档 |
| 读取层——多模态 RL 管线(图像摘要) | 用 PPO 强化学习将事故示意图、放射影像转换为可持久化的结构化文本摘要 | 用于有依据 Q&A 监督的视觉语言模型;RL 训练周期所需算力 | 边缘情况(非常规示意图格式、新型伤情)可能生成低置信度摘要,并静默传导 |
| 事实记录系统(SoR) | 持久化规范案件实体存储,包含溯源图、纠错台账和关系时间线 | Schema 演进管理;跨服务方名称变体的实体链接准确率 | 数百万历史案件上的 schema 变更会带来迁移风险;纠错台账完整性对模型训练质量至关重要 |
| 写作层——有依据文档生成 | 基于 SoR 事实而非开放式 LLM 世界知识起草法律文件(索赔函、诉状、询问书) | 读取层准确率(垃圾进 → 垃圾出);行级引用追溯基础设施 | 基础模型商品化会压窄护城河;约束过强时,生成文本在边缘场景可能显得生硬 |
| 人工复核层(100+ 名员工) | 在交付前验证 Piai™ 输出;纠错回流到读取层模型再训练 | 美国本土人员扩张能力;团队扩大时培训质量一致性 | PLAAS 增长需要线性扩员;招聘快过培训会稀释质量 |
架构描述基于 EvenUp 发布的博客文章 'Building Trustworthy, Scalable Document AI for Legal Tech'(Haixun Wang,VP Engineering / Head of AI)和 Piai™ 产品页。技术风险为作者根据所述系统设计模式作出的评估。
[CE017, CE018, CE019, CE020, CE021, CE023]分层架构覆盖案件数据摄取、System of Record、System of Intelligence,以及面向客户的平台界面。
[CE017, CE018, CE019, CE020, CE021, CE025]端到端案件生命周期流程,展示 Piai™ AI 层与人工复核如何从 CMS 接案一路协作到和解。
PLAAS 流程展示的是托管服务客户;仅使用 SaaS 的客户在「复核 → 谈判 → 解决」环节退出,EvenUp 员工除交付 AI 输出外不再介入。
[CE004, CE011, CE016, CE022, CE031]5.3 CMS 集成、部署与数据依赖图
EvenUp 的集成策略瞄准主流人身伤害案件管理系统:Filevine、Litify、SmartAdvocate 和 CasePeer。公司称可在不到一个工作日内完成完整实施,通过自动同步案件元数据完成上线——客户信息、保单号和接案访谈记录会在上线导入时从 CMS 流入 EvenUp,并在新记录到达时持续更新。 Litify 集成会在文档生成后立刻把生成草稿以直接链接形式送入 CMS,减少诉讼人员切换上下文。Smart Workflows 借助 CMS 连接监控案件状态字段,并在律所定义的条件满足时自动触发提醒和动作(例如距上次治疗经过的时间、缺失记录标记)。Communication Agents™ 会把结果——通话录音、结构化摘要、索赔编号、理赔员分配——直接同步到 CMS 案件文件,无需单独登录。 这种依赖拓扑带来双边风险。数据摄取侧,EvenUp 依赖律所上传记录以及从 CMS 平台同步数据的质量和完整性;接案阶段数据卫生差,会在下游削弱 Piai™ 准确度。平台依赖侧,CMS 供应商可以修改 API、限制第三方集成,或自建竞争性 AI 功能;Filevine 自身 AI 路线图已经显示出这种动态。EvenUp 通过覆盖四大 CMS 平台部分缓释风险,但若某个占主导地位的 CMS 伙伴弃用 API,仍会影响已安装客户基础中相当一部分的服务连续性。 PLAAS 引入第三重依赖:一支 300+ 人的美国案件管理团队,为订阅律所处理诉前工作流。这带来实质性的运营扩展约束——获取新的 PLAAS 客户,需要以高于典型 SaaS 产能扩张的速度招聘、培训并质检法律运营人员。 [CE025, CE026, CE027, CE028]
EvenUp 平台依赖的关键外部依赖和数据流,以及相应风险分类。
风险分类(高 / 中 / 低)是作者基于已发布依赖描述和已知竞争动态的判断,不是正式风险评级。
[CE025, CE031, CE040, CE041]5.4 差异化:自有数据集、反馈闭环与人工审阅层
EvenUp 的核心可防守优势来自自有 PI 数据集、专家标注反馈闭环和人工审阅层;三者叠加,形成通用 AI 竞争者拿不到的数据飞轮。 自有数据集来自早期说服 PI 律所客户贡献匿名化和解与案件结果数据。这个众包语料如今覆盖数十万起伤害案件和数百万份医疗记录,是每个 Piai™ 模型的训练基础。它不只包含公开判决数据,也包含和解前谈判结果与案件策略模式;这些才是 EvenUp 最可防守的竞争差异。 专家反馈闭环每天运转:用户对 Reading Layer 模型输出的数千条纠错(医疗服务方名称标准化、日期修正、治疗分类、账单核对)会直接回流到微调周期。资深 ML 工程师描述了一条有意设计的标注流水线,摄取这些纠错并持续更新各个实体模型。这为专门的 PI 实体带来精度优势——医院系统名称变体、区域保险理赔员行为模式、PI 专属 ICD 代码簇——没有 EvenUp 的真实标注语料,通用 LLM 无法复制。 人工审阅层由 100+ 名美国护士、律师助理、资深理赔员和案件经理组成,会在交付给律所客户前验证 Piai™ 输出。它承担两个功能:质保(在错误到达律所前捕获)和训练信号生成(纠错成为下一轮模型迭代的监督训练数据)。EvenUp 把这一层定位为「远超单靠 LLM 可达到的精确度和准确性」,也是 PLAAS 能提供绩效保证(早期测试追回可用保单限额的 95%)的结构性原因。 客户 ROI 披露虽然全部来自公司,但显示平台确实带来可衡量加速:Jeffcoat Injury Lawyers 需求函产出提升 3×,和解快 30 天;早期 PLAAS 队列调取医疗记录比行业常态快 66 天,交付需求函快 47 天。 [CE029, CE030, CE031, CE032, CE033, CE034]
| 用户任务 | 使用 EvenUp 前的工作流 | EvenUp 方案 | 披露收益 | 限制 / 注意事项 |
|---|---|---|---|---|
| 索赔函准备 | 律师花 3–5 小时审阅记录并起草;不同员工差异大 | Express Demands™ 根据案件记录在数分钟内生成索赔函并提供行级引用;律师复核 | Jeffcoat:索赔函产出量提升 3×;和解提前 30 天(公司来源) | 律师仍需复核;复核走过场会带来执业过失风险 |
| 医疗记录整理 | 律师助理手工编制时间线:每案数小时 | MedChrons™ 在记录上传后数小时内生成经复核的交互式时间线 | EvenUp:Medical Management 时间线可在一小时内生成;未独立验证 | 手写记录和非标准格式的准确率尚无基准测试 |
| 治疗间断监测 | 被动模式:律所通常到临近准备索赔函、索取记录时才发现间断 | Medical Management 标记 30 天治疗间断;Communication Agents 执行回访 | 37% 的回访发现团队本会漏掉的问题(公司来源) | 复杂医疗对话中的沟通智能体准确率未经测试 |
| 保险理赔开立 | 员工致电保险公司;平均每案 16+ 分钟(EvenUp 估计) | Communication Agents 并行自主开立理赔并返回理赔编号 | PLAAS:调取记录快 66 天;每案节省 $1,000(公司来源) | 对抗性保险公司流程下的智能体失败率未披露 |
| 案件策略优先级排序 | 律师每周审阅案件量——手工完成,员工之间不一致 | AI Playbooks™ 在新文件上传时自动标记高价值案件(TBI、商业被告) | EvenUp:每名员工处理 80+ 个案件;AI Playbooks 替代数小时手工阅卷 | 高价值案件误报率未披露;若训练数据过度代表某些案件类型,会有偏差风险 |
| 诉前管理(PLAAS) | 律所内部法律运营团队:调取记录、准备索赔函、谈判分别配人 | PLAAS 托管服务:EvenUp 员工 + Piai 处理完整诉前生命周期 | 早期:95% 保单限额回收;索赔函提前 47 天完成(公司来源,早期队列) | PLAAS 全面部署后的规模和质量尚未验证;仍线性依赖人力 |
所有“披露收益”数字均为 EvenUp 在产品页、博客文章和客户证言中给出的公司来源声明。尚未发现第三方对这些性能指标作独立验证。
[CE005, CE007, CE010, CE011, CE016, CE033]5.5 信任、安全与监管合规
EvenUp 在数万起活跃案件中处理 Protected Health Information(PHI),合规面因此相当重要。其公开 Data Processing Agreement(DPA)记录了个人数据处理的合同框架,并描述了管理 PHI 保护的 Schedule 2 Technical Measures。 EvenUp 承诺签署 Business Associate Agreement(BAA),覆盖在医疗数据场景中运营的律所所承担的 HIPAA 义务。根据 DPA,EvenUp 仅按客户指示处理个人数据,不会向已披露分包处理方以外的第三方出售或共享客户个人数据,并在合同终止时删除数据。安全事件发现后 72 小时内通知,是合同要求。根据 DPA 审计权条款,EvenUp 会应客户要求提供等同 SOC 2 的第三方审计报告。 DPA 记录的技术安全控制包括:逻辑数据隔离和基于角色的访问控制;客户个人数据静态和传输中的加密;企业防火墙和入侵检测基础设施;变更管理流程;业务连续性和灾难恢复流程。DPA 规定,客户身份认证凭证由客户负责。 ABA Model Rule 1.1(Competence)以及各州律师协会关于技术能力的指引,要求律师理解并监督客户代理中使用的 AI 工具。这为 EvenUp 的律所客户制造了结构性合规义务,而不是直接落在 EvenUp 身上;但 EvenUp 的人工审阅层和行级引用系统,是让律师能够监督 AI 输出的主要产品机制。 [CE035, CE036, CE037, CE038, CE039]
| 控制 / 认证 | 状态(May 2026) | 范围 | 尽调缺口 |
|---|---|---|---|
| 数据处理协议(DPA) | 已与所有客户签署 | CCPA 及适用美国州级数据保护法;规范案件记录中 PHI 的处理 | DPA 覆盖美国法律;未处理国际数据传输义务 |
| 业务伙伴协议(BAA) | 可按需提供(DPA 中提及) | 案件文件中医疗记录的 PHI 处理,受 HIPAA 监管 | BAA 条款及分处理方 BAA 链条未公开披露 |
| 安全事件通知(72 小时) | DPA 第 6 节以合同承诺 | 确认客户个人数据泄露后 72 小时内通知客户 | 没有公开的事件披露历史记录 |
| 第三方安全审计(SOC 2 等效) | 根据 DPA 第 7 节,客户可按需获取 | 第三方安全专业人员审计报告;向客户提供保密摘要 | SOC 2 Type II 与 Type I 状态、范围、最近审计日期均未公开确认 |
| 技术安全控制 | 已实施(DPA 附表 2) | 逻辑数据隔离、基于角色的访问、加密、企业防火墙、IDS、事件响应、BCP/DR | 具体云服务商、加密算法标准和渗透测试频率未披露 |
| ABA 胜任能力合规(Rule 1.1) | 客户义务;EvenUp 提供监督机制 | 律师必须复核并监督所有 AI 生成输出;行级引用支持追溯 | 各州律师协会 AI 指引不一;EvenUp 在 AI 规则更严司法辖区的合规立场未确认 |
DPA 和 BAA 引用来自 evenuplaw.com/dpa/ 上公开可得的 EvenUp DPA 文件。SOC 2 状态根据 DPA 审计权利语言推断;直接的 SOC 2 认证尚未独立确认。
[CE035, CE036, CE037, CE038, CE039]5.6 产品风险与反向技术评估
EvenUp 的产品与技术画像包含四类重要风险,需要投资人做超出公司口径业绩主张的尽调。 LLM 商品化是生死攸关的技术风险。EvenUp 的 Writing Layer 能力——文档起草、问答、谈判表生成——可以被 OpenAI、Google 和 Anthropic 的通用模型微调版本部分复制,并应用到 PI 工作流。EvenUp 可防守的护城河在 Reading Layer 和自有 PI 数据集,而不是 Writing Layer 本身。随着基础模型文档理解能力提升、专用微调成本下降,EvenUp 的 Writing Layer 与提示词设计良好的通用 LLM 之间的差距会收窄。Harvey AI、Clio Duo 和 LexisNexis Protégé 已经在把通用 LLM 用于法律文档生成。 数据依赖与准确性风险:Piai™ 的准确度直接取决于上传案件材料的质量。EvenUp 自有分析显示,16.8% 的案件会在三个月内出现 30 天治疗空档,43% 最终会出现这类空档;也就是说,输入数据本身经常不完整,并在下游生成文档中制造准确性风险。根据 ABA Rule 1.1,律师仍对验证 AI 输出承担最终责任,需求函中的一个重大错误就可能让律所暴露在法律过失风险下。 CMS / 平台集中风险:EvenUp 的集成依赖四家 CMS 供应商的 API 访问。这些供应商中任何一家都可能限制 API 访问、推出竞争性 AI 功能,或调整定价并推高 EvenUp 成本。RELX/LexisNexis 的战略投资制造了复杂激励:LexisNexis Protégé 在 AI 辅助法律工作上与 EvenUp 竞争,长期商业条款因此需要追问。 PLAAS 扩展性风险:早期 PLAAS 销售额 $10M+ 显示有需求,但该模型要求美国法律运营人员线性扩张。EvenUp 能否在快速扩大 PLAAS 团队时保持输出质量和盈利能力,仍未在规模化下验证;LawNext 点出了根本问题:PLAAS 是真正的新类别,还是一个重新包装、外包 AI 外壳的托管服务。 [CE040, CE041, CE044, CE045]
5.7 附录
06客户情况
6.1 客户基础分层
EvenUp 的核心客户是美国各类人身伤害律师事务所,从独立执业律师、区域精品所,到大型跨州机构和全国 Top-100 律所。截至 2026 年 5 月,公司服务超过 2,000 家律所;这一数字在 LawNext 的 PLAAS 发布文章、Above the Law 的 $2B 估值报道以及 EvenUp 自有平台信息中得到交叉确认,约占按案件量计算的前 100 家 PI 律所的 30%。客户组合偏向中大型:具名案例研究中的律所拥有 50 到 180+ 名律师和员工(Mama Justice、John K. Zaid & Associates、J&Y Law、ELG Injury Lawyers),而小型律所长尾贡献覆盖广度。 在律所内部,平台服务多个角色。创始律师和管理合伙人决定是否采用;COO 和法律运营负责人推动变革管理与集成;律师助理和案件经理是需求起草、医疗管理和通信代理的日常重度用户;需求函撰写人和和解谈判人使用 AI Drafts 与 Negotiation Sheets。这种多角色部署很重要:它意味着随着单个律所账户员工增加或产品套件扩展,增购和扩张可以在账户内发生,从而降低收入增长对净新增律所获客的依赖。 地理覆盖是全国性的。已记录客户横跨 Oregon 和 Florida(ELG)、Texas(John K. Zaid)、Mississippi、Alabama 和 Tennessee(Mama Justice)、California(J&Y Law、Fielding Law、Lerner and Rowe)及其他市场。多数客户在诉前阶段运营,因为需求函量最高;但诉讼阶段使用也在增长(J&Y Law 使用 AI Playbooks 准备用于证词录取的事实和取证自动化),从而扩大可触达工作流。依赖 CMS 集成获客——Filevine、Litify、SmartAdvocate、CasePeer——把 EvenUp 接入 PI 律所既有技术栈,带来分销倍增,也带来渠道集中风险。 [CU001, CU002, CU003, CU004, CU005, CU006]
| 细分 | 买方 / 用户 / 付款方 | 主要用例 | 律所规模(约) | 收入 / 战略价值 | 关键缺口 |
|---|---|---|---|---|---|
| 前 100 大 PI 律所 | 管理合伙人(买方);COO / 法律运营(运营方);律师助理 / 案件经理(用户) | 全平台——Demands、Medical Mgmt、Playbooks、Communication Agents、Analytics | 50–500+ 名律师 | 单律所收入最高;案件量不成比例地大;品牌可信度 | NRR 未知;头部客户流失会带来集中风险 |
| 中端 PI 律所(11–49 名律师) | 创始律师或运营负责人(买方);案件经理(用户) | 产品:Express Demands、Smart Workflows、Companion、Medical Management | 11–49 名律师 | 核心量级驱动;采用周期更快;多产品扩张潜力 | 公开证据较少;采购有时为单一来源 |
| 小型 PI 律所(1–10 名律师) | 单人律师或小型合伙所(买方兼用户) | Express Demands、基础 Medical Management | 1–10 名律师 | 覆盖广度;市场渗透信号 | 按案收费可能造成预算摩擦;支持带宽有限 |
| 诉前专长律所 | 案件经理和索赔函撰写人员(主要用户);合伙人(付款方) | 产品:Express Demands、Smart Workflows、Settlement Repository | 不一 | 高案件吞吐量;Express Demands ROI 信号强 | 若律所转向纯诉讼模式,可能流失 |
| 诉讼扩张型律所 | 高级律师和庭审团队(用户) | 用于证据开示的 AI Playbooks、MedChrons、用于动议和诉状的 AI Drafts | 不一 | 平台粘性更深;PLAAS 带来更高 ARPU 潜力 | 销售周期更长;集成要求更复杂 |
| PLAAS 早期采用者 | 管理合伙人 / COO(战略买方);EvenUp 员工(运营协作方) | 完整诉前生命周期托管服务 | 法律运营预算 $500K+ 的大型律所 | ARPU 最高;锁定效应最深;早期订阅销售额 $10M+ | 托管服务模式尚未规模化验证;质量控制风险 |
细分定义根据具名案例研究和平台产品描述推断;EvenUp 未披露官方细分分类。收入区间估计为分析推断。
[CU001, CU003, CU008]绘制不同 PI 律所客群如何发现、采用并扩大 EvenUp 在案件生命周期中的使用,从最初的索赔函切入点到采用托管服务 PLAAS。
旅程阶段根据已发布客户案例研究和平台文档推断;EvenUp 尚未发布官方客户旅程或产品采用顺序。
[CU008, CU037]6.2 采用轨迹
EvenUp 已记录的客户增长呈现清晰加速曲线,并与主要融资节点和产品发布绑定。ELG Injury Lawyers 于 2021 年 5 月成为公司第二个客户,构成早期采用拐点。到 2024 年 10 月 Series D 公告时,EvenUp 披露服务 1,000+ 家律所;公司把这一门槛作为规模化产品市场契合的证据。2025 年 5 月 AI Drafts Suite 发布催化下一批客户:EvenUp 报告该发布时服务 1,500+ 家律所,约七个月内客户增长 50%。到 2025 年 10 月,公司已跨过 2,000+ 家律所(Series E 报道与 Unite.ai COO 访谈均确认),这一数字在 2026 年 5 月 PLAAS 发布披露时仍保持。 案件吞吐量与律所数量同步增长。到 2025 年 10 月,EvenUp 已累计解决超过 200,000 起 PI 案件(COO Raymond Mieszaniec,Unite.ai 访谈),当时为受害者追回超过 $10B。截至 2026 年 5 月,每周案件吞吐量为 10,000+ 起,平台累计损害赔偿超过 $14B。这些量级指标显示,平均每家律所的案件负载也随着律所数量增加而提升;这更符合对既有客户工作流的更深渗透,而不只是新律所驱动的增长。 Lightspeed 投资组合介绍确认,EvenUp 在创立两年内达到 $10M ARR,是快速商业化牵引的早期信号。2025 年 7 月,EvenUp 分析了 200 多家律所的数千次 AI 互动,发现 69% 的提示词聚焦医疗治疗分析和伤害评估,30% 的律所进行后续追问;这说明使用是深度、交互式的平台使用,而不是表层采用。2026 年 5 月 PLAAS 发布在早期测试中带来 $10M+ 订阅销售额,在既有按案 SaaS 基础上增加一条托管服务收入向量。 [CU010, CU011, CU012, CU013, CU014, CU015]
| 指标 | 数值 | 日期 / 里程碑 | 来源 | 置信度 | 含义 |
|---|---|---|---|---|---|
| 律所客户数 | 1,000+ | October 2024(Series D) | Bloomberg Law / EvenUp 公告 | 高 | Series E 前最后一个主要披露节点的基线 |
| 律所客户数 | 1,500+ | May 2025(AI Drafts Suite 发布) | EvenUp 新闻稿 / LawNext | 高 | 约 7 个月增长 50%;新定价模式与增长同步推出 |
| 律所客户数 | 2,000+ | October 2025(Series E)和 May 2026(PLAAS 发布) | Unite.ai / LawNext PLAAS 文章 | 高 | 增速较前期放缓;重点转向质量(前 100 大渗透) |
| 前 100 大 PI 律所渗透率 | ~30% | May 2026 | LawNext PLAAS 文章 / EvenUp | 高 | 前 100 大中约 30 家;品牌可信度高,同时收入集中 |
| 累计解决 PI 案件 | 200,000+ | October 2025 | Unite.ai COO 访谈 | 高 | 平均每家律所约 100 个案件;单客户案件深度增加 |
| 每周案件吞吐量 | 10,000+ 个案件 / 周 | May 2026 | LawNext PLAAS 文章 | 高 | 平台覆盖约 $14B+ 损害赔偿额 |
| ARR 里程碑 | 不到 2 年从 $0 到 $10M ARR | 2019–2021 | Lightspeed 投资组合页 | 中 | 早期产品市场匹配信号;仅作历史背景 |
| PLAAS 早期订阅销售额 | $10M+ | 截至 May 2026 发布 | LawNext PLAAS 文章 | 高 | 托管服务扩张已有牵引;能否形成可持续 ARR 还太早 |
EvenUp 不披露 ARR、客户级收入或净留存率。所有客户数和案件量数字都是公司披露指标,见于新闻稿和第三方媒体。“2,000+”最早在 Series E(Oct 2025)时被引用,并在 PLAAS 发布(May 2026)时再次确认;截至 May 2026 的准确数字可能更高,但未公开更新。
[CU010, CU011, CU012, CU013]展示 PI 律所采用 EvenUp 的发现、投产、扩展漏斗,从最初认知到 PLAAS 托管服务合作。
漏斗各阶段量级部分为估计或未披露;只有活跃客户数和 PLAAS 订阅销售额由公司披露。中段漏斗转化率未公开。
[CU001, CU003, CU013]6.3 具名客户验证
EvenUp 至少发布了五个具名客户的详细生产案例研究,每个都给出可验证结果指标。ELG Injury Lawyers 是公司第二个客户(2021 年 5 月),报告在全面采用平台一年内收入增长 400%,并通过 Express Demands 把需求函周转从几个月压缩到几天。ELG 创始人 David Eltringham 直接引用确认,律所在需求起草、通过 Companion 做 AI 驱动反驳,以及用 Medical Management 跟踪护理时间线中都已生产使用。J&Y Law(Los Angeles)用 Companion 和 AI Playbooks 自动化谈判数据收集,每周收回 320 小时案件经理时间,案件生命周期缩短 25%,医疗时间线审阅时间减少 50%。部署接案阶段 AI Playbooks 后,律所月签约案件翻倍。 John K. Zaid & Associates(Houston,180 名律师和员工)在不增加人手的情况下,需求函月环比多发 30%,并在原本报价低于 $10,000 的案件上拿到高出 300% 的和解额。在 2,000 起案件和 7,500 次 AI 执行的客户电话与短信中,EvenUp 的 Communication Agent 主动为 37% 的客户标记急性问题,并识别出 20% 的爽约。Mama Justice(Mississippi / Alabama / Tennessee)实现和解额提高 40%、案件解决快 14%;其 70% 的需求函现在通过 Express Demands 完成,每份耗时 10 分钟,而此前每份需要 1 小时以上。Sweet James PI 律所年结果增长到 $500M+,同比增长 70%;COO Raymond Mieszaniec 把它列为 AI 放大规模、但员工数不成比例增长的旗舰案例。 其他具名验证包括 Fielding Law(California):创始负责人 Clark Fielding 称使用 Medical Management 达到 100% 案件准备度(「我们可以在证词录取中实时调出信息」);以及 Lerner and Rowe 创始合伙人 Glen Lerner,他是早期 PLAAS 客户,称该服务释放了资深员工,让他们专注于更高价值工作。所有结果主张都来自公司发布材料;没有独立第三方审计这些 ROI 数字,这是反向证据部分明确指出的限制。 EvenUp 2025 年 3 月的 Pioneer Awards 项目表彰十家以 AI 驱动创新的 PI 律所,形成正式客户认可与倡导渠道。获奖者获得联合品牌营销、加入 EvenUp Customer Advisory Board 的邀请以及联合活动。这个机制生成公开可见的客户背书,同时加深组织关系,降低头部账户流失风险。 [CU018, CU019, CU020, CU021, CU022, CU023]
| 客户 | 细分 / 规模 | 部署阶段 | 使用产品 | 披露结果 | 来源 / 引用质量 | 限制 |
|---|---|---|---|---|---|---|
| ELG Injury Lawyers(Portland OR / Boca Raton FL 律所) | 中端诉前与诉讼;多办公室 | 生产使用(4+ 年;May 2021 加入,为第 2 位客户) | 产品:Express Demands、Medical Management、Companion | 1 年收入增长 400%;索赔函周转从数月 → 数天 | EvenUp 案例研究;创始人 David Eltringham 直接引述 | 公司发布;收入数字没有独立验证 |
| J&Y Law(Los Angeles CA 律所) | 中大型诉讼;15+ 年律所 | 生产使用(从既有 Demands 关系扩展到全平台) | 产品:AI Playbooks、Medical Management、Companion、Express Demands、MedChrons | 案件生命周期缩短 25%;每周释放 320 小时;医疗审阅时间减少 50%;每月签约案件翻倍 | EvenUp 案例研究;COO Monica Washington Rothbaum 直接引述 | 公司发布;披露指标覆盖面大,需要独立验证 |
| John K. Zaid & Associates(Houston TX 律所) | 大型诉前与诉讼;180 名律师 / 员工 | 生产使用(多产品,包括 Communication Agents、AI Playbooks、Demands) | 产品:AI Playbooks、Medical Management、Communication Agents、Demands、Companion | 每月索赔函增加 30%;低价值案件和解额高 300%;2K 名经 AI 检查的客户中 37% 被标记问题 | EvenUp 案例研究 + LawNext Dec 2025 独立佐证;Zaid 直接引述 | 300% 和解额提升为自报;具体案例($30K vs $10K)属轶事 |
| Mama Justice(Mississippi / Alabama / Tennessee 律所) | 区域性多州;50+ 名员工;5 个办公室 | 生产使用(Express Demands 和全套件) | 产品:Express Demands、Missing Docs Check、Case Companion、Negotiation Prep | 和解金额高 40%;案件解决快 14%;索赔函耗时从 1+ hr 降至 10 min(70% 通过 Express Demands 完成) | EvenUp 案例研究;创始律师 Missy Wigginton 和 CEO Jerry Wigginton 直接引述 | 公司发布;和解结果未获独立验证 |
| Fielding Law(Irvine CA) | 中型诉讼业务;加州 PI | 生产使用(Medical Management) | Medical Management | 案件准备度 100%;庭外取证时实时获取信息 | LawNext 2025 年 12 月文章(独立媒体);Clark Fielding 直接引述 | 仅引用单一产品;未提供案件级指标 |
| Lerner and Rowe(全国) | 大型全国性 PI 律所;百强级别 | 生产使用(PLAAS 早期采用者) | PLAAS 托管服务 | 释放资深员工做更高价值工作;“我们最关键的人反而是最大的支持者” | LawNext 2026 年 5 月文章(独立媒体);Glen Lerner 直接引述 | PLAAS 2026 年 5 月推出;大规模结果数据尚未可得 |
| Sweet James(全国) | 大型全国性 PI 律所;百强级别 | 生产使用(全平台) | Claims Intelligence Platform(全套件) | 年度结果 $500M+;员工数未同比例增加,YoY 增长 70% | Unite.ai COO 访谈(2025 年 11 月);公司来源案例 | 收入数字未经独立审计;“结果”定义不清 |
除 Fielding Law(LawNext)和 Lerner and Rowe(LawNext)外,所有结果数据都来自 EvenUp 发布的案例研究或公司高管陈述。截至 2026 年 5 月,尚无独立第三方审计客户 ROI 主张。“生产使用”部署阶段指律所已在真实案件流程中使用 EvenUp,并有具体运营指标佐证;没有律所自称仅处于试点模式。
[CU018, CU019, CU020, CU021, CU022, CU023]从结果具体性、证据独立性、生产成熟度和留存可见性四个证据质量维度,评估已点名的 EvenUp 客户。
评级是基于来源独立性、指标具体性、部署时长和续约证据可得性的定性评估。未采用标准化评分方法。
[CU028, CU029, CU036, CU025, CU026]6.4 留存与耐久性
EvenUp 没有公开披露 NRR、GRR 或客户流失数据,因此无法直接衡量留存。耐久性只能从代理指标推断:多年客户寿命、既有账户内多产品扩张,以及 Top-100 律所渗透率上升。寿命信号有意义:ELG Injury Lawyers 已经活跃使用超过四年(2021 年 5 月至 2026 年 5 月),J&Y Law 明确从早期只用 Demands 扩张到 EvenUp 全平台套件,说明既有账户增购成功。John K. Zaid & Associates 把 EvenUp 描述为「不只是技术提供商——它是长期伙伴」,Zaid 称公司持续演进的路线图承诺是长期信任和广泛内部采用的基础。 产品架构制造结构性切换成本。PLAAS 把 EvenUp 的美国案件管理人员直接嵌入律所诉前工作流,形成显著高于纯 SaaS 工具的退出门槛。一旦律所把案件生命周期管理迁移到托管服务模式,更换供应商就同时需要技术迁移和运营重建。与 Filevine、Litify、SmartAdvocate 和 CasePeer 的 CMS 集成进一步把 EvenUp 固化在既有技术生态内。2025 年 7 月,EvenUp 对 200+ 家律所的分析发现,30% 会通过后续提示词细化 AI 输出,说明使用是交互式、习惯性的,而不是被动消费。 未披露留存指标是本章的核心尽调缺口。没有 NRR 或 GRR,就无法判断 2,000+ 家律所数量代表高总留存下的净新增,还是快速总新增掩盖了小型律所长尾的显著流失。Pioneer Awards 和 Customer Advisory Board 项目可以缓释头部账户留存风险,但无法说明更广泛客户基础。合同期限、续约率和队列级扩张百分比仍未知。 [CU030, CU031, CU032, CU033, CU034, CU035]
| 指标 | 数值 / 状态 | 分群 | 置信度 | 尽调追问 |
|---|---|---|---|---|
| 净收入留存率(NRR) | 未披露 | 全部 | N/A | 在 Series F 或商业尽调中索取;对标记录系统型 SaaS 120%+ 常规水平 |
| 总收入留存率(GRR) | 未披露 | 全部 | N/A | 要求按律所规模段和获客批次拆分队列级流失 |
| 客户流失率 | 未披露 | 全部 | N/A | 确认 2,000+ 是毛数还是净数;要求 Series D 以来月度新增与流失数据 |
| 多年留存代理指标 | ELG(4+ 年)、Zaid(多年)、J&Y(从既有合作扩展) | 具名头部客户 | 中 | 用全量客户验证留存年限;具名证明存在幸存者偏差 |
| 多产品采用代理指标 | J&Y:7+ 个产品;Zaid:5+ 个产品;ELG:3+ 个产品;Mama Justice:4+ 个产品 | 具名客户 | 中 | 要求披露 2,000+ 全量客户每账户产品数中位数;公司未披露该指标 |
| 平台黏性(PLAAS) | $10M+ 早期订阅销售额;EvenUp 员工嵌入律所工作流 | PLAAS 早期采用者 | 中 | PLAAS 2026 年 5 月推出;跟踪 6 个月和 12 个月订阅续约率 |
| 客户顾问委员会 / Pioneer Awards | 2025 年 3 月表彰 10 家律所;顾问委员会已成立 | 头部客户 | 中 | 顾问委员会成员名单未公开;要求披露构成和获奖客户流失情况 |
| 合同期限 / 续约条款 | 未披露;按案件计费的 SaaS 可能按需终止或年约 | 全部 SaaS 客户 | 低 | 要求披露加权平均合同期限和续约通知期 |
| NPS / CSAT | 未披露 | 全部 | N/A | 要求提供 NPS,并与法律科技同行对标;注意 Trustpilot 截至 2026 年 3 月无评论 |
所有“未披露”条目都反映截至 2026 年 5 月的真实数据缺口;EvenUp 是私营公司,没有披露留存指标的监管义务。留存年限、多产品采用等代理信号偏正面,但样本来自具名高表现客户,不具代表性。
[CU030, CU031, CU032, CU033, CU034, CU035]| 客户 | 加入日期(约) | 截至 2026 年 5 月的服务年限 | 入门产品 | 截至 2026 年 5 月的产品 | 扩张信号 |
|---|---|---|---|---|---|
| ELG Injury Lawyers | 2021 年 5 月 | 4+ 年 | 索赔函起草(早期平台) | 产品:Express Demands、Medical Management、Companion | 服务年限长;多产品扩张;创始人持续站台 |
| J&Y Law | 2024 年前(仅 Demands) | 2+ 年 | 仅 Demands | AI Playbooks、Medical Management、Companion、Express Demands、MedChrons、Negotiation Briefs、面向诉讼的自定义 Playbooks | 深度多产品扩张;COO 平台复盘后明确再次承诺使用 |
| John K. Zaid & Associates | 2024 年(试点 → 生产使用) | 1–2 年 | 怀疑 → 试点 → 全面采用 | 产品:AI Playbooks、Medical Management、Communication Agents、Demands、Companion、Negotiation Sheets | 领导层承诺后,推动内部采用“并非难事”;提到“长期合作关系” |
| Mama Justice | 未说明(暗示多年) | 确认 1+ 年 | Demands(早期) | 产品:Express Demands、Missing Docs、Case Companion、Negotiation Prep | 诉讼用例增长;向更高价值案件扩展 |
| Fielding Law | 2025 年 12 月(Medical Management) | Medical Mgmt 不到 1 年 | Medical Management | Medical Management(引用单一产品) | 新产品早期采用者;显示扩展套件意愿 |
| Lerner and Rowe 律所 | 2026 年(PLAAS 早期测试) | 不到 1 年(PLAAS) | PLAAS | PLAAS | 旗舰 PLAAS 早期采用者;结果验证待定 |
服务年限估计来自公开披露日期;确切合同开始日期未公开。在 NRR/GRR 未披露的情况下,本表用作留存代理指标;具名客户存在幸存者偏差,可能无法反映更广泛 2,000+ 律所客群的流失模式。
[CU029, CU036]6.5 扩张与集中风险
EvenUp 的先落地再扩张动作在所有已记录客户案例中都很清楚:律所从 Demands 或 Express Demands 起步,随着运营信心增强,逐步采用 Medical Management、AI Playbooks、Communication Agents、Companion 和 Executive Analytics。J&Y Law 从只用 Demands 到使用带定制 AI Playbooks 的全平台做诉讼取证,正是这种模式的样本。PLAAS 代表最深的扩张阶段:一种托管服务订阅,让 EvenUp 从软件供应商转为外包运营伙伴;截至 2026 年 5 月,早期测试已带来 $10M+ 订阅销售额。 集中风险存在于多个层面。Top-100 PI 律所渗透率(30%)意味着大量案件量和可能收入集中在大客户中;哪怕流失少数 Top-100 账户,也会对关键指标造成不成比例影响。对 CMS 伙伴的渠道依赖——Filevine、Litify、SmartAdvocate、CasePeer——意味着这些平台的集成政策、定价或竞争定位变化都可能扰乱客户获取管道。EvenUp 自身向 PLAAS 的产品扩张也可能蚕食法律金融伙伴的推荐关系,尤其是托管服务模式与第三方诉前融资服务重叠时。 多平台并用风险是真实存在的。竞争者 Supio 和 Filevine AI 瞄准同一 PI 工作流,并在构建可比的文档生成和案件管理能力。Supio 的博客文章明确批评通用 AI 的准确性,并把 Westlaw 集成定位为差异点;这是一个直接竞争楔子。利润率紧张的小型 PI 律所可能同时采用多个平台,而不是只押注 EvenUp,尤其是在按案定价模型下,把增量案件量放到竞争工具上的切换成本很低。保险理赔员适应也构成额外系统性风险:随着 AI 生成需求函变得无处不在,理赔员可能发展反制手段,削弱律所当前从 EvenUp 数据锚定需求函中获得的谈判溢价。 LawNext 的 PLAAS 发布文章提出了一个尖锐问题:PLAAS 是否「代表一个真正的新类别,还是把既有服务套上 AI 外壳重新包装」;这一定性凸显出尽调要求——在把 $10M 早期销售额视为耐久 ARR 证据前,必须在更长周期内验证 PLAAS 的规模化结果。 [CU037, CU038, CU039, CU040, CU041, CU042]
| 维度 | 风险 / 扩张驱动因素 | 影响等级 | 尽调路径 |
|---|---|---|---|
| Top-100 律所集中度 | ~30 家 top-100 律所;若前 10 大客户占收入 30%+,流失 2–3 家就会显著影响指标 | 高 | 要求按十分位披露收入分布;确认前 10 大客户收入集中度 |
| 先落地再扩张打法 | 具名客户平均采用 4–7 个产品;PLAAS 在 SaaS 之上叠加托管服务层 | 正向驱动 | 按队列测算 NRR;确认产品扩张是否把 NRR 推到 120% 以上 |
| CMS 渠道依赖 | Filevine、Litify、SmartAdvocate、CasePeer 集成带来发现流量;CMS 定价 / 政策变化可能扰乱销售管线 | 中 | 按获客渠道评估收入;确认 CMS 合作伙伴合同条款 |
| PLAAS 模式风险 | 托管服务模式混合 AI 与人工劳动力;质量控制和毛利管理尚未在规模化下验证 | 中-高 | 跟踪托管服务与 SaaS 分群的错误率、客户满意度和毛利率 |
| 多工具并用 / 竞争替代 | Supio、Filevine AI、Clio 瞄准同一工作流;小型律所可能使用多个工具 | 中 | 调查赢单 / 输单原因;跟踪流失客户中的竞争替代率 |
| 保险理算员适应 | AI 生成索赔函增加后,理算员可能部署反制 AI,或给 AI 来源索赔函打折 | 中 | 监测未来 12–24 个月和解结果趋势;跟踪 AI 索赔函溢价是否消退 |
| SMB 采购摩擦 | 按案件计价在规模化时有利,但会给小型律所带来前期成本不确定性 | 中-低 | 按律所规模段分析流失率;面向 10 名律师以下分群提供分层或订阅定价 |
| PLAAS 正当性疑问 | LawNext 质疑 PLAAS 究竟是真正的新服务,还是既有托管服务重新包装;长期价值取决于结果验证 | 中 | 委托第三方在采用后 12 个月研究 PLAAS 队列结果 |
影响等级和风险评级是基于公开信息的分析推断;EvenUp 未披露收入集中度数据或渠道结构。尽调问题代表潜在投资者要化解每项风险所需的最低信息。
[CU037, CU038, CU039, CU040, CU041, CU043]6.6 附录
07风险
7.1 监管与法律风险
EvenUp 处在医疗数据法和律师职业行为规则的交叉口,PLAAS 会让这层层监管暴露进一步加深。 在 HIPAA 下,EvenUp 代表人身伤害律所处理受保护健康信息(PHI)——医疗记录、治疗摘要、账单明细—— 这些数据由律所从受监管的医疗服务提供方取得。这使 EvenUp 成为 HIPAA 商业伙伴, 必须签署商业伙伴协议(BAA),维护行政、物理和技术防护措施,并在发现违规后 72 小时内报告。 自 HITECH Act 以来,HHS OCR 可直接执法商业伙伴;罚款可达每个违规类别每年 $1.9M。 2024 年医疗行业泄露受害者达到 289 million 人——最大单一事件(Change Healthcare)影响 192.7 million 人—— 说明商业伙伴泄露正是 OCR 的重点目标。 律师职业伦理规则又叠加第二层。ABA Model Rule 1.1(能力)要求监督律师理解自己使用的技术; Rule 1.6(保密)限制在未经同意或缺乏运营必要性时向供应商披露客户数据; Rule 5.5 禁止协助未获授权执业。随着 EvenUp 的 PLAAS 把法律运营人员推进到主动案件工作中, 多个州的监管者可能会审查非律师“法律运营”人员是否越界进入法律执业。Texas Bar 和其他州律师协会已经发布伦理指引, 要求律师监督 AI 工具。Mata v. Avianca(S.D.N.Y. 2023)中,法院因律师提交含有虚构判例引用的 AI 生成简报而施加制裁, 说明在律师监督之前出现的 AI 输出错误,会在下游变成职业责任。NIST AI Risk Management Framework 提供了自愿但越来越常被引用的 AI 治理标准, 监管者和法院可能把它用作基准。FTC 对生成式 AI 供应链的竞争审查——尤其是云计算和模型集中度—— 也可能间接影响 EvenUp 的基础设施伙伴。[CR001, CR002, CR003, CR004, CR005, CR006]
| 风险 / 规则 / 案例 | 司法辖区 | 状态 | 可能性 | 严重性 | 缓释措施 | 剩余暴露 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| HIPAA 业务伙伴泄露——OCR 执法,罚款最高 $1.9M/category/yr | 联邦(HHS/OCR) | 持续义务;2024 年医疗健康行业泄露受害者达 289M | 中 | 极高 | 与每家律所客户签署 BAA + DPA;DPA 规定 72h 泄露通知协议;技术保障措施 | 单次泄露即可触发 OCR 调查、客户流失、声誉崩塌 | 核验所有律所客户的 BAA 覆盖;要求 SOC 2 Type II 或第三方审计报告 |
| 律师胜任义务——ABA Model Rule 1.1;监督律师必须理解 AI 输出 | 美国所有司法辖区(ABA 模型规则;多数州已采纳) | 正在执行;多个州律师协会发布 AI 专项指引 | 高 | 高 | 人工审核层;所有索赔函发送前由律师签字 | 若审核流于形式,律师职业过失暴露会转移;EvenUp 品牌受牵连 | 用样本律所客户确认审核协议深度;评估监督人 / 案件比 |
| 客户保密义务——ABA Rule 1.6;与 AI 供应商共享案件数据可能需要知情同意 | 美国所有司法辖区 | 州律师协会伦理意见越来越多地处理 AI 供应商数据共享 | 中 | 高 | DPA 将 EvenUp 的数据使用限制在服务交付;不向第三方出售 / 分享 | 如果律所未记录同意或默示授权分析,可能引发州律师协会投诉 | 核验前 10 大律所客户聘用协议中的律师同意措辞 |
| 未经授权执业——ABA Rule 5.5 及各州同等规则;PLAAS 法律运营员工执行法律任务 | 州级 UPL 法规(50 州) | 截至目前未发现执法行动;PLAAS 于 2026 年 5 月推出 | 低–中 | 高 | PLAAS 定位为“法律运营支持”而非法律意见;需要律师监督 | PLAAS 范围扩大后,UPL 风险上升;AI 辅助法律工作的监管指引仍在演进 | 审查 PLAAS 服务合同范围;评估是否有交付物构成法律意见 |
| AI 幻觉 / 法庭制裁——Mata v. Avianca 先例;律师因未核验 AI 输出被制裁 | 联邦和州法院 | 活跃:法院正在发布 AI 披露和 AI 核验常设令 | 中 | 高 | EvenUp 的 Writing Layer 使用自研模型;人工审核员接受引用事实核查培训 | 如果带有事实错误的索赔函进入诉讼阶段,可能引发律师制裁和职业过失 | 索取 EvenUp 医疗时间线事实核查内部 QA 政策副本 |
| FTC AI 竞争审查——模型提供商集中度风险;可能影响云供应的反垄断行动 | 联邦(FTC) | FTC 正监测 AI 基础设施集中度;未见针对 EvenUp 的直接行动 | 低 | 中 | 多云策略和供应商多元化(如有);公开来源不清楚 | 若 FTC 限制关键模型提供商,EvenUp 的 AI 基础设施成本或访问可能受损 | 尽调 AI 基础设施供应商名单,以及防单边变更的合同保护 |
可能性和严重性是基于截至 2026 年 5 月已发布监管姿态、执法数据和行业先例的定性评估。公开记录中未发现点名 EvenUp 的执法行动。
[CR001, CR002, CR003, CR004, CR005, CR006]以发生可能性 × 影响程度矩阵,呈现 EvenUp 五类主要风险。
可能性和影响为定性估计,依据行业先例、监管指引和公司披露的风险因素。没有可用的量化概率数据。
[CR001, CR010, CR017, CR024, CR029]7.2 运营与质量风险
EvenUp 的核心产品风险在 AI 输出质量:医疗时间线错误、虚构治疗日期、损害赔偿计算虚高或出错, 以及低质量病历扫描件的 OCR 录入失败。公司承认通用 LLM 会产生幻觉,并把人工审核层和自有反馈循环列为缓释措施。 其 2025 年基准报告发现,PI 律所在使用 EvenUp 前平均每 10 个案件遗漏 5 份文件, 而 EvenUp 的 Missing Docs Check 可减少 75%;这也意味着仍有残余错误率,少数案件仍可能被实质性低估。 PLAAS 会放大运营风险:按照公司说法,EvenUp 配备 300+ 名法律运营人员管理外包诉前工作流, 质量控制取决于人工审核员能否随案件量同比例扩张。审核员与案件负载一旦恶化,就会在下游形成系统性错误。 影响 SaaS 平台的云和基础设施故障——认证失败、API 停机、第三方模型提供商中断—— 会让有时效期限的客户律所停下索赔函生产(诉讼时效、理赔员响应窗口)。EvenUp 的 DPA 要求 72 小时内通知泄露, 但不保证正常运行时间。安全事件——例如勒索软件攻击持有 PHI 和机密案件策略的法律科技平台—— 会触发 HIPAA 泄露报告、潜在 OCR 罚款,并严重伤害律所客户中的声誉。[CR010, CR011, CR012, CR013, CR014, CR015]
| 失效模式 | 可能性 | 严重性 | 缓释成熟度 | 剩余暴露 | 未解决缺口 |
|---|---|---|---|---|---|
| 医疗时间线 OCR 错误——手写病历误读,索赔函治疗日期错误 | 中 | 高 | 中:自研 OCR + 每日专家修正进入反馈回路 | 剩余:按公司自有基准,约 25% 缺失文件案件仍会漏掉 | 未见 OCR 错误率公开第三方审计;公司基准为自报 |
| 赔偿叙述中的 AI 幻觉——夸大或编造伤情事实、工资损失计算错误 | 低–中 | 极高 | 中:人工审核层;Writing Layer 基于 PI 数据微调 | 若幻觉通过审核并到达理算员或诉讼阶段,律师可能承担职业过失 | 没有关于 EvenUp 真实输出幻觉率相对既有技术的独立基准 |
| SaaS 平台宕机——诉讼时效窗口内索赔函生产受阻 | 低 | 高 | 低–中:未公开 SLA;DPA 不保证正常运行时间 | 错过截止日期可能损害客户案件;未披露灾备 SLA | 要求提供正常运行时间 SLA 和事故历史;确认故障切换架构 |
| 安全泄露 / 勒索软件——含 PHI 的案件文件暴露;OCR 要求 60 天内通知 | 低 | 极高 | 中:DPA 安全承诺;第三方审计报告可按要求提供 | $1.9M/category OCR 罚款;客户流失;案件主体可能发起集体诉讼 | 未公开披露 SOC 2 Type II 证书;按 DPA §7.2,审计报告保密 |
| PLAAS 审核员质量下滑——法律运营人员增速快于 QA,错误率上升 | 中 | 高 | 低:PLAAS 2026 年 5 月推出;规模化质量基准尚未建立 | 多个案件出现系统性错误;发现滞后;律所声誉和职业过失暴露 | 未公开 PLAAS 质量指标;尽调必须确认审核员 / 主管比 |
严重性分级为定性判断。可能性评级反映当前规模和缓释措施;随着运营规模扩大,PLAAS 相关可能性可能上升。数据来源包括 EvenUp 自有产品博客和 DPA,以及 HIPAA Journal 用于行业基准的泄露统计。
7.3 伙伴与依赖风险
EvenUp 的分发高度依赖少数案件管理平台集成:Filevine、Litify、CasePeer/Clio 和 SmartAdvocate 合计覆盖多数中大型 PI 律所。 任何平台弃用 API、调整定价或签排他安排,都可能打断 EvenUp 的数据摄取管线,并迫使公司付出高成本重新集成。 Filevine 已经搭建自有 AI 层(LOIS),让宿主平台和插件式 AI 供应商之间存在潜在冲突。 EvenUp 的战略投资者之一 LexisNexis/RELX,是 Lexis+ AI(现 Lexis+ Protégé)的母公司; 该产品直接竞争 AI 辅助法律研究赛道,也有分发触达能力扩展到 PI 索赔函生成。 RELX 的多数股权地位制造了利益冲突场景:该投资者可能复用 EvenUp IP 或客户数据学习, 也可能在投资周期结束后直接自建竞争产品。 云和模型提供商集中度是另一层依赖。EvenUp 的 Writing Layer 依赖基础 LLM(很可能是 GPT-4 级或类似模型); 任何模型提供商弃用、涨价,或服务条款变化限制使用 PHI 相邻的法律数据,都会迫使公司付出高成本迁移模型。 FTC 已经指出,少数云提供商主导 AI 计算基础设施,给 AI 初创公司带来系统性集中风险。 EvenUp 客户集中在前 100 大 PI 律所中的 30%,意味着即便流失两三个旗舰账户,也会对渠道背书和增长轨迹造成实质伤害。[CR017, CR018, CR019, CR020, CR021, CR022]
| 依赖 | 对手方 | 角色 | 集中度 | 失效场景 | 严重性 | 缓释措施 | 剩余暴露 |
|---|---|---|---|---|---|---|---|
| CMS 集成——主要数据摄取路径 | Filevine(拥有 LOIS AI) | 案件数据管线;摄取医疗记录和案件元数据 | 高:Filevine 是最大的 PI CMS;LOIS 是直接 AI 竞争者 | Filevine 限制 / 重新定价 API 访问,或构建原生索赔函生成能力 | 高 | 多套 CMS 集成(Litify、SmartAdvocate、CasePeer);不存在单平台锁定 | 若 Filevine 原生捆绑 AI,EvenUp 将同时失去推荐渠道和数据管线 |
| 战略投资者——董事会访问权、IP 可见性 | RELX / LexisNexis(Series D/E 投资者) | 资本提供方和潜在收购方;LexisNexis 在法律 AI(Lexis+ Protégé)领域竞争 | 高:RELX 可能持有董事会观察员席位或董事席位 | RELX 复制 EvenUp 的 PI 数据集;推出竞争性 LexisNexis PI 索赔函产品 | 高 | EvenUp 的自研模型和数据护城河;合同 IP 保护条款(未核验) | RELX 有分销规模,可凭既有 LexisNexis 关系以低价冲击 EvenUp |
| 基础 LLM 供应商——AI Writing Layer | 可能是 OpenAI 或类似前沿模型供应商(未披露) | 索赔叙述的核心文本生成和推理 | 高:未披露;Writing Layer 可能依赖单一供应商 | 模型供应商弃用版本、提高 API 成本,或限制 PHI 相邻法律数据使用 | 中 | EvenUp 的自研模型和微调能降低但不能消除依赖 | 模型迁移需数月;API 成本上涨会直接传导到毛利率 |
| 云基础设施 | AWS、GCP 或 Azure(未披露) | 承载所有 AI 工作负载和含 PHI 案件文件的计算、存储与服务 | 高:AI 工作负载通常因延迟 / 成本选择单云 | 云宕机、涨价,或 AUP 变化限制 PHI 工作负载 | 中 | 标准云 SLA;是否有多区域或多云架构不清楚 | 重大云宕机(例如 AWS us-east-1)可能让 EvenUp 离线数小时;若数据暴露则有 HIPAA 泄露风险 |
EvenUp 未公开披露 LLM 和云供应商对手方名称。未披露依赖的集中度评级基于类似规模 AI 初创公司的行业常态估计。
EvenUp 在 CMS 平台、AI 提供商、数据和投资人上的关键上游依赖。
[CR017, CR018, CR019, CR020, CR021]7.4 人员与执行风险
EvenUp 的三位联合创始人(Rami Karabibar、Raymond Mieszaniec、Saam Mashhad)以及首席科学家 Haixun Wang(ACM Fellow) 带来显著关键人物集中度。任何一位联合创始人离开,都会影响产品愿景、融资可信度和客户信心。 招募专业法律 AI 人才——具备 NLP 专长的 ML 工程师、愿意进入法律运营的前律师、符合 HIPAA 要求的数据标注人员—— 竞争激烈且成本高。EvenUp 自己的工程博客文章(Emre Yamangil、Fatemeh Torabi Asr)显示公司有意识地建设文化、留住人才, 但火热 AI 劳动力市场中的人才流失仍是持续风险。 扩张 PLAAS 需要同步增加受训法律运营审核员。质量恶化是滞后指标:索赔函中的错误可能要到理赔员谈判暴露不准时才浮出水面, 届时系统性缺陷可能已经扩散到大量案件。监督律师带宽是约束因素——律所合伙人必须审核 AI 输出以满足能力义务; 如果 EvenUp 的扩张速度快过客户律所的监督能力,伦理和执业过失暴露就会转嫁给这些律所, 从而拉低 EvenUp 的净推荐值和续约率。[CR024, CR025, CR026, CR027, CR028]
| 角色 / 职能 | 依赖或缺口 | 可能性 | 严重性 | 缓释措施 | 尽调路径 |
|---|---|---|---|---|---|
| 联合创始人(3 人)— Karabibar、Mieszaniec、Mashhad | 产品愿景、融资可信度、客户信任都押在创始人身上 | 低 | 高 | 股权归属安排(推定);C 级高管以下的管理梯队深度(未验证) | 评估非创始人高管团队深度;确认归属悬崖期和留任协议 |
| 首席科学家 — Haixun Wang(ACM Fellow) | 核心 AI 研究路线图;向 AI 人才招聘渠道释放信号 | 低–中 | 中 | 博客文章显示 ML 团队文化较强;已招聘多名资深 ML 工程师 | 确认雇佣合同期限;评估研究路线图文档和继任安排 |
| 法律运营审核员 — PLAAS 规模 | 质控吞吐量;规模化医疗时间线审核准确率 | 中 | 高 | 培训项目;专家每天向 AI 模型反馈的闭环 | 索取 PLAAS 人员爬坡计划、审核员 / 主管比例、与客户约定的错误率 SLA |
| 专门 ML 工程师 — NLP、法律 AI | 人才市场竞争激烈;Harvey 估值 $11B,招聘攻势很猛 | 中 | 中 | 使命驱动文化、股权上行空间;公开博客释放留才信号 | 审阅流失数据;确认未招岗位与计划的差距;查看 Glassdoor/Blind 口碑 |
| 律师顾问 / 在职前律师 | 为 PLAAS 产品和审核员培训提供专业能力与伦理指引 | 低 | 中 | 产品博客称,公司配有前律师和律师助理 | 核验人数和资质;确认其在审核员质量监督中的职责 |
可能性和严重性均为定性判断。EvenUp 没有公开可用的员工流失数据。PLAAS 审核员人数来自公司说法,未获独立验证。
7.5 财务与模型风险
EvenUp 是私营公司,不披露 ARR 或烧钱速度,投资者视线受阻。公司已按 $2B+ 估值累计融资 $385M(2025 年 10 月 Series E 轮), 资本底座达到 $385M;但 PLAAS 需要配备 300+ 名法律运营人员,本质上比纯 SaaS 更吃资本和人头,压缩毛利率。 如果毛利率落在 50% 以下(对托管服务混合模式来说并非不合理),公司就需要相当大的收入规模, 才能继续支撑研发和 AI 基础设施。 按案定价让收入波动绑定 PI 案件量,而案件量又与经济周期、诉讼趋势和保险理赔员做法相关。 如果保险公司通过系统性折价或更严格审查 EvenUp 式文件来适应 AI 生成索赔函,每案收入可能被侵蚀。 来自 Harvey(2026 年估值 $11B)、Anthropic 和 OpenAI 进入法律垂直领域,以及 Thomson Reuters(CoCounsel)和 LexisNexis 这类资本充足在位者的竞争, 都可能压缩市场价格。未来融资风险真实存在:估值回调或 AI 板块下行,可能削弱公司按当前条款融资的能力; 公司也没有披露盈利路径。[CR029, CR030, CR031, CR032, CR033, CR034]
7.6 缓释、监控与终止标准
EvenUp 已经布置了几项结构性缓释措施:(1)BAA 和公开 DPA,覆盖 CCPA 与数据处理义务,并承诺 72 小时内通知泄露; (2)在 AI 输出和客户交付之间插入人工审核层(法律运营 + 律师签字);(3)自有反馈循环每天吸收数千条专家修正,用于重新训练领域模型; (4)专注 PI 垂直(而不是通用法律 AI),相较通用 LLM 降低幻觉率;(5)多项 CMS 集成,降低单一平台锁定。 即便有这些缓释,几项投资逻辑失效触发点仍会实质削弱投资案例:(a)确认发生大规模 HIPAA 泄露,引发 OCR 罚款和客户流失; (b)法院制裁或律师协会投诉点名 EvenUp AI 输出是律师不当行为的近因,形成反向先例并触发律所级禁用; (c)Filevine 或 Clio 转向原生捆绑竞争性索赔生成 AI,在进件入口切走 EvenUp; (d)RELX/LexisNexis 利用董事会访问权限复制 EvenUp 的 PI 数据集并以 LexisNexis 品牌推出竞争产品; (e)PLAAS 扩张后毛利率长期低于 40%,说明业务无法达到类 SaaS 经济性。监控指标包括点名 EvenUp 的 OCR 投诉文件、 律师协会明确提示 AI 索赔函服务风险的指引、CMS 平台定价或 API 政策变化,以及 RELX 关于 PI 产品扩张的公开表态。[CR035, CR036, CR037, CR038, CR039, CR040]
| 风险 | 可监控触发信号 | 阈值 / 事件 | 行动含义 |
|---|---|---|---|
| HIPAA 数据泄露 | HHS OCR 泄露门户公开列名 EvenUp;或媒体报道 EvenUp 数据事件 | 任何已向 OCR 报告、影响 500+ 人的确认泄露 | 全面暂停尽调;评估影响范围、客户通知义务、OCR 和解风险;考虑退出或重新定价 |
| AI 幻觉导致律师被制裁 | 法院命令、律师协会投诉或公开判决意见点名 EvenUp 输出是制裁原因 | 任何公开的不利裁定明确引用 EvenUp 索赔函内容 | 立即复查产品;重做律师监督流程;投资人可能面临责任敞口 |
| CMS 平台捆绑竞品 AI | Filevine、Clio 或 Litify 在产品路线图中宣布原生索赔函 AI | 前三大 CMS 平台任一家正式发布竞品原生功能 | 模型转向情景:评估 EvenUp 的工作流优势是否耐久;重新评估护城河 |
| RELX/LexisNexis 产品冲突 | LexisNexis 大规模推出面向 PI 的索赔函或法律运营产品 | 公开产品公告披露拿下 PI 律所客户 | 评估董事会观察员权利和 IP 保护条款;评估战略替代方案 |
| 毛利率压缩至 40% 以下 | 任何披露的财务数据表明 PLAAS 服务成本超过 PLAAS 收入的 60% | 毛利率连续两个季度低于 40% | 重新评估 PLAAS 单位经济;判断 SaaS 分层定价能否抵消服务拖累 |
| AI 法律工具监管规则制定 | ABA、州律师协会或联邦机构发布有约束力规则,要求律师披露或限制 AI 生成法律文件 | 有约束力规则,合规期限在 12 个月内 | 评估合规成本;判断规则会形成护城河(合规级产品)还是构成障碍 |
否决标准不是数学确定性;每个触发信号都需要按严重性和可恢复性做定性判断。阈值只是供投资人复核的指示性监测阈值。
EvenUp 的主要风险向收入、利润率和估值传导的路径。
[CR005, CR018, CR030, CR031, CR034, CR035]7.7 展示材料
08估值
8.1 投资建议与估值立场
EvenUp 作为专为人身伤害打造的 AI 赛道领导者,位置有防御性;但财务画像不透明,让投资决策蒙上阴影。 公司累计披露股权融资 $385M,最近一轮是 2025 年 10 月由 Bessemer Venture Partners 领投的 $150M Series E 轮, 估值超过 $2 billion。任何公开来源都没有出现 ARR、毛利率、EBITDA、烧钱速度或客户净收入留存。 唯一具体收入数据点,是 EvenUp 在 2026 年 5 月产品发布时自行宣布的 $10M+ PLAAS 早期订阅销售; 这更能证明管线验证,而不是证明规模。 建议是继续研究。置信度为中:产品证据和品类领导力很强,但没有披露单位经济性,估值无法做压力测试。 风险评级为高,主要由三点驱动:300+ 名美国本土法律运营人员造成的毛利率压缩(按标准美国法律运营薪酬估计年人工成本 $45–75M)、 HIPAA 和律师职业伦理规则下的监管暴露,以及 PLAAS 可能代表公司向服务重型经济性发生结构性转向, 从而长期压低软件式毛利率。 估值立场为昂贵。$2B+ 投后估值要求至少满足一项:(a)~$100M+ ARR 且增长 50%+; (b)存在可信路径,通过 IPO 或战略收购实现 $10B+ 退出;或(c)随着赛道成熟,法律 AI 倍数出现实质重估。 截至 2026 年 5 月,现有公开证据无法验证任何一项。Bessemer 面向法律 AI 的 Roads to $100M ARR playbook 暗示该赛道能够诞生很大的软件公司, 但 EvenUp 转向托管服务带来纯软件同行没有的成本结构不确定性。Lightspeed 投资组合页面披露 EvenUp 创立不到两年就达到 $10M ARR—— 说明早期速度很快——但那是在 PLAAS 之前、成本结构更轻的公司阶段。 投资纪律上的含义是:新投资者若按 $2B+ 进入,应在承诺前要求经审计的 ARR、毛利率和烧钱速度披露; 定价时也应按 50–65% 的混合毛利率(而非纯 SaaS 的 70–80%)来反映 PLAAS 的托管服务成分。 在计算回报情景前,还必须把 $385M 优先股资本和未量化期权池授予带来的稀释建模进去。 [CV001, CV002, CV003, CV004, CV005, CV006]
| 维度 | 评估 | 依据 |
|---|---|---|
| 投资建议 | 继续研究 | 产品/市场匹配强;单位经济不透明,无法直接放行 |
| 信心 | 中 | 品类领导地位证据充分;ARR / 毛利率未披露 |
| 风险评级 | 高 | 毛利率压缩、监管风险、AI 商品化、$385M 优先股压力 |
| 估值立场 | 昂贵 | $2B+ 私募估值;没有披露 ARR / EBITDA,无法验证倍数 |
| 决策含义 | 以财务披露为条件 | 承诺前要求经审计 ARR、毛利率和 PLAAS 单位经济 |
评估基于截至 2026-05-20 的公开证据;公司没有披露财务指标。
[CV001, CV002, CV003, CV004]从规模、证据、风险和估值证据推导到继续研究建议的链条。
流程节点位置为示意;连接权重未量化。
[CV001, CV004]投委会可用评分,覆盖市场、证据、护城河、经济性、风险、估值和证据质量维度。
评分为分析师判断(1=差,5=优秀),依据截至 2026-05-20 的公开可得证据。
[CV001, CV002, CV003]8.2 当前融资背景与进入纪律
EvenUp 的融资历史横跨从种子轮到 Series E 的五轮已披露融资,其中 Series D($135M,2024 年 10 月宣布,2024 年 6 月公开报道,估值 $1B+) 和 Series E($150M,2025 年 10 月,估值 $2B+)在 Bloomberg Law、EvenUp 自家博客以及多家行业出版物中相互印证。 GlobeNewswire 的 Series D 新闻稿和 EvenUp 2024 年 10 月 8 日博客都确认了 $135M 融资、$1B+ 估值, 以及当时累计资本 $235M。Series E 将已披露股权融资总额提高到 $385M。 对 “EvenUp, Inc.” 做 EDGAR 全文搜索,会返回四份 ServiceTitan DRS/A 文件(CIK 0001638826), 其中把 EvenUp, Inc. 标识为一家 Delaware 私营公司,并自 2023 年 4 月起有董事会代表——尚未识别出 EvenUp 自己独立提交的 Form D, 这与 Regulation D 豁免不要求提交 Form D,或实体名称变体尚未定位相一致。 $2B 估值意味着大约 12 个月内较 $1B+ Series D 抬升约 100%。Series E 参与方包括 Bessemer(领投)、 REV(RELX/LexisNexis 的 VC 部门)、B Capital,以及从前几轮延续的 Bain Capital。 RELX/REV 的参与具有战略含义:LexisNexis 母公司 RELX 在法律 AI 研究工具(Lexis+ AI,现 Lexis+ Protégé)上有独立竞争利益。 这笔投资一方面给 EvenUp 带来战略选择权信号——未来可能被收购;另一方面也引入利益冲突, 需要从治理和 IP 条款上审查。 $385M 优先股堆叠带来有意义的清算优先权压力。如果 EvenUp 以 $1B 被收购, 单纯优先股清算优先权很可能吸收全部交易对价,具体取决于未公开的参与权和转换机制。 按 Series E 价格($2B+ 估值)进入,需要可观退出溢价才能产生风险投资级回报。 联合创始人明确提到 IPO 是一种可能性(据 Axios 2025 年 10 月报道),说明管理层在朝公开退出管理公司; 但截至 2026 年 5 月,没有 S-1 或 IPO 流程启动。Filevine、Litify、CasePeer/Clio 和 SmartAdvocate 是主要案件管理分发集成; 上市前,任何平台排他或定价变化都可能削弱收入可预测性。 [CV010, CV011, CV012, CV013, CV014, CV015]
| 维度 | 乐观论点 | 悲观反驳 | 哪些证据会改变判断 |
|---|---|---|---|
| 市场领导力 | PI AI 品类领导者;2,000+ 家律所、前 100 大 PI 律所渗透率 30%、200,000+ 起案件 | 集中在美国单一执业领域;Harvey 或 OpenAI 可扩张到 PI | Harvey 或 OpenAI 推出 PI 产品,索赔函质量相当且单案价格更低 |
| PLAAS 模型 | $10M+ 早期订阅销售验证托管服务需求;把固定劳动力转成收入 | 300+ 名美国法律运营员工形成 $45-75M 年成本底线;毛利率不透明 | 披露的 PLAAS 毛利率低于 45%,或第 1 年采用 PLAAS 的律所少于 50 家 |
| 数据护城河 | 自研 Piai 模型用数十万起 PI 案件和 $14B+ 赔偿额训练 | LLM 提供商(OpenAI、Anthropic)快速商品化基础能力;护城河变窄 | 竞品不靠自有数据集也能以 <$50/案做出相当输出质量 |
| 退出 / IPO | 联合创始人称 IPO 在考虑范围内;RELX 投资带来 M&A 可选项 | 未提交 S-1;$385M 优先股资本让普通股股东的分配瀑布更复杂 | IPO 推迟到 2028 年以后,或战略收购方退出;迫使公司降轮 |
乐观 / 悲观论点基于公开报道证据和已披露指标。估计成本数字由分析师推导,并非公司披露。
[CV005, CV006, CV007, CV008]8.3 可比估值分析
EvenUp 位于法律 AI(偏软件)和法律托管服务(人力资本密集)的交叉点,同行选择很难。 最相关的可比集合横跨专用法律 AI 初创公司、法律工作流 SaaS 平台、M&A 参考点和垂直 SaaS 类比对象。 所有可比都只具方向性;EvenUp 财务不透明,无法做严谨收入倍数分析。 Harvey AI(通用法律 AI,据 Bloomberg Law 报道 2025 年 2 月 $300M Series D 后估值约 $3B)是最接近公开市场的类比对象, 同为后期私营法律 AI 公司;但 Harvey 服务 BigLaw 和企业客户,而不是 PI 原告律所, 收入模型没有托管服务成分,也不大规模处理 PHI。Harvey 可比说明,市场会给有清晰企业牵引力的专用法律 AI 平台高溢价倍数; 但它解释不了 EvenUp 的软件 / 服务混合经济性。Harvey 估值 $3B,比 EvenUp 的 $2B+ 高 50%, 虽然服务的可触达市场不同且可说更宽;这意味着在调整市场差异后,EvenUp 可能按相对规模与 Harvey 持平或小幅折价估值。 Casetext/CoCounsel(法律研究 AI,2023 年 6 月被 Thomson Reuters 以 $650M 收购)是相关 M&A 参考。 $650M 的收购价把一家尚未达到 ARR 规模、尚未盈利的法律 AI 公司定在很高溢价上——但远低于 EvenUp 当前 $2B+ 估值。 Casetext 主要做法律研究(运营复杂度较低、无 PHI、无法律运营人员),而 EvenUp 的集成式 PLAAS 产品结构不同, 可以说更难复制,但也更难估值。Casetext 退出验证了法律 AI 的 M&A 退出选择权,但也给出一个更低水位; EvenUp 的估值已经超过它 3x。 Clio(法律执业管理 SaaS,据 LawNext 报道 2021 年 Series F 估值 $3B)是最突出的独立法律工作流 SaaS。 Clio 广泛服务普通律所,披露过收入增长和盈利里程碑,并在约 $100M+ ARR 时达到 $3B 估值。 EvenUp 在未披露 ARR 的情况下估值 $2B+,意味着三种可能之一:(a)ARR 规模相当或更大; (b)更高增长倍数反映更快增长;或(c)市场为 Clio 没有的托管服务叠加付溢价。 Clio 参考给出一个粗略“可比 SaaS”基准:$3B 估值约为 30x ARR。 ServiceTitan(家居 / 商业服务垂直 SaaS,据 SEC DRS/A 文件,2024 年 12 月 IPO 估值约 $8B)是相关垂直 SaaS 类比。 ServiceTitan 运营复杂度高于纯软件 SaaS——它嵌入现场服务工作流——IPO 时估值约为前瞻收入的 12–16x。 将 ServiceTitan 倍数应用到 EvenUp 估计 ARR,意味着 EvenUp 的 $2B+ 估值需要 ARR 处在 $125–170M 区间、收入倍数为 12–16x; 这一数字显著高于任何公开披露的 EvenUp 收入指标,更像管理层预测,而不是已确认实际值。 Grand View Research 法律 AI 市场预测(2024 年 $1.45B,到 2030 年增至 $3.90B,CAGR 17.3%) 和 MarketsAndMarkets 法律 AI 软件预测(2025 年 $3.11B,到 2030 年 $10.82B,CAGR 28.3%) 都显示市场顺风强劲;但 EvenUp 的 $2B+ 估值已经把一个年轻、分散市场中的重要份额嵌入进去。 按较高 CAGR 估计,EvenUp 需要到 2030 年在约 $3–11B 的法律 AI 软件市场中拿下并守住有意义份额, 才能证明当前隐含退出估值合理。 [CV018, CV019, CV020, CV021, CV022, CV023]
| 可比对象 | 指标 / 事件 | 估值 / 价格 | 与 EvenUp 的相关性 | 关键局限 |
|---|---|---|---|---|
| Harvey AI | Series D,Feb 2025 | 融资 $300M,对应 ~$3B | 最接近的后期私有法律 AI 可比公司;投资人验证强 | 面向 BigLaw 的通用 AI,不专注 PI;没有托管服务组件;不处理 PHI |
| Casetext / CoCounsel | 被 Thomson Reuters 收购,Jun 2023 | $650M 收购 | 法律 AI 的 M&A 退出参照;验证战略收购方胃口 | 未达到规模化,仅做法律研究;EvenUp 估值已超过其 3x |
| Clio | Series F / 最近一轮融资 ~2021 | ~$100M+ ARR 对应 ~$3B 估值 | 规模化法律业务管理 SaaS;~30x ARR 倍数提供收入倍数锚点 | 通用法律业务管理 SaaS,不专注 PI;没有托管服务层;并非 AI 优先定价 |
| ServiceTitan | IPO,Dec 2024 | IPO 时企业价值 ~$8B | 垂直 SaaS IPO 参照;12-16x 前瞻收入倍数提供公开市场锚点 | 面向现场服务,不是法律;监管画像不同;IPO 后流动性不适用于 EvenUp |
| EvenUp | Series E,Oct 2025 | $150M 融资,对应 $2B+(累计 $385M) | 标的公司;按披露估值,为 Casetext 的 3x、Harvey 的 67%、Clio 的 67% | 未披露 ARR / 毛利率;隐含倍数区间取决于 ARR 假设,为 13-40x |
Harvey $3B 来自 Bloomberg Law 报道(Feb 2025);Casetext $650M 来自多家新闻来源;Clio $3B 来自 LawNext / 多家来源;ServiceTitan IPO 来自 SEC DRS/A 文件。所有倍数均为估计或第三方报道;Harvey、Casetext 退出时或 EvenUp 均没有确认 ARR。
[CV018, CV019, CV020, CV021, CV022, CV023]低 / 基准 / 高退出估值,以及乐观、基准和悲观情景下 Series E 回报倍数示例。
所有估值均为分析师推导的估计,依据可比交易、增长轨迹和 PLAAS 采用假设。公司未披露财务目标。
[CV029, CV030, CV031]8.4 乐观 / 基准 / 悲观情景与估值敏感性
三个情景框住截至 2026 年 5 月的 EvenUp 投资案例。所有情景都基于估计; 没有公司披露的 ARR、毛利率或增长率可用来锚定数字。收入估计来自 $385M 累计融资、 已披露增长轨迹(Series D 时 YoY >100%)、Lightspeed 发布的 ARR 轨迹(不到两年从 $0 到 $10M), 以及 PLAAS $10M+ 早期订阅数字。 乐观情景下,PLAAS 快速扩张,足以覆盖 300+ 名法律运营人员的人头成本, 到 2026 年底推动混合 ARR 达到 $150–200M,毛利率 65–70%。在这一情景中, 收入增长 80%+ 会把收入倍数压缩到 10–13x,与 IPO 时高增长垂直 SaaS 相当。 通过 IPO 或战略收购(最可能是 RELX/LexisNexis、Thomson Reuters 或大型保险公司)以 $4–6B 企业价值退出, 会让 Series E 进入价产生 2–3x 回报。来自 200,000+ 个案件和 $14B+ 损害赔偿的数据护城河, 会成为商品化 LLM 提供商难以复制的结构性壁垒。 基准情景下,PLAAS 到 2026 年底在 50–100 家企业级 PI 律所中获得牵引, 产生 $50–80M 混合 ARR,毛利率 55–60%。随着 PLAAS 销售周期拉长、 按案 SaaS 平台在当前前 100 大律所客群中趋近饱和,收入增长放缓到 40–60%。 按 $3–4B 企业价值 IPO(考虑当前市场条件,2027–2028 年现实可行), Series E 可产生 1.5–2x 回报。托管服务成分在收入中占比上升后,倍数压缩风险增加。 悲观情景下,受律师-客户特权顾虑、HIPAA 合规摩擦,以及大型 PI 律所自建还是购买的竞争决策影响, PLAAS 采用速度慢于预期。ARR 停滞在 $40–60M。与此同时,AI 商品化周期——OpenAI 法律 AI 工具、 Filevine 的 LOIS AI,以及 Harvey 可能扩展到 PI——压缩按案 SaaS 产品的定价权。 $45–75M 年度法律运营成本造成每年 $20–40M 经营亏损。 以 $800M–$1.2B(低于 Series E 估值)战略退出,或下轮降价融资,都会让 Series E 投资者亏损。 估值敏感性分析集中在三个驱动因素:(1)PLAAS ARR 爬坡速度,(2)混合毛利率(由法律运营成本吸收情况驱动), 以及(3)市场退出倍数。在 $100M ARR、10x 收入倍数下,毛利率变化 10 个百分点(例如 55% 对 65%) 会让隐含企业价值变化 $300–500M。毛利率敏感性是主导风险因素,也是投资案例中最重要、尚未验证的变量。 [CV027, CV028, CV029, CV030, CV031, CV032]
| 情景 | 关键假设 | ARR 区间(估计) | 退出估值区间(估计) | Series E 回报(估计) | 概率信号 |
|---|---|---|---|---|---|
| 乐观 | PLAAS 扩至 100+ 家律所,毛利率 65-70%,2027-2028 年 IPO 或战略退出 | 2026 年底 $150-200M | $4-6B 企业价值 | $2B 进入估值下 2-3x | 取决于 PLAAS 采用速度跑赢计划且毛利率得到确认 |
| 基准 | PLAAS 拿下 50-100 家律所,毛利率 55-60%,2027-2028 年 IPO | 2026 年底 $50-80M | $3-4B 企业价值 | $2B 进入估值下 1.5-2x | 需要 PLAAS 温和爬坡,倍数也只是温和压缩 |
| 悲观 | PLAAS 采用缓慢,AI 商品化压缩单案价格,ARR 持平 | 2026 年底 $40-60M | $800M-1.2B 退出(低于 Series E 进入估值) | $2B 进入估值下亏损 | 由竞品价格压力或重大 HIPAA / 伦理监管事件触发 |
所有 ARR 和估值均由分析师根据公开数据点推导;公司没有披露收入数字。回报估计假设简单清算优先权,不包括参与型优先股。
[CV027, CV028, CV029, CV030]在 15x 收入倍数下,隐含企业价值对 ARR 水平和毛利率假设的敏感性。
EV = ARR × 15x(基准倍数);实际倍数随增长率和可比公司变化。所有 ARR 数值均为分析师估计;公司未披露 ARR。
[CV027, CV028]8.5 投资逻辑失效触发点与最终尽调要求
EvenUp 投资逻辑建立在三项结构性主张上:(1)EvenUp 是 PI 法律 AI 中可持续的品类领导者,拥有自我强化的数据护城河; (2)PLAAS 创造一条新的高价值收入线,把固定人工成本转化为订阅收入;(3)公司存在可信路径, 可以通过 IPO 或战略退出,以足以支撑 $2B+ Series E 价格的估值实现退出。 三项主张都与现有公开证据一致,但没有机密财务披露,任何一项都无法独立验证。 主要投资逻辑失效触发点包括:(a)PLAAS 毛利率低于 45%,说明托管服务成分无法在规模化时吸收法律运营人工成本; (b)前 100 大 PI 律所年客户流失率超过 15%,表明产品锁定弱于公司表述; (c)竞争进入者(Harvey、OpenAI 或 Filevine)以显著更低成本推出实质可比的 PI 索赔函工作流, 削弱 EvenUp 按案定价权;以及(d)EvenUp 或使用 EvenUp AI 输出的客户律所发生重大 HIPAA 泄露或律师伦理执法行动, 触发监管和声誉伤害并加速客户流失。 Supio AI 竞争对手博客明确指出法律 AI 的幻觉风险,以及绑定来源核验的重要性——这也是 EvenUp 索赔函和医疗时间线输出面临的同一质量风险。 虽然 EvenUp 声称人工审核层是其质量缓释措施,但 Mata v. Avianca 制裁案(S.D.N.Y. 2023)已经确立: 当输出错误在法庭浮出水面时,律师监督并不能让 AI 供应商免受声誉伤害。 一次高知名度、对 EvenUp 生成索赔函的对抗性法律挑战,就可能加速客户审查并拉长 PLAAS 销售周期。 最终尽调要求包括:经审计的 ARR 和毛利率披露(至少覆盖 Series E 交割前过去 12 个月); PLAAS 单位经济性(每家律所每年收入、每个 PLAAS 订户毛利率、每份合同所需法律运营人头,以及每名员工案件产出); SaaS 和 PLAAS 两条线的客户净收入留存;RELX/REV 投资的治理条款(竞业限制、数据共享权、优先要约权); 以及 Piai 模型的 IP 归属和授权结构(尤其是任何第三方 LLM 提供商依赖、微调数据权利,以及控制权变更后仍可存续的可迁移条款)。 没有这五组数据,现有证据无法支持在 $2B+ 估值下完整放行。 [CV036, CV037, CV038, CV039, CV040, CV041]
| 触发信号 | 阈值 | 对投资逻辑的传导 | 行动含义 |
|---|---|---|---|
| PLAAS 毛利率低于阈值 | < 45% PLAAS 收入混合毛利率 | 托管服务组件无法自我覆盖法律运营团队成本;软件经济性逻辑失效 | 暂停或缩小仓位;按服务公司倍数(3-5x 收入)重新承销 |
| 前 100 大 PI 律所流失率飙升 | > 15% 前 100 大 PI 律所队列年流失率 | 数据护城河和切换成本叙事失效;竞争替代已经启动 | 立即要求披露 NRR;若无法进入可接受的数据室,降低仓位 |
| 竞品 PI AI 发布 | Harvey、OpenAI 或 Filevine 部署质量相当的 PI 索赔函能力,价格 < $50/案 | 单案定价权被侵蚀;EvenUp 在前 100 大律所的份额承压 | 压力测试收入模型;若竞品签下 EvenUp 前 10 大参考客户律所,则重新评估 |
| 重大 HIPAA 泄露或伦理执法 | OCR 执法行动或律师制裁可追溯到 EvenUp AI 输出 | HIPAA 监管风险兑现;企业级 PI 市场可能冻结 PLAAS 采用 | 若泄露影响 > 1,000 名客户,或监管制裁跨多个州,则全额退出持仓 |
| RELX 投资人冲突显性化 | RELX 宣布竞品 PI AI 产品,并使用来自 EvenUp 董事会接触的数据洞察 | 竞争护城河受损;数据权利和治理结构需要复核 | 法律审查投资条款;与独立董事接触 |
阈值是分析师建议的尽调基准,不是公司披露目标。
[CV036, CV037, CV038, CV039]| 主题 | 缺失证据 | 重要性 | 尽调路径 |
|---|---|---|---|
| ARR 与收入构成 | 经审计的过去 12 个月 ARR,并拆分单案 SaaS 与 PLAAS 订阅 | 没有 ARR,无法计算或对标 $2B+ 估值倍数 | 要求进入数据室;要求经审计财报或 Big-4 商定程序 |
| 毛利率与 PLAAS 单位经济 | 混合毛利率;PLAAS 每合同年收入;每个 PLAAS 订户对应的法律运营成本 | 300+ 名美国法律运营员工意味着 $45-75M 年成本;若 PLAAS 毛利率 < 50%,软件经济性逻辑失效 | CFO 单位经济说明材料;对标法律流程外包可比公司 |
| 客户净收入留存 | 单案 SaaS 年度 NRR;前 100 大 PI 律所队列流失率;PLAAS 试点转合同率 | NRR > 120% 可验证增购逻辑;< 100% 则推翻留存主张 | 客户背调电话(5-10 家大型律所);审阅合同续约数据 |
| RELX/REV 治理条款 | REV 投资文件中的竞业限制条款、数据共享权和优先要约权 | RELX 拥有 LexisNexis;条款可能限制 EvenUp 接触其他战略收购方 | 索取并审阅投资人附函;独立法律审查共同投资人权利 |
| Piai 模型 IP 与依赖 | 第三方 LLM 提供商合同、微调数据所有权、控制权变更时的可迁移性 | 若模型被 OpenAI 或类似供应商锁定,收购或 IPO 时会产生成本和 IP 风险 | 技术尽调模型架构;审阅 LLM 提供商 MSA 和数据处理附录 |
尽调要求按对估值判断的重要性排序。第 1 和第 2 项为阻断项;第 3-5 项为重要项。
[CV040, CV041, CV042, CV043]8.6 展示材料
免责声明
本报告仅基于截至 2026-05-20 收集的公开可得信息。EvenUp 未审阅或批准本分析。报告仅用于支持尽职调查,不构成投资建议。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | EvenUp was founded in late 2019 in San Francisco, California. | 高 | SO001, SO002, SO009 |
| CO002 | EvenUp has three co-founders: Rami Karabibar (CEO), Raymond Mieszaniec (COO), and Saam Mashhad (Chief of Product & Legal Operations). | 高 | SO002, SO009, SO010 |
| CO003 | EvenUp serves more than 2,000 personal injury law firms as of May 2026. | 高 | SO001, SO012, SO013 |
| CO004 | EvenUp has approximately 30% penetration among the top 100 US personal injury law firms as of May 2026. | 中 | SO012 |
| CO005 | EvenUp has raised $385 million in total disclosed capital as of October 2025. | 高 | SO013, SO004 |
| CO006 | EvenUp closed a $150 million Series E in October 2025 at a valuation above $2 billion, led by Bessemer Venture Partners. | 高 | SO013, SO007, SO011 |
| CO007 | EvenUp closed a $135 million Series D in June 2024 at approximately $1 billion valuation. | 中 | SO021, SO027 |
| CO008 | EvenUp has cumulatively resolved more than 200,000 personal injury cases. | 中 | SO009, SO011 |
| CO009 | EvenUp has processed more than $14 billion in personal injury damages cumulatively as of May 2026. | 中 | SO012 |
| CO010 | EvenUp holds SOC 2 Type II and HIPAA compliance certifications. | 高 | SO001, SO007, SO009 |
| CO011 | EvenUp was named to the Forbes Cloud 100 list in 2025. | 高 | SO017, SO005 |
| CO012 | EvenUp's product suite includes the Claims Intelligence Platform, Piai model, AI Drafts Suite, Smart Workflows, Medical Bill Summary, Mirror Mode, Case Companion, Voice Agent, Express Demands, AI Playbooks, Medical Management, and PLAAS. | 高 | SO007, SO006, SO009 |
| CO013 | EvenUp processes approximately 10,000 personal injury cases per week as of May 2026. | 中 | SO012 |
| CO014 | EvenUp launched PLAAS (Pre-Litigation as a Service) in May 2026, with $10 million in subscriptions pre-sold. | 高 | SO012, SO004 |
| CO015 | CEO Rami Karabibar's founding motivation came from personal experience with a family member's personal injury case. | 中 | SO009 |
| CO016 | COO Raymond Mieszaniec previously co-founded fintech startup EquitySim and worked in PwC's risk consulting practice in Hong Kong. | 中 | SO009 |
| CO017 | Saam Mashhad serves as Chief of Product and Legal Operations, a dual role reflecting EvenUp's product-legal integration thesis. | 高 | SO002, SO010 |
| CO018 | Series E co-investors include REV Ventures (RELX/LexisNexis's venture arm), B Capital Group, and Bain Capital Ventures. | 高 | SO011, SO013 |
| CO019 | EvenUp employs a 300+ person legal operations team that provides human review of AI-generated outputs. | 中 | SO009 |
| CO020 | EvenUp launched the Medical Management product in December 2025 to identify treatment gaps in active PI cases. | 中 | SO011, SO016 |
| CO021 | EvenUp launched Express Demands in January 2025 as the first product in its AI Documents series. | 中 | SO023, SO004 |
| CO022 | EvenUp uses a per-case pricing model introduced alongside the AI Drafts Suite launch in May 2025. | 高 | SO007, SO006 |
| CO023 | The AI Drafts Suite was launched on May 15, 2025, bundling Smart Workflows, Medical Bill Summary, and AI demand drafts. | 高 | SO007, SO003 |
| CO024 | At the time of the AI Drafts Suite launch in May 2025, EvenUp had processed $7 billion in personal injury claims. | 中 | SO007 |
| CO025 | EvenUp is headquartered in San Francisco, California. | 高 | SO001, SO002 |
| CO026 | EvenUp operates a flexible hybrid work model with San Francisco as its primary hub. | 中 | SO005 |
| CO027 | Piai™ is EvenUp's proprietary AI model trained on a large corpus of personal injury case data. | 高 | SO007, SO009 |
| CO028 | Pre-Series D investors in EvenUp include Lightspeed Venture Partners, NFX, and DCM Ventures. | 中 | SO007, SO014 |
| CO029 | SignalFire led EvenUp's seed round, making it the company's first institutional investor. | 中 | SO010 |
| CO030 | Critics have raised concerns that AI legal tools like EvenUp's may introduce algorithmic bias and contribute to inflated demand letters. | 中 | SO008 |
| CO031 | Above the Law raised a specific concern that EvenUp's AI tools may further burden the court system with lower-quality or algorithmically-generated demands. | 中 | SO008 |
| CO032 | Express Demands is described by EvenUp as the first in its AI Documents product series. | 中 | SO023, SO004 |
| CO033 | The Medical Management product helps PI attorneys identify treatment gaps in ongoing client cases to optimize recovery outcomes. | 中 | SO011, SO016 |
| CO034 | PLAAS had $10 million in subscriptions pre-sold at its May 2026 launch, and EvenUp positioned it as a shift from software vendor to outsourced legal operations partner. | 中 | SO012 |
| CO035 | At the AI Drafts Suite launch in May 2025, EvenUp reported serving 1,500+ personal injury law firms. | 中 | SO007 |
| CO036 | Bloomberg Law confirmed EvenUp's total capital raised reached $385 million as of the October 2025 Series E announcement. | 高 | SO013, SO022 |
| CO037 | EvenUp's $135 million Series D was publicly reported in June 2024 and later announced on company channels in October 2024 at approximately $1 billion valuation, with Bain Capital Ventures named as lead investor. | 中 | SO021, SO027 |
| CO038 | EvenUp's customer base grew from 1,500+ firms (May 2025) to 2,000+ firms (Oct 2025) to 30% of top-100 PI firms (May 2026), indicating continued market penetration. | 中 | SO007, SO013, SO012 |
| CO039 | No material leadership departures, board disputes, or regulatory actions against EvenUp were identified during the research period. | 中 | SO004, SO009 |
| CO040 | RELX/LexisNexis's venture arm REV participated in EvenUp's Series E, signaling potential strategic alignment around legal data and distribution. | 中 | SO011, SO013 |
| CM001 | EvenUp's addressable market is plaintiff personal injury legal operations AI, defined as software for demand preparation, case valuation, medical-record analysis, intake triage, and treatment tracking for contingency-fee PI firms. | 中 | SM001, SM022, SM009 |
| CM002 | The PI legaltech AI market explicitly excludes defense-side legal AI, corporate contract-review tools, insurance-carrier claims platforms, and general-purpose AI assistants not calibrated to PI workflows. | 中 | SM001, SM003 |
| CM003 | The contingency fee model means PI attorneys earn a percentage of settlement recovery, so AI tools that increase settlement values or reduce hours-per-case produce direct margin gains with no billable-hour revenue offset. | 中 | SM011, SM009 |
| CM004 | Status-quo substitutes that EvenUp displaces include manual paralegal demand drafting, generic word-processor templates, standalone medical-record summary vendors, and off-the-shelf LLMs used without PI-specific training or HIPAA controls. | 中 | SM001, SM011, SM006 |
| CM005 | Grand View Research valued the global legal AI market at $1.45 billion in 2024. | 中 | SM003 |
| CM006 | Grand View Research projects the global legal AI market will reach $3.90 billion by 2030, growing at a CAGR of 17.3% from 2025 to 2030. | 中 | SM003 |
| CM007 | MarketsandMarkets sized the global legal AI software market at $3.11 billion in 2025, approximately double the Grand View Research figure for legal AI broadly, reflecting a broader scope including generative AI agents and drafting tools. | 中 | SM004 |
| CM008 | MarketsandMarkets projects the legal AI software market will reach $10.82 billion by 2030, growing at a CAGR of 28.3%—significantly higher than GVR's 17.3% estimate, driven by a broader market definition. | 中 | SM004 |
| CM009 | North America accounted for over 46% of the global legal AI market revenue in 2024, making it the largest regional market. | 中 | SM003 |
| CM010 | IBISWorld tracks Personal Injury Lawyers & Attorneys as a distinct US sub-industry, noting the sector is highly fragmented with no firm holding more than 5% market share. | 中 | SM023 |
| CM011 | The US had 864,800 licensed lawyers as of 2024 with a median annual wage of $151,160, projected to grow 4% through 2034 per the BLS Occupational Outlook Handbook. | 高 | SM005, SM018 |
| CM012 | PI law firms are highly fragmented with no single firm holding more than 5% market share; the vast majority operate as practices of fewer than 10 attorneys. | 中 | SM023, SM019 |
| CM013 | EvenUp processes more than 10,000 PI cases per week as of May 2026, per the company's PLAAS product launch disclosure. | 中 | SM013 |
| CM014 | EvenUp reports its platform represents over $14 billion in damages across active cases as of May 2026. | 低 | SM013 |
| CM015 | EvenUp's early testing of its PLAAS managed service resulted in more than $10 million in subscriptions before the general launch in May 2026. | 低 | SM013 |
| CM016 | PI firm owners and managing partners are the economic buyers who control software budgets and make the final subscription decisions for AI platforms. | 中 | SM011, SM013 |
| CM017 | Case managers and paralegals are the primary daily users of EvenUp's platform, handling the highest-volume workflows: medical records coordination, client follow-up, and document intake. | 中 | SM012, SM001 |
| CM018 | Demand writers—staff attorneys and senior paralegals—are the primary users of EvenUp's AI drafting tools, generating and reviewing demand packages and medical chronologies. | 中 | SM001, SM022 |
| CM019 | The majority of US PI law firms operate as practices with fewer than 10 attorneys, creating a fragmented and price-sensitive buyer universe for legal AI. | 中 | SM023, SM005 |
| CM020 | Large national PI firms represent a small share of total firm count but a disproportionate share of annual case volume and EvenUp's enterprise revenue opportunity. | 中 | SM013, SM014 |
| CM021 | Intake teams and legal-ops staff at PI firms use AI for case triage, coverage verification, and statute-of-limitations risk screening as secondary but growing workflow applications. | 中 | SM013, SM001 |
| CM022 | EvenUp's PLAAS (Pre-Litigation as a Service) model shifts the buyer relationship from a per-seat SaaS subscription to a per-case managed-service contract, expanding the addressable buyer segment to firms that lack internal implementation capacity. | 中 | SM013 |
| CM023 | The contingency fee model creates a direct margin incentive for PI attorneys: every hour saved on case preparation that does not reduce settlement value is pure profit, unlike hourly-billed markets where AI savings reduce revenue. | 中 | SM011, SM003 |
| CM024 | EvenUp's own data analysis shows 43% of PI cases eventually experience treatment gaps longer than 30 days, reducing case value; real-time monitoring of these gaps is a direct value driver for AI adoption. | 中 | SM012 |
| CM025 | Medical record volume per PI case has grown substantially due to more complex treatment histories, electronic health records, and multi-provider care paths, increasing the labor cost of manual review and the value of AI-assisted analysis. | 中 | SM012, SM017 |
| CM026 | Insurance companies have become more sophisticated in claims reserve-setting, requiring better-documented, data-backed demand packages; this directly benefits AI platforms that improve demand quality. | 中 | SM011, SM010 |
| CM027 | Settlement cycle compression—insurers pushing for faster case resolution to reduce reserves—creates demand for faster and more accurate demand preparation, which AI platforms can deliver. | 中 | SM013, SM011 |
| CM028 | Labor shortages in legal support roles—particularly paralegals and medical record reviewers—have intensified since 2020, accelerating demand for automation in PI case preparation workflows. | 中 | SM005, SM013 |
| CM029 | ABA Model Rule 1.1 (Competence) requires attorneys to maintain sufficient understanding of AI tools they use in client representations, including Comment 8 on technological competence; this creates an ongoing supervision obligation that cannot be delegated to AI alone. | 高 | SM006, SM007 |
| CM030 | ABA Model Rule 1.6 (Confidentiality) prohibits lawyers from disclosing client information without informed consent, requiring that AI platforms handling PI case data operate under strict data isolation and contractual confidentiality obligations. | 高 | SM007, SM006 |
| CM031 | PI cases routinely involve protected health information (PHI) from medical records, making HIPAA compliance mandatory for any AI platform that processes this data; EvenUp has achieved HIPAA attestation to address this requirement. | 中 | SM001, SM017 |
| CM032 | AI hallucinations—confidently incorrect factual assertions or fabricated legal citations—represent a specific malpractice and sanctions risk when embedded in demand letters or legal filings, creating an irreducible supervision requirement. | 中 | SM011, SM016 |
| CM033 | Healthcare data breach rates remain elevated, with 7,419 large breaches reported to OCR between 2009 and January 2026, providing context for the data security risk that PI AI platforms inheriting medical records must manage. | 中 | SM017 |
| CM034 | Above the Law commentary identifies ongoing concerns about algorithmic bias, data privacy, and the risk that AI-generated demand letters could 'gum up the court system' by flooding courts with standardized filings. | 中 | SM011 |
| CM035 | No independent analyst firm publishes a standalone market-size estimate for the PI plaintiff legaltech AI sub-market; all available figures are for broader legal AI or legaltech categories. | 中 | SM003, SM004, SM023 |
| CM036 | The PI-specific SAM estimated at $150M–$600M (US) is a bottom-up analyst approximation based on firm count and ARPU assumptions, not a published figure; it contains significant uncertainty. | 低 | SM005, SM013, SM023 |
| CM037 | EvenUp's top-100 PI firm penetration figure increased from 20% (reported at the October 2025 Series E per Bloomberg Law and careers page) to 30% (reported at the May 2026 PLAAS launch per LawNext), indicating continued market share gains in the 7-month period. | 中 | SM013, SM014, SM020 |
| CM038 | Small PI firms (under 10 attorneys) represent the majority of the firm count but face greater cost sensitivity and implementation friction than large national firms, limiting their AI adoption rate relative to their numeric share of the market. | 中 | SM023, SM011 |
| CM039 | EvenUp's proprietary PI case dataset—built on settlements, verdicts, and medical records from thousands of firms—constitutes a network-effect data moat that reinforces its market positioning as case volume grows. | 中 | SM011, SM013 |
| CM040 | ABA Model Rule 5.5 prohibits assisting in the unauthorized practice of law, raising a compliance question for AI systems that produce legal strategy recommendations without licensed attorney review at each decision point. | 中 | SM008, SM016 |
| CM041 | BLS projects US lawyer employment to grow by 35,900 jobs from 2024 to 2034 (4% growth), with approximately 31,500 annual openings, indicating continued but modest labor force expansion in the legal sector. | 中 | SM005 |
| CM042 | The US personal injury attorneys industry is classified by IBISWorld under NAICS code OD4812 and is characterized by high fragmentation with no firm exceeding 5% share, confirming the predominantly small-business buyer profile. | 中 | SM023 |
| CM043 | The countrywide average US auto insurance expenditure was $1,127 per vehicle in 2022 per NAIC data, reflecting the volume of premiums underlying the auto-accident claims pool that drives a large share of PI case volume. | 中 | SM010 |
| CM044 | EvenUp's analysis of AI usage across 200+ firms in July 2025 found that 69% of prompts focused on medical treatment analysis and injury assessment, and 75% of drafting prompts centered on litigation documents. | 中 | SM012 |
| CM045 | As of May 2026, EvenUp is used by 30% of the top 100 PI firms in the US, up from 20% at the October 2025 Series E, per LawNext and Bloomberg Law reporting. | 高 | SM013, SM014 |
| CP001 | EvenUp's competitive landscape spans five categories: direct PI specialist AI peers (primarily Supio), workflow-platform CMS incumbents (Filevine, Litify, Clio, CasePeer, SmartAdvocate), horizontal legal AI platforms (Harvey, CoCounsel, Lexis+ Protégé), general-purpose LLMs, and manual status-quo substitutes. | 高 | SP001, SP002, SP009, SP010 |
| CP002 | As of May 2026, EvenUp is used by 2,000+ PI law firms and 30% of the top 100 PI firms in the US, making it the scale leader among direct PI AI peers. | 高 | SP001, SP005 |
| CP003 | EvenUp's proprietary AI system, Piai, is trained on 250,000+ PI verdicts and settlements, a dataset scale that no publicly identified competitor has claimed to match as of May 2026. | 中 | SP006, SP001 |
| CP004 | EvenUp has raised $385M total ($135M Series D, June 2024; $150M Series E, October 2025) at a $2B+ valuation, giving it substantially more disclosed financial resources than any identified direct PI AI peer. | 高 | SP005, SP007 |
| CP005 | No published independent competitive benchmark comparing EvenUp, Supio, and CMS AI platforms on PI demand-letter accuracy, settlement improvement, or processing speed has been identified as of May 2026. | 高 | SP001, SP002 |
| CP006 | Supio was founded in 2021 by Jerry Zhou (CEO) and Kyle Lam (CTO), is headquartered in Seattle with a San Francisco office (opened February 2026), and describes itself as 'the only agentic legal AI platform built for plaintiff law and mass torts cases' as of May 2026. | 高 | SP002, SP003 |
| CP007 | Supio's pricing model offers two tiers—Case Subscription (volume-based, lower per-case costs) and Unlimited Firm Access (lowest per-case cost, unlimited cases)—with no lock-in contracts or platform fees, directly paralleling EvenUp's per-case pricing approach. | 中 | SP004 |
| CP008 | Supio's platform spans pre-litigation (Case Engine: medchrons, demands, bill reconciliation) and litigation (Case Bench: AI chat, drafting, interrogatory responses), covering a comparable PI case lifecycle to EvenUp's platform. | 高 | SP004, SP002 |
| CP009 | Supio's website states 'join hundreds of personal injury firms' as of May 2026, compared to EvenUp's 2,000+ firms—indicating Supio is approximately 5–10× smaller by customer count. | 中 | SP003 |
| CP010 | Supio announced a partnership with Thomson Reuters in April 2026 to integrate Westlaw Advantage directly into its platform, providing attorneys with verified legal-citation access without switching tools—a capability EvenUp does not currently offer. | 中 | SP008 |
| CP011 | Supio has achieved SOC2 Type 2 certification and HIPAA and GDPR compliance, matching EvenUp's compliance posture and making compliance a table-stakes parity rather than a differentiating advantage for either platform. | 高 | SP002, SP003 |
| CP012 | Supio launched Supio Agent in May 2026 as an agentic AI system covering case-level and firm-level operations, including intake, case building, and firm analytics, positioning itself as natively agentic rather than a document-generation tool with AI retrofitted. | 高 | SP008, SP002 |
| CP013 | Darrow AI (founded in Israel, New York office) targets legal exposure management—surfacing regulatory risk and litigation signals before harm materializes—and is not a direct PI plaintiff demand-workflow competitor to EvenUp. | 中 | SP009 |
| CP014 | Filevine added FilevineAI medical-chronology generation in April 2025 and offers LOIS (agentic AI layer), court-reporting and deposition tools (August 2025), deadline management (October 2025), and document generation as part of its enterprise CMS platform. | 高 | SP010, SP011 |
| CP015 | Litify's Agentic Case Expert (ACE) provides AI intake agents, source-linked medical chronologies, demand packet drafting, case-value forecasting, and settlement scenario modeling—making it a functionally broad PI AI competitor integrated into the Litify CMS. | 中 | SP012 |
| CP016 | Clio launched Clio Work in late 2025 as 'the only AI that understands your cases, their context, and the law,' covering matter analysis, legal research integration, and drafting—overlapping conceptually with EvenUp's full-lifecycle AI platform. | 中 | SP013 |
| CP017 | CasePeer is a PI-specialist CMS acquired by Clio offering turnkey practice management for small/mid PI firms, with AI capabilities governed by Clio's product roadmap rather than PI-specific training, giving Clio a bundling advantage. | 中 | SP014 |
| CP018 | SmartAdvocate offers SmartIntelligence as built-in AI tools for case summarization, medical-record summarization, multi-document analysis, email refinement, translation, and transcription—narrower in scope than EvenUp's full lifecycle platform. | 中 | SP015 |
| CP019 | EvenUp's Smart Workflows integrate with SmartAdvocate, Litify, and CasePeer per EvenUp's own blog disclosures, confirming that CMS integration partners currently serve as distribution enablers for EvenUp rather than pure substitutes. | 中 | SP006 |
| CP020 | Filevine's 2026 pricing is fully custom-built per team and Litify's pricing is enterprise custom; neither publishes list pricing, indicating enterprise sales motion with higher switching costs and longer sales cycles than EvenUp's per-case model. | 高 | SP011, SP025, SP012 |
| CP021 | Filevine's 2026 'Legal AI Trust Index' marketing materials argue that AI trust comes from data grounded in the firm's own workflows—a direct counter-positioning to EvenUp's argument that a specialized PI dataset provides superior outcomes. | 中 | SP011 |
| CP022 | EvenUp's per-case pricing model, launched in May 2025, replaced feature-tiered pricing to provide one clear predictable cost per case across the full Claims Intelligence Platform, creating direct comparability with Supio's case-based model. | 中 | SP006 |
| CP023 | Harvey AI (founded 2022, San Francisco) has raised an estimated $300M+ through 2025, targets AmLaw 200 and BigLaw firms, and focuses on corporate counsel, M&A, and enterprise litigation—not PI plaintiff law firms. | 中 | SP016, SP017 |
| CP024 | CoCounsel (Thomson Reuters, acquired from Casetext in 2023 for approximately $650M) is grounded in Westlaw and Practical Law databases, targeting enterprise litigation and corporate counsel—not PI plaintiff firms. | 高 | SP018, SP019 |
| CP025 | Lexis+ with Protégé (formerly Lexis+ AI, rebranded in 2026 per LexisNexis product page) provides legal drafting, research, and analysis grounded in LexisNexis databases with Shepard's citation verification and DMS integrations—not calibrated for PI plaintiff workflows. | 中 | SP020 |
| CP026 | Lexis+ with Protégé explicitly supports drafting discovery documents, deposition questions, and litigation documents from uploaded materials—overlapping conceptually with EvenUp's discovery and demand modules—but lacks PI-specific verdict/settlement benchmarking. | 中 | SP020 |
| CP027 | Thomson Reuters simultaneously owns CoCounsel (its enterprise legal AI) and has a Westlaw Advantage partnership with Supio (April 2026), a potential channel conflict that could limit EvenUp's access to Thomson Reuters' legal-citation infrastructure. | 中 | SP008, SP019 |
| CP028 | Harvey's 2026 product direction includes 'Legal Agents for Every Matter, Tailored to You'—agentic capabilities that could theoretically extend to PI workflows, though no public evidence of PI targeting has been identified as of May 2026. | 低 | SP016 |
| CP029 | None of the three horizontal legal AI platforms (Harvey, CoCounsel, Lexis+ Protégé) has disclosed PI-specific training data, verdict/settlement corpora, or explicit PI plaintiff firm targeting as of May 2026. | 高 | SP016, SP017, SP018, SP020 |
| CP030 | Manual demand drafting in PI cases historically requires 5–20 hours of paralegal or staff-attorney time per demand letter, representing $250–$2,000 of internal labor cost at prevailing rates—the primary cost EvenUp's per-case pricing is designed to displace. | 中 | SP006, SP021 |
| CP031 | EvenUp customers have reported 3× more demand letters generated (Jeffcoat Injury Lawyers using Express Demands), 30-day faster case settlement, 300% settlement improvement (John K. Zaid & Associates), and 400%+ revenue growth (ELG Injury Lawyers), per EvenUp's homepage and blog. | 中 | SP001, SP006 |
| CP032 | Offshore legal support services provide document processing and basic chronology preparation at equivalent rates of $15–$40 per hour, representing a cost-competitive alternative for small PI firms that cannot afford in-house staff or per-case AI pricing. | 低 | SP021 |
| CP033 | EvenUp's blog post listing '5 Reasons Claude Isn't Right for Your Personal Injury Law Firm' explicitly addresses the general-purpose LLM competitive threat, noting lack of PI training data, absent HIPAA data isolation, no line-level citation traceability, no PI verdict benchmarking, and no proactive workflow agents. | 高 | SP022, SP001 |
| CP034 | As large language model capabilities improve over 2026–2028, the document-generation component of EvenUp's value proposition faces commoditization risk from general-purpose LLMs, leaving compliance, dataset quality, and proactive agents as the primary differentiators. | 中 | SP023, SP022 |
| CP035 | EvenUp's proprietary statistic—42% of demand letters are sent more than 100 days after last treatment—drives Smart Workflows' demand-timing alerts, a workflow-agent capability that cannot be replicated by general-purpose LLMs without PI case data access. | 中 | SP006 |
| CP036 | EvenUp's primary competitive moat is its proprietary PI case dataset (250,000+ verdicts/settlements); no identified competitor has disclosed a comparable PI-specific corpus, providing a training advantage estimated at 3–5 years to replicate at scale. | 中 | SP001, SP003, SP006 |
| CP037 | EvenUp grew from approximately 1,500 firms (mid-2025) to 2,000+ (May 2026) and from roughly 20% of top-100 PI firms at Series E (October 2025) to 30% by May 2026, indicating sustained market-share gains in a competitive environment. | 高 | SP005, SP007, SP024 |
| CP038 | EvenUp integrates with CMS platforms (SmartAdvocate, Litify, CasePeer) that are themselves adding AI features that replicate EvenUp's core demand-drafting and medical-chronology capabilities, creating a structural channel-conflict risk that escalates as CMS AI depth increases. | 高 | SP006, SP010, SP012, SP015 |
| CP039 | SmartAdvocate currently integrates with EvenUp—confirmed by EvenUp's blog noting Smart Workflows connect to SmartAdvocate—while simultaneously offering SmartIntelligence AI tools, demonstrating the co-opetition dynamic between EvenUp and its CMS integration partners. | 高 | SP006, SP015 |
| CP040 | EvenUp's Mirror Mode (firm-specific style replication) directly addresses the demand-letter commoditization risk where insurance adjusters learn to recognize and discount standardized AI-generated demands—a risk noted in Above the Law commentary on EvenUp's scale. | 中 | SP023, SP001 |
| CP041 | Filevine's 2025–2026 product launches include medical-record AI analysis (April 2025), court-reporting and deposition management (August 2025), and deadline-management tools (October 2025), indicating continuous AI expansion into PI-relevant workflow steps that EvenUp currently covers. | 中 | SP010 |
| CP042 | Clio's acquisition of CasePeer and development of Clio Work means a single parent company controls both a PI-specialist CMS and a cross-practice-area AI platform, creating a bundling capability that could undercut EvenUp's standalone PI AI economics for small PI firms. | 中 | SP013, SP014 |
| CP043 | As of May 2026, EvenUp has confirmed CMS integrations with SmartAdvocate, Litify, and CasePeer (per its blog), with Filevine's integration status unclear; Filevine's AI expansion represents the most significant potential platform-conflict scenario given its enterprise scale. | 中 | SP006, SP010 |
| CP044 | Five CMS platforms with active 2024–2026 AI launches in the PI space (Filevine, Litify, Clio, CasePeer, SmartAdvocate) collectively represent both EvenUp's primary distribution channel and its primary medium-term encroachment risk—a dual role that creates strategic exposure unique to vertical AI companies embedded in incumbent platforms. | 中 | SP010, SP011, SP012, SP013, SP014, SP015 |
| CI001 | EvenUp transitioned its revenue model in May 2025 from tiered feature licensing to a unified per-case pricing model—one predictable cost per case covering the full Claims Intelligence Platform (demand drafting, medchrons, Smart Workflows, Medical Bill Summary, Express Demands). | 高 | SI010, SI017 |
| CI002 | EvenUp's per-case price for the Claims Intelligence Platform is not publicly disclosed; analyst estimates place average revenue per firm at $5,000–$20,000 per year based on firm count and capital raised, but this is not a company-disclosed figure. | 中 | SI010, SI009 |
| CI003 | EvenUp launched PLAAS (Pre-Litigation as a Service) in May 2026, combining AI technology with 300+ US-based legal operations staff to handle the full pre-litigation case lifecycle including claim setup, care coordination, records retrieval, demand preparation, settlement negotiation, and optional lien resolution. | 高 | SI012, SI015, SI009 |
| CI004 | PLAAS generated $10M+ in subscription sales during early testing as of May 2026, making it the only concrete revenue figure publicly disclosed for EvenUp across its full operating history. | 中 | SI012 |
| CI005 | EvenUp's revenue is exclusively US-based as of May 2026; no international expansion plans have been publicly disclosed. | 中 | SI009, SI015 |
| CI006 | EvenUp's Medical Management product (launched December 2025) adds a proactive treatment-gap identification and care-coordination service that addresses under-treatment risk before demands are filed, enabling a higher potential settlement and thus a higher-value case for EvenUp to process. | 高 | SI013, SI016 |
| CI007 | EvenUp's PLAAS subscription model differs structurally from its per-case SaaS pricing: PLAAS is an all-case subscription covering the full pre-litigation lifecycle rather than a per-document or per-case fee, creating a recurring revenue stream with higher average contract value per firm. | 高 | SI012, SI015 |
| CI008 | EvenUp has raised $385M in total disclosed venture capital—approximately $100M across Seed through Series C (implied, not publicly confirmed by round), $135M in Series D (October 2024), and $150M in Series E (October 2025). | 高 | SI001, SI003, SI004, SI018, SI019 |
| CI009 | EvenUp's Series D ($135M) was led by Bain Capital Ventures and announced on October 8, 2024, with participation from Premji Invest, Lightspeed Venture Partners, Bessemer Venture Partners, SignalFire, and B Capital Group, at a post-money valuation exceeding $1 billion. | 高 | SI001, SI003, SI019 |
| CI010 | EvenUp's Series E ($150M) was led by Bessemer Venture Partners and announced October 1, 2025, with participation from REV (the venture arm of RELX/LexisNexis), B Capital Group, and Bain Capital, at a post-money valuation exceeding $2 billion. | 高 | SI004, SI005, SI013, SI018 |
| CI011 | The total disclosed capital raised at Series D ($135M + ~$100M prior rounds) was approximately $235M, implying prior-round capital of approximately $100M across Seed through Series C before October 2024, per EvenUp's own announcement. | 中 | SI001 |
| CI012 | SignalFire was an early seed investor in EvenUp and also participated in the Series D, confirming continued insider support from a data-science-focused early-stage fund. | 高 | SI021, SI003, SI019 |
| CI013 | REV, the venture arm of RELX plc (parent company of LexisNexis), participated in EvenUp's Series E, creating a structural competitive conflict: RELX owns a direct legal AI competitor while holding a minority equity stake in EvenUp. | 高 | SI004, SI013, SI018 |
| CI014 | EvenUp's legal entity is incorporated as EvenUp, Inc., a privately held Delaware software company, as confirmed by the ServiceTitan DRS/A filing on SEC EDGAR (October 2024), which names a board member with EvenUp service starting April 2023. | 中 | SI002 |
| CI015 | Lightspeed Venture Partners participated in EvenUp's Series D but did not participate in the Series E; the reason for this change has not been publicly disclosed. | 高 | SI003, SI004, SI019 |
| CI016 | EvenUp's 300+ US-based legal operations team creates structural gross margin pressure: at $100K–$150K fully loaded per employee, this headcount alone implies approximately $45M–$75M in annual labor COGS before AI infrastructure, sales, and G&A expenses. | 中 | SI011, SI012, SI015 |
| CI017 | At the Series D (October 2024), EvenUp disclosed a team of 100+ engineers and product professionals, implying R&D headcount that would represent a significant second cost line in addition to the legal operations team. | 中 | SI001 |
| CI018 | EvenUp's AI infrastructure costs—GPU compute for Piai model training and inference across 200,000+ cases and ~10,000 cases/week—are not publicly disclosed; analyst estimates for comparable legal AI platforms suggest $5M–$20M annually depending on model architecture and prompt complexity. | 低 | SI011, SI009 |
| CI019 | EvenUp holds SOC2 Type II certification and HIPAA Business Associate Agreement compliance, adding ongoing audit and compliance overhead (estimated tens of thousands to hundreds of thousands annually) that is atypical for pre-revenue-stage AI companies but standard for HIPAA-adjacent healthcare data processors. | 中 | SI017, SI024 |
| CI020 | PLAAS converts EvenUp's legal operations team from a cost center to a direct revenue-generating workforce; the financial viability of PLAAS depends on whether per-case or subscription pricing covers the approximately $45M–$75M in attributed annual labor cost plus AI infrastructure. | 中 | SI012, SI015, SI011 |
| CI021 | If PLAAS subscription pricing covers the fully-loaded cost of the legal operations team, EvenUp's residual SaaS platform gross margin could approach 70–80%+; if not, blended company gross margin will remain materially below pure-software AI peers. | 中 | SI012, SI011 |
| CI022 | At Series D (October 2024), EvenUp reported: revenue growth exceeding 100% year-over-year; more than 1,000 demand documents processed per week; workforce doubled in the prior twelve months; and 20% of customers using multiple products. | 中 | SI001 |
| CI023 | EvenUp's firm customer count grew from 1,000+ at Series D (October 2024) to 2,000+ as of May 2026, with 30% penetration of the top 100 US personal injury law firms. | 高 | SI009, SI018, SI011 |
| CI024 | EvenUp processed 200,000+ cumulative PI cases as of October 2025 (per the Unite.ai COO interview), generating $10B+ in aggregate damages; by May 2026 the homepage states $14B+ in damages and ~10,000 cases/week. | 高 | SI011, SI009 |
| CI025 | EvenUp's Series E valuation of $2B+ implies an approximate ARR range of $40M–$200M at typical 2025 vertical AI SaaS revenue multiples of 10–50×, though this range is entirely inferential—no ARR figure has been disclosed. | 低 | SI004, SI018, SI020 |
| CI026 | PLAAS early-testing results reported by LawNext (May 2026): 95% of available third-party policy limits achieved; medical records retrieved 66 days faster than industry norm; demand packets prepared 47 days faster; $1,000 per case in cost savings to PI firms. | 中 | SI012 |
| CI027 | Sweet James PI firm reported $500M+ in annual results with 70% year-over-year growth using EvenUp, per COO Raymond Mieszaniec's October 2025 interview with Unite.ai—a company-sourced customer testimonial with no independent corroboration. | 中 | SI011 |
| CI028 | Jeffcoat Injury Lawyers reported producing 3× more demand letters and settling cases 30 days faster using EvenUp's AI Drafts Suite, per EvenUp's May 2025 product launch blog—a company-controlled customer testimonial with no independent corroboration. | 中 | SI010 |
| CI029 | Customer Dwuan Hammond's law firm reported 300% top-line revenue growth attributable to EvenUp per the May 2025 AI Drafts Suite launch blog, a company-controlled testimonial with no independent corroboration or methodology disclosure. | 低 | SI010 |
| CI030 | Clio's 2025 Legal Trends Report documents accelerating AI adoption among law firms, with PI firms showing above-average technology investment rates—a market tailwind that supports EvenUp's customer growth trajectory but does not constitute independent financial evidence for EvenUp specifically. | 中 | SI008 |
| CI031 | EvenUp has not disclosed ARR, revenue, gross margin, operating loss, burn rate, runway, or EBITDA for any period in any public source as of May 2026—making its financial profile materially opaque relative to its $2B+ valuation. | 高 | SI001, SI004, SI009 |
| CI032 | The $10M+ PLAAS early-testing subscription figure represents a new product in early testing and cannot be used as a proxy for EvenUp's total recurring revenue base, which remains unknown. | 中 | SI012 |
| CI033 | All published EvenUp customer ROI figures (300% revenue growth, $500M+ Sweet James results, 3× demand volume for Jeffcoat) are sourced from EvenUp's own marketing materials or company-controlled interview channels with no independent validation, median-vs-outlier disclosure, or control-group comparisons. | 高 | SI010, SI011 |
| CI034 | REV (RELX/LexisNexis venture arm) participation in EvenUp's Series E creates an undisclosed governance risk: standard Series E investor rights (board observer, information rights, ROFR) could provide RELX/LexisNexis with non-public financial and product intelligence that constitutes competitive advantage. | 中 | SI004, SI013, SI014 |
| CI035 | AboveTheLaw's October 2025 Series E coverage identified three standard AI risk concerns specifically applicable to EvenUp's financial model: algorithmic bias in demand generation, data privacy for PHI, and demand inflation potentially training insurance adjusters to apply systematic discounts over time. | 中 | SI014 |
| CI036 | LawNext's May 2026 PLAAS launch article noted that PLAAS' long-term value depends on the depth of firm integration and scale of outcomes, and that the distinction between 'genuinely new category' and 'AI wrapper around existing services' remains an open question—a structural financial risk if PLAAS fails to sustain its early-testing premium. | 中 | SI012 |
| CI037 | No Form D filing for EvenUp, Inc. has been identified in SEC EDGAR as of May 2026 under any searched name variant (EvenUp, EvenUp Inc, EvenUp Law, EvenUp Inc., Mighty Law); this is unusual for a company that has raised $385M in equity under Regulation D. | 中 | SI002 |
| CI038 | Bessemer Venture Partners' 'Roads to $100M ARR in Legal AI' framework positions EvenUp as a category-defining legal AI company on a trajectory toward $100M ARR, representing Bessemer's public endorsement of EvenUp's growth narrative as the lead Series E investor. | 中 | SI020 |
| CI039 | EvenUp's Axios interview (October 2025) disclosed co-founder Rami Karabibar discussing IPO prospects, suggesting leadership is contemplating a liquidity event—consistent with the capital scale ($385M) and valuation ($2B+) but without a confirmed timeline. | 中 | SI022 |
| CI040 | At EvenUp's $2B+ Series E valuation and assuming 2,000+ firm customers, implied ARPU needed to justify a 20× ARR multiple ranges from $50,000–$100,000+ per firm—substantially above the $5,000–$20,000 analyst estimate—suggesting either higher actual ARPU than estimated, higher-than-20× multiple, or reliance on PLAAS revenue materially lifting total ARR. | 低 | SI009, SI004, SI020 |
| CE001 | EvenUp's Claims Intelligence Platform™ is powered by Piai™, a proprietary AI system trained on hundreds of thousands of personal injury cases and millions of medical records. | 高 | SE001, SE019 |
| CE002 | Piai™ encompasses a coordinated suite of specialized AI models—not a single model—each optimized for specific PI legal workflow tasks such as medical entity extraction, case timeline analysis, and document drafting. | 高 | SE001, SE010 |
| CE003 | The Claims Intelligence Platform™ provides rich business insights, AI workflow automation, and best-in-class document creation capabilities for personal injury law firms. | 高 | SE001, SE018 |
| CE004 | EvenUp serves 2,000+ PI firms as of May 2026 and processes more than 10,000 cases per week, representing over $14 billion in damages managed through the platform. | 中 | SE020, SE026 |
| CE005 | Express Demands™ was introduced in January 2025 as EvenUp's first AI Draft template, enabling generation of demand letters from case files with court-ready exhibits and comparable verdict analysis. | 高 | SE002, SE019 |
| CE006 | AI Drafts Suite™ launched in May 2025 and covers the full document lifecycle: demand letters, complaints, medical summaries, negotiation sheets, and responses to interrogatories. | 高 | SE019, SE012 |
| CE007 | Smart Workflows provides data-driven case lifecycle automation using EvenUp's proprietary dataset showing that 42% of demands are sent more than 100 days after last treatment, enabling firms to trigger alerts and actions at optimal case moments. | 中 | SE019, SE004 |
| CE008 | Medical Bill Summary, part of EvenUp's Case Financials solution, automates charge amount tracking and enables staff to validate and exclude unrelated charges against available policy limits. | 中 | SE006, SE019 |
| CE009 | MedChrons™ organizes raw medical records into an interactive, professionally reviewed chronology allowing attorneys to evaluate diagnoses, ICD codes, interventional treatments, and objective tests in a single structured view. | 高 | SE003, SE002 |
| CE010 | Medical Management, launched December 2025, provides an interactive real-time treatment timeline—including history, upcoming appointments, medical expenses, and communications—available within one hour of case file upload. | 中 | SE021, SE006 |
| CE011 | Communication Agents™ are AI-powered voice and SMS agents that autonomously open insurance claims, verify liability and coverage, conduct client treatment check-ins, follow up on medical record requests, and verify medical balances—saving teams 9+ hours per case according to EvenUp. | 中 | SE005, SE021 |
| CE012 | AI Playbooks™, launched July 8, 2025, automatically analyzes case files and extracts case-critical insights each time new documents are uploaded, flagging liability issues, TBI indicators, commercial defendants, and DUI indicators without attorney prompting. | 高 | SE012, SE018 |
| CE013 | Voice Agent™ is a conversational AI available in early access as of July 2025 that supports PI firms 24/7 across the case lifecycle, starting with care management outreach, capturing structured summaries and full transcripts of each interaction. | 中 | SE012, SE005 |
| CE014 | Settlement Repository™ uses Piai™ to ingest a firm's own historical settlement data and provides firmwide benchmarks—filterable by injury, treatment, and policy limits—so attorneys can support demand valuation with comparable case outcomes. | 高 | SE007, SE001 |
| CE015 | Executive Analytics™ provides firm leadership with KPI dashboards comparing performance against peer firms and filterable by time period, office, role, or individual team member, with automated weekly email digest reporting. | 中 | SE008, SE018 |
| CE016 | PLAAS (Pre-Litigation as a Service), launched May 2026, is a managed-services subscription combining Piai™ AI with EvenUp's US-based case management staff to handle the full pre-litigation lifecycle—including claim setup, records retrieval, demand preparation, settlement negotiation, and optional lien resolution. | 中 | SE020, SE019 |
| CE017 | Piai™ is built on a two-layer architecture: a Reading Layer that extracts structured information from raw unstructured documents, and a Writing Layer that grounds document generation in the extracted facts rather than general LLM world knowledge. | 高 | SE001, SE011 |
| CE018 | The Reading Layer processes the full spectrum of PI case materials—PDFs with varied layouts, scanned documents, handwritten notes, ICD-coded medical forms, police reports, and accident images—extracting case entities such as medical providers, dates of service, ICD codes, treatment descriptions, and bill line items. | 高 | SE001, SE010 |
| CE019 | The Writing Layer generates legal document outputs (demands, complaints, interrogatories, negotiation sheets) grounded in facts extracted by the Reading Layer rather than in general world knowledge, with every assertion constrained by the System of Record. | 高 | SE001, SE011 |
| CE020 | EvenUp's architecture follows a System of Record → System of Intelligence → System of Action progression: persistent normalized case data enables reusable downstream automation, then proactive workflow actions triggered from case entity state. | 高 | SE010, SE001 |
| CE021 | EvenUp uses PPO (Proximal Policy Optimization) reinforcement learning to train image-to-text summarization models: the system generates image-grounded question sets, evaluates text-only QA answerability from the summary, and trains until answerability matches pixel-grounded accuracy—with summaries stored once at ingestion in the System of Record. | 高 | SE010, SE001 |
| CE022 | Every Piai™ output includes line-level citations that link each assertion to the specific source document, page, and line of origin, enabling attorneys to verify AI-generated facts and fulfill their ABA Rule 1.1 supervisory obligation. | 高 | SE001, SE011 |
| CE023 | EvenUp receives thousands of expert input and data updates daily per entity extraction model, contributed by both end-users (lawyers, paralegals, case managers) and internal legal experts, creating a continuous and targeted model fine-tuning loop. | 中 | SE011, SE010 |
| CE024 | EvenUp's VP of Engineering and Head of AI Haixun Wang is an ACM Fellow and IEEE Fellow recognized for foundational work on graph-based systems and text understanding, with prior senior roles at Amazon, Meta, Microsoft, Google, and IBM and 200+ academic publications. | 高 | SE013, SE010, SE027 |
| CE025 | EvenUp integrates natively with four major PI case management systems—Filevine, Litify, SmartAdvocate, and CasePeer—with the company claiming implementation can be completed in less than one business day. | 中 | SE004, SE019 |
| CE026 | The Litify CMS integration delivers a link to the EvenUp-generated draft directly within the Litify interface upon document creation, eliminating the need for attorneys to switch applications. | 中 | SE004, SE019 |
| CE027 | Smart Workflows leverages CMS integrations to automatically scan case data and trigger actions when firm-defined conditions are met, with firms able to customize rules and track adherence through analytics. | 中 | SE019, SE004 |
| CE028 | EvenUp's CMS integrations pre-populate each new case with client details, policy numbers, and intake transcripts from the CMS, and maintain continuous synchronization as new records and bills arrive. | 中 | SE004, SE023 |
| CE029 | EvenUp's proprietary PI dataset was built by crowdsourcing anonymized settlement and case-outcome data from PI firm customers, providing a corpus of pre-settlement negotiation patterns and case-strategy outcomes unavailable to general-purpose AI competitors. | 中 | SE001, SE022 |
| CE030 | Unlike general-purpose LLMs that operate on extracted text summaries, Piai™ operates directly on raw case records—PDFs, scans, handwriting—enabling superior accuracy for PI-specific entity extraction tasks compared to document-summary-based RAG approaches. | 中 | SE001, SE011 |
| CE031 | EvenUp's quality review layer includes 100+ US-based nurses, paralegals, experienced adjusters, case managers, and attorneys who validate Piai™ outputs before delivery and whose corrections feed back into Reading Layer model retraining. | 高 | SE001, SE010 |
| CE032 | EvenUp's expert correction signal—thousands of targeted entity-model updates per day contributed by firm users and internal experts—creates a data flywheel that produces continuously improving PI-specific model precision unavailable to general-purpose LLM competitors without equivalent domain feedback. | 中 | SE011, SE014 |
| CE033 | Jeffcoat Injury Lawyers, using Express Demands™, reported producing 3× more demand letters per period and settling cases 30 days faster, with the firm growing top-line revenue approximately 300% over the same timeframe (all company-sourced). | 中 | SE019, SE022 |
| CE034 | Early PLAAS results reported by EvenUp show: firms recovering 95% of available third-party policy limits; medical records requested 66 days faster than industry baseline; demands delivered 47 days faster; and approximately $1,000 per case in cost savings—all company-sourced from early-cohort testing. | 中 | SE020, SE019 |
| CE035 | EvenUp executes a Data Processing Agreement (DPA) with all customers covering CCPA and applicable US state data protection laws, with data security controls including logical data segregation, role-based access controls, and encryption for customer personal data. | 高 | SE009, SE001 |
| CE036 | EvenUp's DPA commits to notifying customers of a confirmed Security Incident within 72 hours of discovery and to making third-party security audit reports (SOC 2-equivalent) available to customers upon request. | 高 | SE009, SE001 |
| CE037 | EvenUp processes PHI (Protected Health Information) from case medical records under a Business Associate Agreement (BAA) that governs HIPAA-regulated data handling obligations, as referenced in the company's DPA. | 高 | SE009, SE001 |
| CE038 | EvenUp's documented technical security measures include enterprise firewalls and intrusion detection systems, change management procedures, incident response procedures, and business continuity and disaster recovery programs per Schedule 2 of the company's DPA. | 高 | SE009, SE001 |
| CE039 | ABA Model Rule 1.1 (Competence) requires attorneys to maintain the technological knowledge and supervisory oversight necessary for competent representation, creating a structural obligation for PI firm customers using EvenUp to review all AI-generated outputs rather than rely on them uncritically. | 高 | SE025, SE022 |
| CE040 | Third-party coverage identifies algorithmic bias in EvenUp's PI training dataset and data privacy risks from concentrated PHI handling as standard concerns for AI tools operating in personal injury legal practice. | 中 | SE022, SE025 |
| CE041 | General-purpose LLMs from OpenAI, Google, and Anthropic pose a commoditization risk to EvenUp's Writing Layer capabilities; the defensible moat resides primarily in the proprietary PI dataset and the specialized Reading Layer models, not in document generation itself. | 中 | SE022, SE011 |
| CE042 | According to EvenUp's proprietary data, 42% of demand letters are sent more than 100 days after the plaintiff's last medical treatment, illustrating the scale of workflow delay that Smart Workflows is designed to address. | 中 | SE019, SE021 |
| CE043 | EvenUp's own analysis of platform cases found that 16.8% of plaintiffs develop a 30-day treatment gap within the first three months, rising to 32.4% after six months, and 43% of cases eventually experience gaps longer than 30 days. | 中 | SE021, SE019 |
| CE044 | Independent journalism flagged that whether PLAAS represents a genuinely new managed-service category or a repackaging of existing EvenUp AI capabilities with added headcount remains open, with outcomes at scale not yet validated. | 中 | SE020, SE022 |
| CE045 | EvenUp's public GitHub organization contains legacy DevOps infrastructure repositories (Puppet modules for backup management, DNS, log shipping, and security auditing) rather than active ML model or platform code repositories, indicating core AI development occurs in private repositories not publicly accessible. | 中 | SE017, SE016 |
| CU001 | EvenUp serves more than 2,000 personal injury law firms as of May 2026. | 高 | SU008, SU012 |
| CU002 | EvenUp processes more than 10,000 personal injury cases per week as of May 2026, representing over $14 billion in aggregate damages on the platform. | 中 | SU008 |
| CU003 | Approximately 30% of the top 100 US personal injury law firms use EvenUp as of May 2026. | 中 | SU008, SU007 |
| CU004 | ELG Injury Lawyers became EvenUp's second customer in May 2021. | 中 | SU002 |
| CU005 | John K. Zaid & Associates is a Houston, Texas PI firm with 180 attorneys and case staff that adopted EvenUp after a peer referral, beginning as a skeptic before committing to full platform deployment. | 高 | SU003, SU009 |
| CU006 | Mama Justice is a multi-state regional PI firm (Mississippi, Alabama, Tennessee) with 50+ employees across five offices that uses EvenUp's platform extensively for demand drafting and case management. | 中 | SU004 |
| CU007 | J&Y Law is a Los Angeles-based PI firm founded over 15 years ago that re-committed to EvenUp's full platform suite after a COO-led evaluation of their technology stack. | 中 | SU005 |
| CU008 | EvenUp's platform is used across multiple roles within PI firms: founding attorneys and managing partners (adoption decisions), COOs and legal ops leaders (change management), and paralegals, case managers, and demand writers (daily users). | 中 | SU002, SU003, SU004, SU005 |
| CU009 | EvenUp served more than 1,000 law firms at the time of its Series D announcement in October 2024. | 中 | SU018 |
| CU010 | EvenUp reported serving 1,500+ law firms at the AI Drafts Suite launch in May 2025, concurrent with the introduction of per-case pricing. | 高 | SU017, SU010 |
| CU011 | EvenUp's firm customer count grew from 1,000+ (October 2024) to 1,500+ (May 2025) to 2,000+ (October 2025 and reaffirmed May 2026), representing approximately 100% growth in 12 months. | 高 | SU008, SU018, SU012 |
| CU012 | EvenUp had resolved more than 200,000 cumulative personal injury cases as of October 2025, with victims recovering more than $10 billion. | 中 | SU012 |
| CU013 | Lightspeed Venture Partners' portfolio profile for EvenUp states the company grew from $0 to $10M in ARR in under two years after founding in 2019. | 中 | SU013 |
| CU014 | In July 2025, EvenUp analyzed thousands of AI interactions across more than 200 firms, finding 69% of prompts focused on medical treatment analysis and 30% of firms engaged in interactive follow-up prompting. | 中 | SU009 |
| CU015 | EvenUp's internal case data found that 16.8% of plaintiffs develop a 30-day treatment gap in the first three months, rising to 43% of all cases experiencing gaps longer than 30 days, informing the Medical Management product's design. | 中 | SU009 |
| CU016 | ELG Injury Lawyers reported 400% revenue growth within one year of fully adopting EvenUp's platform. | 中 | SU002 |
| CU017 | ELG Injury Lawyers reduced demand-letter turnaround from several months to a matter of days using EvenUp's Express Demands product. | 中 | SU002 |
| CU018 | John K. Zaid & Associates sends 30% more demand letters month-over-month without adding staff or overhead after deploying EvenUp's Demands and Express Demands. | 中 | SU003 |
| CU019 | John K. Zaid & Associates reported securing settlements 300% higher on low-value cases; a specific case that previously would have resolved below $10,000 received $30,000 in policy limits. | 中 | SU003 |
| CU020 | EvenUp's Communication Agents, deployed across 2,000 active Zaid cases and 7,500 client calls and texts, flagged acute issues requiring case manager follow-up in 37% of clients and identified missed appointments in 20%. | 高 | SU003, SU009 |
| CU021 | Mama Justice achieved 40% higher settlements and 14% faster case resolution without adding headcount after adopting EvenUp's Claims Intelligence Platform. | 中 | SU004 |
| CU022 | Mama Justice now routes 70% of demands through Express Demands, reducing per-demand time from over one hour to approximately 10 minutes. | 中 | SU004 |
| CU023 | J&Y Law reclaimed 320 hours per week of case manager capacity by automating negotiation data-gathering using Companion and AI Playbooks. | 中 | SU005 |
| CU024 | J&Y Law achieved a 25% decrease in case lifecycle timelines and a 50% reduction in medical chronology review time after full platform deployment. | 中 | SU005 |
| CU025 | Clark Fielding, founding principal of Fielding Law (Irvine, CA), stated in LawNext's December 2025 coverage that EvenUp's Medical Management enables 100% case readiness and real-time information access during depositions. | 中 | SU009 |
| CU026 | Glen Lerner, founding partner of Lerner and Rowe, stated as a PLAAS early customer that EvenUp 'delivers better case development and moves [cases] off the desk sooner, freeing up our best people to get more value on our biggest cases.' | 中 | SU008 |
| CU027 | Sweet James PI firm grew to over $500 million in annual results with 70% year-over-year growth using EvenUp without proportional headcount increases, as cited by COO Raymond Mieszaniec. | 中 | SU012 |
| CU028 | EvenUp has not publicly disclosed net revenue retention (NRR), gross revenue retention (GRR), or customer churn rate as of May 2026. | 高 | SU008, SU012 |
| CU029 | ELG Injury Lawyers has been an active EvenUp customer for more than four years, from May 2021 through at least May 2026, making it one of the longest-tenured documented customers. | 中 | SU002 |
| CU030 | J&Y Law explicitly expanded its EvenUp engagement from an earlier Demands-only relationship to the company's full platform suite, representing documented upsell within an existing account. | 中 | SU005 |
| CU031 | EvenUp launched its Personal Injury Pioneer Awards in March 2025, honoring ten PI firms for AI-driven innovation, creating a formal customer advocacy and loyalty mechanism with co-branded marketing and Customer Advisory Board membership. | 高 | SU006, SU016 |
| CU032 | In July 2025, EvenUp's analysis of 200+ firms found that 30% of firms engaged in follow-up AI prompting to refine outputs, indicating habitual and interactive platform usage rather than passive consumption. | 中 | SU009 |
| CU033 | PLAAS creates structural switching costs by embedding EvenUp's US-based case management staff directly into firm pre-litigation workflows, making exit more operationally disruptive than typical SaaS churn. | 中 | SU008 |
| CU034 | EvenUp's Customer Advisory Board, associated with the Pioneer Awards program, gives top customers formal product influence, which deepens organizational commitment and reduces churn risk for the highest-value accounts. | 中 | SU006 |
| CU035 | No multi-product adoption rate for EvenUp's full 2,000+ firm base has been publicly disclosed; named case study accounts show 3–7 products used, but this sample is non-representative and survivorship-biased. | 中 | SU002, SU003, SU004, SU005 |
| CU036 | PLAAS early testing generated more than $10 million in subscription sales as of the May 2026 launch announcement, according to EvenUp. | 中 | SU008 |
| CU037 | LawNext's coverage of the PLAAS launch raised the question of whether PLAAS represents a genuinely new managed-service category or a repackaging of existing managed legal services with an AI wrapper. | 中 | SU008 |
| CU038 | EvenUp's ~30% penetration of the top-100 PI firms creates revenue concentration risk; the highest-case-volume firms likely represent a disproportionate share of platform case throughput and revenue. | 中 | SU008, SU011 |
| CU039 | EvenUp's customer acquisition is partly channel-dependent through CMS integrations with Filevine, Litify, SmartAdvocate, and CasePeer; changes to those platforms' policies could disrupt discovery pipelines. | 中 | SU024, SU008 |
| CU040 | Supio, Filevine AI, and other AI legal tools target the same PI workflow as EvenUp, creating multi-homing risk where firms may adopt multiple platforms simultaneously rather than committing exclusively to EvenUp. | 中 | SU014, SU011 |
| CU041 | Insurance adjusters may adapt to the proliferation of AI-generated demand letters by developing counter-measures or discounting AI-sourced demands, potentially eroding the negotiating premium that EvenUp customers currently enjoy. | 低 | SU011, SU014 |
| CU042 | All published EvenUp customer ROI figures (400% revenue growth ELG, 300% settlement uplift Zaid, $500M Sweet James results, 40% higher settlements Mama Justice, 320 hours/week J&Y) are company-published testimonials or executive interviews with no independent third-party audit. | 中 | SU002, SU003, SU004, SU005, SU012 |
| CU043 | EvenUp's per-case pricing model, introduced in May 2025, may create budget friction for smaller PI firms (1–10 attorneys) where per-case costs are harder to forecast and absorb relative to flat SaaS fees. | 中 | SU017, SU010 |
| CR001 | EvenUp qualifies as a HIPAA business associate because it processes protected health information (medical records, treatment summaries, billing statements) on behalf of PI law firms that obtain that data from healthcare providers. | 高 | SR001, SR013 |
| CR002 | The HIPAA Security Rule requires EvenUp, as a business associate, to implement administrative, physical, and technical safeguards to protect electronic PHI, with direct OCR enforcement authority since the HITECH Act. | 高 | SR002, SR001 |
| CR003 | Under the HIPAA Breach Notification Rule, EvenUp as a business associate must notify covered-entity clients within 60 days of discovering a breach of unsecured PHI; EvenUp's own DPA commits to notification within 72 hours. | 高 | SR003, SR013 |
| CR004 | EvenUp launched PLAAS (Pre-Litigation as a Service) in May 2026, expanding from software vendor to outsourced legal-operations partner and increasing exposure to UPL, HIPAA, quality, and people risks simultaneously. | 中 | SR027 |
| CR005 | In Mata v. Avianca (S.D.N.Y. 2023), a federal court imposed sanctions on attorneys who submitted AI-generated legal briefs containing fabricated case citations, establishing a precedent for attorney liability tied to unverified AI outputs. | 高 | SR008, SR020 |
| CR006 | ABA Model Rule 1.1 (Competence) requires attorneys to maintain competence in the technology they use; state bars and the ABA have issued guidance interpreting this to include understanding AI tools used in legal practice. | 高 | SR010, SR023 |
| CR007 | ABA Model Rule 1.6 (Confidentiality) restricts attorneys from disclosing client information to third parties including AI vendors without informed consent or a showing that the disclosure is impliedly authorized to carry out the representation. | 高 | SR011, SR009 |
| CR008 | ABA Model Rule 5.5 prohibits lawyers from assisting another person in the unauthorized practice of law; as EvenUp's PLAAS model involves non-attorney legal-ops staff performing case analysis, state bars may scrutinize whether those activities constitute UPL. | 中 | SR012, SR009 |
| CR009 | No public enforcement actions, bar complaints, or court sanctions specifically naming EvenUp or its AI outputs as the cause of attorney misconduct have been identified as of May 2026. | 中 | SR023, SR024 |
| CR010 | EvenUp's own 2025 benchmarks report found that PI firms miss an average of five documents and bills per 10 cases before using EvenUp, implying a material pre-existing error rate in the market EvenUp targets. | 中 | SR014 |
| CR011 | EvenUp's Missing Docs Check feature reportedly delivers a 75% reduction in missing documents per case, but the company's own benchmarks acknowledge residual gaps in approximately one of every four previously-missing documents. | 中 | SR014 |
| CR012 | EvenUp's DPA defines a Security Incident as a material breach leading to unauthorized third-party access to Customer Personal Data, and commits to notifying clients within 72 hours; but PHI is governed by a separate BAA, not the DPA. | 中 | SR013 |
| CR013 | AI hallucination in legal document generation is an industry-wide risk; Supio (a competing PI AI platform) publicly acknowledges that any AI output carries a non-zero probability of error and that legal hallucinations can destroy a case or cost an attorney their license. | 高 | SR025, SR008 |
| CR014 | EvenUp's engineering blog describes a proprietary human-in-the-loop feedback system in which thousands of expert corrections are ingested daily to retrain domain models, functioning as the primary mitigation against AI hallucination at scale. | 中 | SR015, SR016 |
| CR015 | EvenUp's human reviewer layer is the load-bearing mitigation across regulatory, quality, and legal risk categories; its effectiveness depends on reviewer-to-case ratio remaining adequate as PLAAS scales. | 中 | SR014, SR015 |
| CR016 | EvenUp's DPA does not include a service-level agreement guaranteeing platform uptime; a cloud or platform outage could halt demand-letter production during time-sensitive legal deadlines. | 中 | SR013 |
| CR017 | Filevine, EvenUp's largest CMS integration partner, has developed its own AI layer called LOIS, creating a latent competitive conflict between the CMS platform and EvenUp as an add-on AI vendor. | 中 | SR029, SR027 |
| CR018 | RELX, a strategic investor in EvenUp's Series D and E rounds, is the parent company of LexisNexis, which competes directly in the legal AI space with its Lexis+ Protégé product; this creates a structural conflict of interest between investor and investee. | 中 | SR017, SR031, SR018 |
| CR019 | LexisNexis's Lexis+ Protégé offers AI-assisted drafting, research, and analysis for legal professionals, covering functionality adjacent to EvenUp's demand-letter and case-preparation products. | 高 | SR031, SR018 |
| CR020 | The FTC has flagged that a handful of cloud providers dominate AI compute infrastructure and that control over these inputs could allow incumbents to dampen competition in generative AI markets. | 高 | SR006, SR007 |
| CR021 | EvenUp does not publicly disclose its LLM provider(s), cloud infrastructure vendor, or multi-cloud architecture; the concentration of its AI infrastructure on undisclosed counterparties is an unresolved diligence gap. | 中 | SR014, SR015 |
| CR022 | EvenUp integrates with Filevine, Litify, CasePeer/Clio, and SmartAdvocate; these four platforms collectively serve the majority of mid-to-large PI firms, making EvenUp's data-ingestion pipeline dependent on their API policies. | 中 | SR029, SR027 |
| CR023 | Harvey, a competing legal AI platform, was valued at approximately $11 billion in 2026 and is actively targeting the broader legal-AI market; its entry into PI-specific applications would represent a major competitive threat to EvenUp's pricing power. | 中 | SR019, SR030 |
| CR024 | EvenUp's three co-founders — Rami Karabibar, Raymond Mieszaniec, and Saam Mashhad — and its chief scientist Haixun Wang (ACM Fellow) represent significant key-person concentration in a company whose moat depends on proprietary AI research and client relationships. | 中 | SR015, SR016 |
| CR025 | EvenUp actively recruits specialized ML engineers from competitive AI firms, and blog posts highlight competitive hiring; attrition in a market where Harvey and other well-capitalised AI firms are recruiting remains a persistent execution risk. | 中 | SR015, SR016, SR019 |
| CR026 | EvenUp's engineering culture blog posts describe a collaborative, mission-driven environment that retains talent; these are company-sourced and do not represent independent verification of attrition rates or employee satisfaction. | 低 | SR015, SR016 |
| CR027 | Scaling PLAAS requires parallel ramp of trained legal-ops reviewers; quality degradation is a lagging indicator and errors in demand letters may not surface until adjuster negotiations reveal inaccuracies already propagated across many cases. | 中 | SR014, SR027 |
| CR028 | Law firm partners who adopt EvenUp must review AI outputs to satisfy ABA Rule 1.1 competence obligations; if EvenUp's PLAAS scales faster than law firm supervisory capacity, the ethics exposure transfers to those firms and affects renewal rates. | 中 | SR010, SR014 |
| CR029 | EvenUp has raised $385M total across multiple rounds, reaching a $2B+ valuation after its October 2025 Series E; no ARR, gross margin, or burn rate data has been publicly disclosed. | 中 | SR027 |
| CR030 | PLAAS — a managed legal-operations service requiring 300+ staff — is fundamentally more capital-intensive than pure SaaS, and gross margins for managed-services businesses typically run 30–50%, well below typical SaaS gross margins of 70–80%. | 中 | SR027, SR014 |
| CR031 | Harvey AI, valued at approximately $11B, and frontier AI providers (Anthropic, OpenAI) actively expanding into legal verticals represent competitive pricing pressure on EvenUp's per-case fee structure. | 中 | SR019, SR030 |
| CR032 | If insurance adjusters systematically identify AI-generated demand letters and apply discounts or heightened scrutiny, the settlement-value uplift that drives EvenUp's ROI case for law firms would erode, potentially reducing renewal rates. | 低 | SR014, SR028 |
| CR033 | Healthcare data breach statistics show that the number of individuals affected by healthcare breaches reached 289 million in 2024, and that business associates are a primary attack surface; this benchmark frames the probability of a breach incident affecting EvenUp. | 高 | SR026, SR003 |
| CR034 | EvenUp's per-case revenue model creates financial volatility correlated with PI caseload volumes, which in turn depend on auto-accident rates, economic cycles, and litigation trends — factors outside the company's control. | 中 | SR014, SR028 |
| CR035 | EvenUp's DPA requires clients to accept responsibility for securing systems EvenUp uses to deliver the service and for backing up Customer Personal Data, partially shifting operational security obligations to client firms. | 中 | SR013 |
| CR036 | EvenUp's proprietary feedback loop — ingesting thousands of daily expert corrections into domain models — is a structural mitigation against hallucination but is self-reported; no independent audit or third-party validation of error-rate reduction has been identified. | 中 | SR014, SR015 |
| CR037 | NIST's AI Risk Management Framework (AI RMF 1.0) is a voluntary standard for trustworthy AI that regulators and courts increasingly cite as a governance benchmark; EvenUp's compliance posture relative to this framework is unknown. | 中 | SR005 |
| CR038 | The FTC's GenAI competition analysis identifies data, talent, and compute concentration as key barriers that could allow incumbents to leverage control over AI inputs to dampen competition; RELX/LexisNexis holds all three relative to EvenUp's PI dataset. | 中 | SR006, SR017, SR018 |
| CR039 | EvenUp's DPA commits to providing clients with SOC 2 or equivalent audit reports upon request but classifies them as confidential; no publicly accessible SOC 2 Type II certificate has been identified. | 中 | SR013 |
| CR040 | State bars and the ABA have increasingly issued guidance on attorney use of AI tools, with the ABA Journal reporting in May 2026 that a lawyer was disqualified in a morgue-scandal case because of past AI errors; this signals courts and bars actively monitoring AI misconduct. | 中 | SR023 |
| CR041 | EvenUp's BAA template for PHI processing is separate from its DPA and is not publicly available; the adequacy of BAA terms — particularly indemnification, cure obligations, and breach-response timelines — cannot be assessed from public sources. | 中 | SR013, SR001 |
| CR042 | The HHS OCR breach portal provides a mechanism for individuals or organizations to report HIPAA breaches affecting 500+ people; any breach naming EvenUp would become public record on the OCR 'Wall of Shame' database. | 高 | SR004, SR003 |
| CV001 | EvenUp's May 2026 recommendation is research-more; confidence is medium due to strong qualitative product-market fit evidence but no disclosed ARR, gross margin, or EBITDA. | 中 | SV001, SV005, SV006 |
| CV002 | EvenUp's risk rating is high, driven by margin compression risk from 300+ US legal-ops staff, HIPAA and attorney-ethics regulatory exposure, and AI commoditization from Harvey, OpenAI, and Filevine. | 中 | SV006, SV024 |
| CV003 | EvenUp's valuation stance is expensive: a $2B+ private valuation with no disclosed ARR implies a revenue multiple of 13–40x depending on ARR assumption, which exceeds public-market legal SaaS comparables except for the highest-growth tier. | 中 | SV003, SV005, SV009 |
| CV004 | A new investor at $2B+ should demand audited ARR, gross-margin, and burn-rate disclosure, and should price the round assuming blended gross margins of 50–65% to reflect PLAAS's managed-services component. | 中 | SV001, SV006, SV010 |
| CV005 | EvenUp serves 2,000+ personal injury law firms as of May 2026, including approximately 30% of the top-100 PI firms by revenue. | 中 | SV006, SV029 |
| CV006 | EvenUp processed over 200,000 cases and represents over $14 billion in damages as of the PLAAS launch in May 2026. | 中 | SV006, SV035 |
| CV007 | EvenUp's PLAAS managed services combined AI with 300+ US-based legal operations staff who handle the full pre-litigation lifecycle including claim setup, care coordination, records retrieval, demand preparation, and settlement negotiation. | 高 | SV006, SV035 |
| CV008 | PLAAS early testing generated over $10 million in subscription sales before the May 2026 public launch, the only concrete revenue data point publicly disclosed by EvenUp. | 高 | SV006, SV029, SV035 |
| CV009 | EvenUp has raised $385 million in total disclosed venture equity capital across at least five rounds (Seed through Series E) as of May 2026; all capital is equity with no disclosed debt financing. | 高 | SV001, SV003, SV006 |
| CV010 | EvenUp raised $135 million in Series D funding on October 8, 2024, at a post-money valuation exceeding $1 billion; the round was led by Bain Capital Ventures with participation from Premji Invest, Lightspeed, Bessemer, SignalFire, and B Capital. | 高 | SV001, SV002, SV004 |
| CV011 | The Series D raised total capital to $235 million with $220 million raised in the eighteen months preceding October 2024, indicating accelerating funding pace. | 高 | SV001, SV002 |
| CV012 | EvenUp raised $150 million in Series E funding on October 1, 2025, at a post-money valuation exceeding $2 billion; the round was led by Bessemer Venture Partners with new participation from REV (RELX/LexisNexis VC arm) and continuing participation from B Capital and Bain Capital. | 高 | SV003, SV005, SV033 |
| CV013 | An EDGAR full-text search for 'EvenUp, Inc.' returns four ServiceTitan DRS/A filings, identifying EvenUp as a Delaware private company with board representation beginning April 2023; no independent Form D for EvenUp has been identified in SEC EDGAR. | 高 | SV020, SV021 |
| CV014 | RELX, the parent of LexisNexis, invested in EvenUp's Series E through its venture arm REV, creating a strategic-investor relationship with a direct competitor in legal research and AI-assisted legal services. | 高 | SV012, SV025, SV033 |
| CV015 | The RELX/REV investment in EvenUp's Series E introduces a conflict-of-interest scenario: RELX parent LexisNexis competes in the AI-assisted legal services market, and REV's board or observer access gives RELX insight into EvenUp's proprietary PI dataset and PLAAS economics. | 中 | SV025, SV005 |
| CV016 | EvenUp's Series E represents a 100%-plus step-up from the Series D valuation (from $1B+ to $2B+) in approximately twelve months, implying strong investor confidence in the PLAAS model and continued growth. | 高 | SV001, SV003 |
| CV017 | EvenUp's co-founder and CEO referenced an IPO as a possible future path, per Axios reporting from October 2025; no S-1 or IPO process has been publicly initiated as of May 2026. | 中 | SV005, SV029 |
| CV018 | Harvey AI, a general legal AI platform serving BigLaw and enterprise clients, was valued at approximately $3 billion after raising $300 million in a Series D round in February 2025 per Bloomberg Law reporting. | 高 | SV014, SV015 |
| CV019 | Casetext, an AI-powered legal research platform, was acquired by Thomson Reuters for $650 million in June 2023, establishing a floor M&A reference for legal AI acquisitions. | 高 | SV016, SV032 |
| CV020 | CoCounsel, the Thomson Reuters AI product built on the Casetext acquisition, competes in the broader legal AI market but focuses on legal research rather than personal injury case management and PLAAS services. | 高 | SV016, SV032 |
| CV021 | Clio, the leading legal practice management SaaS, was valued at approximately $3 billion in its most recent disclosed funding round; analyst references suggest ARR at or above $100 million at that valuation, implying a ~30x ARR multiple. | 中 | SV019, SV027 |
| CV022 | ServiceTitan, a vertical SaaS company for field service businesses, IPO'd in December 2024 at approximately $8 billion enterprise value, at roughly 12–16x forward revenue—a useful public-market anchor for vertical SaaS. | 高 | SV020, SV021 |
| CV023 | The global legal AI market was valued at $1.45 billion in 2024 and is projected to reach $3.90 billion by 2030 at a 17.3% CAGR, per Grand View Research. | 中 | SV009 |
| CV024 | The legal AI software market is estimated at $3.11 billion in 2025 and projected to reach $10.82 billion by 2030 at a 28.3% CAGR, per MarketsAndMarkets. | 中 | SV010 |
| CV025 | The US personal injury lawyers and attorneys industry had market revenue of approximately $61.7 billion in 2025, per IBISWorld, reflecting a large but slowly-growing addressable market. | 中 | SV011, SV012 |
| CV026 | EvenUp's $2B+ valuation represents approximately 0.27% of the $61.7 billion US PI services market, implying a market-penetration valuation hypothesis rather than a current-revenue multiple. | 中 | SV003, SV011 |
| CV027 | In the bull scenario, EvenUp reaches $150–200 million in blended ARR by end of 2026 as PLAAS scales to 100+ firms, supporting a $4–6 billion exit at 2–3x Series E returns; this assumes 65–70% gross margin. | 低 | SV006, SV013 |
| CV028 | In the base scenario, EvenUp reaches $50–80 million in blended ARR by end of 2026 with PLAAS at 50–100 firms and 55–60% gross margins, supporting a $3–4 billion exit at 1.5–2x Series E returns. | 低 | SV001, SV013 |
| CV029 | In the bear scenario, PLAAS adoption is slow, AI commoditization compresses per-case pricing, and blended ARR stagnates at $40–60 million; an exit at $800M–$1.2B would produce a loss for Series E investors. | 低 | SV024, SV032 |
| CV030 | The valuation sensitivity analysis shows that a 10-percentage-point change in gross margin at $100M ARR and a 15x revenue multiple changes implied enterprise value by approximately $300–500 million. | 低 | SV009, SV010 |
| CV031 | The Lightspeed portfolio page confirms EvenUp reached $10 million in ARR in under two years from founding (by approximately 2021), demonstrating early revenue velocity before the product scaled to its current size. | 高 | SV013, SV001 |
| CV032 | EvenUp's $385 million in total preferred capital creates meaningful liquidation preference overhang; a $1 billion acquisition would likely be fully absorbed by preferred liquidation preferences, and Series E returns require a premium exit. | 中 | SV001, SV003 |
| CV033 | EvenUp's 300+ US-based legal operations staff implies an estimated annual fully-loaded labor cost of $45–75 million, a structural gross margin drag absent from pure-software legal AI peers like Harvey. | 中 | SV006, SV007 |
| CV034 | If PLAAS subscription revenue covers the fully-loaded legal-ops cost of $45–75 million annually, the underlying SaaS platform could approach 70–80% gross margins; if not, blended margins remain below 55%, a structural weakness at $2B+ valuation. | 中 | SV006, SV023 |
| CV035 | EvenUp's co-founder and CEO Rami Karabibar stated the company processes more than 10,000 cases per week as of the PLAAS launch in May 2026. | 中 | SV006, SV029 |
| CV036 | The primary thesis-break trigger is PLAAS gross margin falling below 45%, which would indicate the managed-services layer cannot cover legal-ops labor costs and would justify revaluing EvenUp as a services company at 3–5x revenue rather than software multiples. | 中 | SV006, SV023 |
| CV037 | Customer churn exceeding 15% annually among the top-100 PI firm cohort would invalidate EvenUp's switching-cost claims and suggest competitive substitution is underway. | 中 | SV005, SV024 |
| CV038 | Supio AI explicitly markets its legal AI platform as source-linked and verifiable to address attorney hallucination risk, representing a direct competitive response to the quality risk that applies to EvenUp's demand-letter and medical-chronology outputs. | 中 | SV024, SV032 |
| CV039 | The specific governance terms of RELX/REV's EvenUp investment—including anti-compete covenants, data-sharing rights, and right-of-first-offer—have not been publicly disclosed; this is a material diligence gap for incoming Series E investors. | 中 | |
| CV040 | Audited ARR and gross-margin disclosure is the highest-priority diligence item for an incoming Series E investor; without it, the $2B+ implied revenue multiple cannot be validated or benchmarked against legal-tech peers. | 高 | SV003, SV009 |
| CV041 | PLAAS unit economics—specifically revenue per contract per year, gross margin per PLAAS subscriber, legal-ops headcount per contract, and case yield per staff member—are not publicly disclosed and are the second-priority diligence item. | 高 | SV006, SV035 |
| CV042 | Customer net revenue retention for both the per-case SaaS product and PLAAS is not publicly disclosed; NRR above 120% would validate the upsell thesis while NRR below 100% would indicate churn-driven revenue risk. | 中 | SV005, SV006 |
| CV043 | The IP ownership and licensing structure for EvenUp's Piai model—including third-party LLM provider contracts, fine-tuning data rights, and portability on change of control—has not been publicly disclosed and is a material diligence item for IPO or M&A scenarios. | 中 | SV001, SV035 |
| CV044 | Harvey AI serves BigLaw and enterprise clients rather than the personal injury plaintiffs' market, meaning Harvey's $3B valuation reflects a different addressable market, customer profile, and absence of PHI-handling complexity relative to EvenUp. | 高 | SV014, SV015 |
| CV045 | Bessemer Venture Partners participated in EvenUp's Series D and led the Series E, representing a confidence re-up by a lead investor with deep vertical SaaS expertise and the Roads to $100M ARR legal AI playbook. | 高 | SV001, SV003, SV023 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | EvenUp | EvenUp — Homepage | 2,000+ personal injury firms trust EvenUp to handle their pre-litigation workflows. |
| SO002 | EvenUp | EvenUp — About Us | Co-founders: Rami Karabibar (CEO), Raymond Mieszaniec (COO), Saam Mashhad (Chief of Product & Legal Operations). |
| SO003 | EvenUp | EvenUp — Blog | |
| SO004 | EvenUp | EvenUp — Newsroom | |
| SO005 | EvenUp | EvenUp — Careers | Forbes Cloud 100 named; flexible hybrid work model; 20% of top 100 US PI firms. |
| SO006 | EvenUp | EvenUp — Products | |
| SO007 | EvenUp | Introducing AI Drafts Suite, Smart Workflows, Medical Bill Summary, and Case-Based Pricing | Backed by Bessemer, BCV, Lightspeed, SignalFire, NFX, DCM. 1,500+ PI firms. $7B in claims generated. |
| SO008 | Above the Law | From Startup to $2 Billion: EvenUp Is Transforming Personal Injury Practice | Of course, the standard concerns with AI tools remain: algorithmic bias, data privacy, and for EvenUp, whether its tools will further gum up the court system. |
| SO009 | Unite.AI | Raymond Mieszaniec, Co-Founder & COO of EvenUp — Interview Series | EvenUp was founded in 2019. Co-founders: Rami Karabibar, Raymond Mieszaniec. Series E $150M at $2B+. 200,000+ cases resolved. 300+ legal ops team. $10B+ recovered. |
| SO010 | SignalFire | EvenUp — SignalFire Portfolio | Founders: Rami Karabibar, Raymond Mieszaniec, Saam Mashhad. |
| SO011 | LawnNext | EvenUp Launches Medical Management Tool to Address Treatment Gaps in Personal Injury Cases | Series E led by Bessemer with REV (RELX/LexisNexis), B Capital, Bain Capital. CEO: Rami Karabibar. 200,000+ cases. |
| SO012 | LawnNext | EvenUp Extends Beyond Software With Launch of Pre-Litigation as a Service Offering for PI Law Firms | 30% of top 100 PI firms. 10,000 cases/week. $14B in damages. $10M PLAAS subscriptions already sold. |
| SO013 | Bloomberg Law | AI Legal-Tech EvenUp Raises $150 Million at $2 Billion Valuation | Total capital raised = $385M. 2,000+ firms including 20% of Top 100 US PI firms. |
| SO014 | Bessemer Venture Partners | Ten Principles for Building Strong Vertical AI Businesses | |
| SO015 | Bessemer Venture Partners | Roads to $100M ARR — Legal AI | |
| SO016 | Lawsites (Robert Ambrogi) | EvenUp Launches Medical Management Tool | |
| SO017 | Forbes | Forbes Cloud 100 — 2025 List | |
| SO019 | Reuters | AI Boom Fuels Fresh Wave of Legal Tech Investments | |
| SO020 | Crunchbase | EvenUp — Crunchbase Organization Profile | |
| SO021 | Bloomberg Law | EvenUp Raises $135 Million in Series D Funding | |
| SO022 | Bloomberg Law | Investors Pour Cash Into AI Startups for Plaintiffs Lawyers | |
| SO023 | PR Newswire | EvenUp Introduces Express Demands, the First in Its Series of AI Documents | |
| SO024 | EvenUp — LinkedIn Company Page | ||
| SO025 | LawnNext | LawnNext Search Results — EvenUp Series D | |
| SO027 | AP News | EvenUp Announces $135 Million Series D | |
| SM001 | EvenUp | EvenUp Homepage — Proactive AI for Personal Injury Firms | Why 2,000+ Personal Injury Firms Choose EvenUp. #1 Proactive AI Platform for PI. Built on the Largest PI Dataset. |
| SM002 | EvenUp | EvenUp Blog — Articles and Case Studies | ELG Injury Lawyers Achieves 400%+ Revenue Growth Using AI Tech Built for Personal Injury Firms |
| SM003 | Grand View Research | Legal AI Market Size, Share & Trends — Industry Report, 2030 | The global legal AI market size was valued at USD 1.45 billion in 2024 and is projected to reach USD 3.90 billion by 2030, growing at a CAGR of 17.3% from 2025 to 2030. |
| SM004 | MarketsandMarkets | Legal AI Software Market by Offering — Global Forecast to 2030 | The legal AI software market is expected to grow from USD 3.11 billion in 2025 to USD 10.82 billion by the year 2030, at a CAGR of 28.3% during the forecast period. |
| SM005 | U.S. Bureau of Labor Statistics | Lawyers — Occupational Outlook Handbook | Number of Jobs, 2024: 864,800. Job Outlook, 2024–34: 4% (As fast as average). 2024 Median Pay: $151,160 per year. |
| SM006 | American Bar Association | Model Rules of Professional Conduct — Rule 1.1: Competence | A lawyer shall provide competent representation to a client. Competent representation requires the legal knowledge, skill, thoroughness and preparation reasonably necessary for the representation. |
| SM007 | American Bar Association | Model Rules of Professional Conduct — Rule 1.6: Confidentiality of Information | A lawyer shall not reveal information relating to the representation of a client unless the client gives informed consent... |
| SM008 | American Bar Association | Model Rules of Professional Conduct — Rule 5.5: Unauthorized Practice of Law | A lawyer shall not practice law in a jurisdiction in violation of the regulation of the legal profession in that jurisdiction, or assist another in doing so. |
| SM009 | Cornell Law School Legal Information Institute | Personal Injury — LII Wex Legal Encyclopedia | Personal injuries include every variety of injury to a person's body, emotions, or reputation, as contradistinguished from injury to property rights. |
| SM010 | Insurance Information Institute | Facts + Statistics: Auto Insurance | The countrywide average auto insurance expenditure increased 6.1 percent to $1,127 in 2022 from $1,062 in 2021. |
| SM011 | Above the Law | From Startup to $2 Billion: EvenUp Is Transforming Personal Injury Practice | The standard concerns with AI tools remain: algorithmic bias, data privacy, and for EvenUp, whether its tools will further gum up the court system. |
| SM012 | LawNext | EvenUp Launches Medical Management Tool to Address Treatment Gaps in Personal Injury Cases | EvenUp's analysis of cases on its platform found that 16.8% of plaintiffs develop a 30-day gap in treatment within the first three months of their case. That figure rises to 32.4% after six months, with 43% of cases eventually experiencing gaps longer than 30 days. |
| SM013 | LawNext | EvenUp Extends Beyond Software with Launch of 'Pre-Litigation-as-a-Service' Offering For PI Law Firms | EvenUp says it is used by 30% of the top 100 PI firms and processes more than 10,000 cases per week, representing over $14 billion in damages. |
| SM014 | Bloomberg Law | AI Legal Tech EvenUp Raises $150 Million at $2 Billion Valuation | EvenUp's Claims Intelligence Platform is used for drafting, reviewing, and strategizing across the case lifecycle. It said it has more than 2,000 firms, including 20% of the Top 100 U.S. personal injury firms, among its... |
| SM015 | Legal IT Insider | Legal IT's Leading News Source — Homepage | |
| SM016 | Stanford Law School — CodeX | CodeX: The Stanford Center for Legal Informatics | CodeX is a global epicenter for research and development of computational law. |
| SM017 | HIPAA Journal | Healthcare Data Breach Statistics — Updated for 2026 | Between October 21, 2009, when OCR first started publishing summaries of data breach reports on its 'Wall of Shame', and January 31, 2026, 7,419 large healthcare data breaches have been reported to OCR. |
| SM018 | U.S. Courts | Caseload Statistics Data Tables | |
| SM019 | EvenUp | EvenUp About Us Page | |
| SM020 | EvenUp | EvenUp Careers Page | Thousands of firms, including 20% of the Top 100 U.S. personal injury firms, depend on EvenUp's platform. |
| SM021 | EvenUp | EvenUp Newsroom | |
| SM022 | EvenUp | EvenUp Products Page | |
| SM023 | IBISWorld | Personal Injury Lawyers & Attorneys in the US — Industry Analysis 2025 | The Personal Injury Lawyers & Attorneys industry in the United States is highly fragmented with no companies holding a market share greater than 5%. |
| SM024 | Bloomberg Law | Investors Pour Cash Into AI Startups for Plaintiffs' Lawyers | |
| SM025 | ABA Journal | ABA Journal — Homepage and Recent Coverage | |
| SP001 | EvenUp | EvenUp Homepage — Proactive AI for Personal Injury Firms | Why 2,000+ Personal Injury Firms Choose EvenUp. #1 Proactive AI Platform for PI. Built on the Largest PI Dataset. |
| SP002 | Supio | Supio Homepage — Legal AI for Personal Injury Law Firms | Supio is the only agentic legal AI platform built for plaintiff law and mass torts cases. |
| SP003 | Supio | About Supio — Mission, Team, and Offices | Founded 2021. Offices: Seattle, San Francisco. Join hundreds of personal injury firms using Supio. |
| SP004 | Supio | Supio Pricing Plans — Case Subscription and Unlimited Firm Access | Risk free: We don't require lock-in contracts or platform fees. 2 ways to buy Supio: Case Subscription and Unlimited Firm Access. |
| SP005 | Bloomberg Law | AI Legal Tech EvenUp Raises $150 Million at $2 Billion Valuation | EvenUp raised $150 million at a $2 billion+ valuation in its Series E round. |
| SP006 | EvenUp | Introducing AI Drafts, Smart Workflows, Medical Bill Summary, and Case-Based Pricing | Smart Workflows leverage seamless integrations with CRMs such as SmartAdvocate, Litify, and CasePeer. 42% of Demands are sent >100 days after last treatment. |
| SP007 | Bloomberg Law | EvenUp Raises $135 Million in Series D Funding | EvenUp raised $135 million in a Series D funding round. |
| SP008 | Supio | Why AI in Legal Needs a Higher Standard — Supio and Thomson Reuters Westlaw Advantage | Our partnership with Thomson Reuters is built to solve citation verification. We announced the next step: providing a direct channel for Supio users to access Westlaw Advantage without leaving the platform. |
| SP009 | Darrow AI | Darrow AI Homepage — Legal Exposure Management | Darrow is the leader in Legal Exposure Management — working upstream to surface hidden signals and transform them into structured legal intelligence, before harm compounds. |
| SP010 | Filevine | Filevine Features — AI Medchron, Court Reporting, Timely | Apr 2025: FilevineAI analyzes your uploaded medical records to surface what matters most — automatically. Aug 2025: Court Reporting Software — manage the entire deposition lifecycle in a single AI-powered platform. |
| SP011 | Filevine | Filevine Homepage — Legal AI and Operating Intelligence | A singular system of truth for the modern practice, powered by LOIS. The 2026 Legal AI Trust Index: trust does not come from general-purpose tools operating in isolation. |
| SP012 | Litify | Litify Plaintiff Practice Management — Agentic Case Expert (ACE) | Introducing Litify ACE: Agentic Case Expert. Instantly generate accurate, source-linked medical chronologies and draft demand packets. |
| SP013 | Clio | Clio Work — AI Legal Platform for Case Analysis, Research, and Drafting | The only AI that understands your cases, their context, and the law. Clio Work brings together everything relevant to your matter. |
| SP014 | CasePeer | CasePeer — Personal Injury Case Management Software | CasePeer is the only case management software driving better outcomes for your clients, and better outcomes for your firm. Turnkey platform for plaintiffs' law firms, no customization required. |
| SP015 | SmartAdvocate | SmartAdvocate — Legal Case Management with Built-In AI | SmartIntelligence, SmartAdvocate's Built-In AI Tools, are revolutionizing case management by boosting precision, streamlining routine work. |
| SP016 | Harvey | Harvey Blog — Legal Agents for Every Matter | Legal Agents for Every Matter, Tailored to You. We're introducing improved agentic capabilities in Harvey. |
| SP017 | Harvey | Harvey AI Homepage | Harvey — enterprise legal AI platform targeting AmLaw firms and corporate counsel. |
| SP018 | Thomson Reuters | CoCounsel Homepage — AI-Powered Legal Assistant | CoCounsel — Thomson Reuters AI legal assistant grounded in Westlaw and Practical Law. |
| SP019 | Bessemer Venture Partners | Roads to $100M ARR: Legal AI | Thomson Reuters acquired Casetext (CoCounsel) for approximately $650M in 2023. |
| SP020 | LexisNexis | Lexis+ with Protégé — Legal AI for Drafting, Research, and Analysis | Lexis+ with Protégé is a legal AI workflow solution for drafting, research, and analysis. It combines the LexisNexis Protégé AI assistant with trusted sources. |
| SP021 | LawNext | EvenUp Extends Beyond Software with Launch of Pre-Litigation-as-a-Service | EvenUp's pre-litigation-as-a-service offering expands beyond software—charging per case for fully managed demand preparation. |
| SP022 | EvenUp | EvenUp Blog — 5 Reasons Claude Isn't Right for Your Personal Injury Law Firm | 5 Reasons Claude Isn't Right for Your Personal Injury Law Firm — blog post title listed in EvenUp blog, May 2026. |
| SP023 | Above the Law | From Startup to $2 Billion: EvenUp Is Transforming Personal Injury Practice | Above the Law notes concerns about standardized AI demands potentially gumming up courts and insurance adjusters developing systematic responses to EvenUp-style letters. |
| SP024 | LawNext | EvenUp Launches Medical Management Tool for Personal Injury Cases | EvenUp's Medical Management product addresses treatment gaps in personal injury cases. |
| SP025 | Filevine | Filevine Pricing — Custom-Built Packages | All packages are custom built for your team's needs. LOIS — Agent-assisted drafting, in-document execution, source-linked recommendations. |
| SI001 | EvenUp | EvenUp Announces $135 Million Series D | EvenUp has raised $135 million in Series D funding led by Bain Capital Ventures. EvenUp's team doubled in size in the past 12 months, and the company saw more than 100% year-over-year revenue growth. |
| SI002 | U.S. Securities and Exchange Commission (EDGAR) | ServiceTitan, Inc. DRS/A Registration Statement (Exhibit identifying EvenUp board member) | EvenUp, Inc., a privately held software company, since April 2023 |
| SI003 | Bloomberg Law | EvenUp Raises Series D at $1B Valuation for Personal Injury AI | EvenUp raised a Series D round at a more than $1 billion valuation for its personal-injury AI platform. |
| SI004 | Bloomberg Law | EvenUp Raises $150 Million for AI That Helps Personal Injury Firms | EvenUp raised $150 million led by Bessemer Venture Partners, with participation from REV, the venture arm of RELX, parent of LexisNexis. |
| SI005 | Bloomberg Law | EvenUp Legal AI Raises $150 Million in Series E Funding Round | EvenUp, a legal AI company focused on personal injury law, raised $150 million in a Series E round. |
| SI006 | Grand View Research | Legal Services Market Size, Share & Trends Report | The global legal services market is valued in the hundreds of billions annually with significant AI-driven automation opportunity. |
| SI007 | IBISWorld | Personal Injury Lawyers & Attorneys in the US — Market Size Statistics | The US Personal Injury Lawyers & Attorneys industry generates over $60 billion in annual revenue across tens of thousands of firms. |
| SI008 | Clio | Clio Legal Trends Report — Annual Legal Industry Benchmarking | AI adoption among law firms accelerated significantly in 2025, with personal injury firms showing above-average technology investment rates. |
| SI009 | EvenUp | EvenUp Homepage — Proactive AI for Personal Injury Firms | Why 2,000+ Personal Injury Firms Choose EvenUp. ~10,000 cases per week. $14B+ in damages for our clients' clients. |
| SI010 | EvenUp | Introducing AI Drafts, Smart Workflows, Medical Bill Summary, and Case-Based Pricing | One clear, predictable cost per case. Access the full Claims Intelligence Platform with one clear, predictable cost per case. Jeffcoat Injury Lawyers: 3x more demand letters. Settled cases 30 days faster. |
| SI011 | Unite.ai | Raymond Mieszaniec, Co-Founder & COO of EvenUp — Interview Series | EvenUp has raised $150M in Series E at a valuation greater than $2 billion. Sweet James has achieved $500M+ in annual results with 70% year-over-year growth. |
| SI012 | LawNext | EvenUp Extends Beyond Software with Launch of Pre-Litigation as a Service (PLAAS) for PI Law Firms | EvenUp has already sold more than $10 million in subscriptions to PLAAS during early testing. Early results include 95% of available third-party policy limits, records retrieved 66 days faster, demand packets 47 days faster, $1,000 per case in cost savings. |
| SI013 | LawSitesBlog | EvenUp Launches Medical Management Tool to Address Treatment Gaps in Personal Injury Cases | EvenUp's Series E was led by Bessemer Venture Partners and included REV, the venture arm of RELX (parent company of LexisNexis), B Capital, and Bain Capital. |
| SI014 | Above the Law | From Startup to $2 Billion: EvenUp Is Transforming Personal Injury Practice | Standard concerns with AI tools remain: algorithmic bias, data privacy, and for EvenUp, whether its tools will further gum up the court system. |
| SI015 | EvenUp | EvenUp About Page — PLAAS and Company Overview | EvenUp is the only Pre-Litigation as a Service provider combining proactive AI with a team of US-based legal operations professionals. |
| SI016 | LawNext | EvenUp Launches Medical Management Tool to Address Treatment Gaps | EvenUp's Medical Management product uses proactive AI check-in agents to identify and address treatment gaps before they reduce case value. |
| SI017 | EvenUp | EvenUp Products Page — Claims Intelligence Platform | The Claims Intelligence Platform: demand letters, medchrons, Smart Workflows, Medical Bill Summary, Express Demands — all included in one per-case price. |
| SI018 | Bloomberg Law | AI Legal Tech EvenUp Raises $150 Million at $2 Billion Valuation | EvenUp raised $150 million at a $2 billion valuation in its Series E round, cementing its position as the leading AI platform for personal injury law firms. |
| SI019 | Bloomberg Law | EvenUp Raises $135 Million in Series D Funding | EvenUp raised $135 million in Series D funding at a valuation exceeding $1 billion, led by Bain Capital Ventures. |
| SI020 | Bessemer Venture Partners | Roads to $100M ARR — Legal AI | EvenUp represents the category-defining vertical AI platform in personal injury legal services, on a trajectory toward $100M ARR. |
| SI021 | SignalFire | EvenUp Portfolio Company Page | SignalFire portfolio: EvenUp — AI-powered legal platform for personal injury law firms. |
| SI022 | Reuters | AI Boom Fuels Fresh Wave of Legal Tech Investments | The legal AI investment wave shows no signs of slowing; EvenUp's $150M Series E is among the largest rounds in the sector. |
| SI023 | PR Newswire | EvenUp Introduces Express Demands — First in AI Documents Series | EvenUp introduces Express Demands, the first in its AI Documents series — a rapid-turnaround demand letter product. |
| SI024 | EvenUp | EvenUp Newsroom | EvenUp newsroom: series D, series E, PLAAS launch, medical management, AI Drafts suite — all milestones documented. |
| SI025 | EvenUp | EvenUp Careers — Team and Hiring | EvenUp is hiring legal operations professionals, engineers, and product managers across our US-based team. |
| SE001 | EvenUp | Piai™ — How EvenUp's AI System Works | Piai™ is built on hundreds of thousands of injury cases and millions of medical records, enhanced by in-house legal and medical expertise. |
| SE002 | EvenUp | Expert Demands — What's Included in EvenUp's Professionally Reviewed Demands | |
| SE003 | EvenUp | MedChrons™ — Understand Treatment, Damages, and Risks in One Interactive View | |
| SE004 | EvenUp | Integrations — Your Cases, Ready When You Are | Go live in less than a day. Matters flow in automatically from there, so casework starts immediately. |
| SE005 | EvenUp | Communication Agents™ — End Hold Times, Reclaim Your Time | Agents check in with clients via text or call to track appointments. 37% of the time, they surface problems your team would have missed. |
| SE006 | EvenUp | Case & Negotiation Prep — Optimize Every Step From Intake to Resolution | |
| SE007 | EvenUp | Settlement Repository™ — Trusted by the Top Personal Injury Law Firms | |
| SE008 | EvenUp | Executive Analytics™ — Turn Performance Data Into Action | |
| SE009 | EvenUp | Data Processing Agreement (DPA) — EvenUp Customer Data Processing Terms | EvenUp will notify Customer of the Security Incident without undue delay and in any event within seventy-two (72) hours after becoming aware of it. |
| SE010 | EvenUp (Haixun Wang, VP Engineering and Head of AI) | Building Trustworthy, Scalable Document AI for Legal Tech | Piai™ extracts and structures data from the most complex raw records rather than AI-extracted summaries. This includes poorly legible handwriting and images, capturing nuanced language with superior accuracy. |
| SE011 | EvenUp | How Accurate Are Legal AI Tools for PI Firms — EvenUp vs. General Purpose AI | Piai™ is built on a Reading Layer and a Writing Layer. The Reading Layer features the initial processing of unstructured, messy legal documents. |
| SE012 | EvenUp | EvenUp Announces AI Playbooks™ and Voice Agent™ to Redefine Case Analysis and Client Communication | AI Playbooks analyzes your case files and extracts the insights you need every time you upload new documents. |
| SE013 | EvenUp | EvenUp Law Celebrates Haixun Wang's Recognition as a 2024 ACM Fellow | Haixun was recognized for his foundational work on graph-based systems and their application to text understanding. |
| SE014 | EvenUp | Disrupting the World of Legal Tech: Fatemeh Torabi Asr Uses Machine Learning to Help EvenUp Customers | We're building the foundational models that power EvenUp's solutions, from classifying and tagging medical documents to innovating new LLM-based models to extract and consolidate information. |
| SE015 | EvenUp | Driving Innovation with Human-Centered AI: Meet EvenUp's Emre Yamangil | We moderate AI outputs to avoid errors that could impact clients' lives. |
| SE016 | EvenUp | EvenUp Careers — Meet Our Engineering Team | Our engineers build fast, solve hard problems, and ship technology that makes a real difference. |
| SE017 | EvenUp (GitHub) | EvenUp GitHub Organization — Public Repository Index | Popular repositories: Puppet modules for backup, DNS management, log shipping, and security auditing. |
| SE018 | EvenUp | Why EvenUp — The First and Best Proactive AI Platform for Personal Injury | |
| SE019 | EvenUp | Introducing AI Drafts™, Smart Workflows, Medical Bill Summary, and Case-Based Pricing | According to EvenUp's proprietary data, 42% of Demands are sent >100 days after last treatment. |
| SE020 | LawNext | EvenUp Extends Beyond Software with Launch of 'Pre-Litigation-as-a-Service' Offering For PI Law Firms | Whether PLAAS represents a genuinely new category or a repackaging of existing services with an AI wrapper will likely depend on how deeply firms integrate it and what the outcomes look like at scale. |
| SE021 | LawNext | EvenUp Launches Medical Management Tool to Address Treatment Gaps in Personal Injury Cases | 16.8% of plaintiffs develop a 30-day gap in treatment within the first three months of their case. That figure rises to 32.4% after six months. |
| SE022 | Above the Law | From Startup To $2 Billion: EvenUp Is Transforming Personal Injury Practice | The standard concerns with AI tools remain: algorithmic bias, data privacy, and for EvenUp, whether its tools will further gum up the court system. |
| SE023 | SmartAdvocate | SmartAdvocate Case Management System — Features and Integrations | |
| SE024 | Filevine | Filevine Features — Case Management and Integrations | |
| SE025 | American Bar Association | ABA Model Rule 1.1: Competence | A lawyer shall provide competent representation to a client. Competent representation requires the legal knowledge, skill, thoroughness and preparation reasonably necessary for the representation. |
| SE026 | Bloomberg Law | EvenUp Raises $150 Million for AI That Helps Personal Injury Firms | |
| SE027 | Association for Computing Machinery (ACM) | 2024 ACM Fellows Celebrated for Transformative Contributions to Computing Science and Technology | |
| SE028 | Fortune | EvenUp raises $150 million at over $2 billion valuation as AI reshapes personal injury law | |
| SU001 | EvenUp | Customer Success Stories | How ELG Injury Lawyers Went From 'Firefighting' to 400% Revenue Growth with EvenUp |
| SU002 | EvenUp | How ELG Injury Lawyers Went From 'Firefighting' to 400% Revenue Growth with EvenUp | Our business is up over 400% of what it was just a year ago. |
| SU003 | EvenUp | How John K. Zaid & Associates Used EvenUp's Proactive AI Agents to Catch Issues Early in 37% of Client Conversations | We've had attorneys come into my office saying, 'Sir, we just got policy limits—$30,000—on a case we never would have received before.' |
| SU004 | EvenUp | How Mama Justice Used EvenUp to Hit to 40% Higher Settlements and 14% Faster Cases Without Adding Headcount | We had a case where we went in asking for $1 million and got over $900,000. Without EvenUp, we could have easily settled it for around $600,000. |
| SU005 | EvenUp | How J&Y Law Shaved 25% Off Their Case Lifecycle with EvenUp | The partnership with EvenUp has been very impactful. The team is incredibly responsive, and they make our employees feel important. That's what made this change work. |
| SU006 | EvenUp | Personal Injury Pioneer Awards | Advisory Board Membership — Opportunity to join EvenUp's elite Customer Advisory Board, contributing to product innovations. |
| SU007 | EvenUp | Why Choose EvenUp | Trusted by thousands of PI firms for mission-critical work across pre-litigation and litigation |
| SU008 | LawNext | EvenUp Extends Beyond Software with Launch of 'Pre-Litigation-as-a-Service' Offering For PI Law Firms | Whether PLAAS represents a genuinely new category or a repackaging of existing services with an AI wrapper will likely depend on how deeply firms integrate it and what the outcomes look like at scale. |
| SU009 | LawNext | EvenUp Launches Medical Management Tool to Address Treatment Gaps in Personal Injury Cases | Clark Fielding, founding principal at Fielding Law in Irvine, Calif., said the technology helps his firm be better advocates. 'We can pull up information in real time during depositions and get the ammunition we need to be the best advocates possible.' |
| SU010 | LawNext | How It Works: My 15 Minute Demo with EvenUp | With over 1,000 law firms utilizing its services, EvenUp has helped collect over $500 million in damages |
| SU011 | Above the Law | From Startup To $2 Billion — EvenUp Is Transforming Personal Injury Practice | Of course, the standard concerns with AI tools remain: algorithmic bias, data privacy, and for EvenUp, whether its tools will further gum up the court system. |
| SU012 | Unite.ai | Raymond Mieszaniec, Co-Founder & COO of EvenUp — Interview Series | Firms like Sweet James have grown to over $500 million in annual results with 70% year-over-year growth, all without adding headcount. |
| SU013 | Lightspeed Venture Partners | EvenUp Portfolio Company Profile | a rapidly growing startup ($0 to $10M in ARR in <2 years) |
| SU014 | Supio | Why AI in Legal Needs a Higher Standard | The firms that will lead are not the ones that adopt AI fastest. They're the ones who adopt it most carefully, with systems that make verification easy, outputs that are auditable, and a clear understanding of where human judgment remains essential. |
| SU015 | EvenUp | How Accurate Are Legal AI Tools for PI Firms | EvenUp firms using our Missing Docs Check feature see a 75% reduction in missing documents per case and a 69% greater likelihood of reaching policy limits on every case. |
| SU016 | EvenUp | EvenUp Newsroom | |
| SU017 | EvenUp | Introducing AI Drafts, Smart Workflows, Medical Bill Summary, and Case-Based Pricing | 1,500+ PI firms |
| SU018 | Bloomberg Law | EvenUp Raises $135 Million in Series D Funding | more than 1,000 law firms |
| SU019 | Bloomberg Law | AI Legal Tech EvenUp Raises $150 Million at $2 Billion Valuation | |
| SU020 | PR Newswire | EvenUp Announces $150 Million Series E Funding Round | |
| SU021 | SignalFire | EvenUp Portfolio | |
| SU022 | LawSites (Bob Ambrogi) | EvenUp Launches Medical Management Tool | |
| SU023 | Clio | Legal Trends Report | |
| SU024 | EvenUp | EvenUp Products Overview | |
| SU025 | Bloomberg Law | Investors Pour Cash Into AI Startups for Plaintiffs Lawyers | |
| SR001 | U.S. Department of Health & Human Services (HHS) | Business Associates — HIPAA Privacy Rule Guidance | A business associate is a person or entity that performs certain functions or activities that involve the use or disclosure of protected health information on behalf of, or provides services to, a covered entity. |
| SR002 | U.S. Department of Health & Human Services (HHS) | Summary of the HIPAA Security Rule | |
| SR003 | U.S. Department of Health & Human Services (HHS) | HIPAA Breach Notification Rule | If a breach of unsecured protected health information occurs at or by a business associate, the business associate must notify the covered entity without unreasonable delay and no later than 60 days from the discovery of the breach. |
| SR004 | HHS Office for Civil Rights (OCR) | OCR HIPAA Breach Report Portal | |
| SR005 | National Institute of Standards and Technology (NIST) | Artificial Intelligence Risk Management Framework (AI RMF 1.0) | |
| SR006 | Federal Trade Commission (FTC) | Generative AI Raises Competition Concerns | If a single company or a handful of firms control one or several of these essential inputs, they may be able to leverage their control to dampen or distort competition in generative AI markets. |
| SR007 | Federal Trade Commission (FTC) | Office of Technology | |
| SR008 | Justia / U.S. District Court, Southern District of New York | Mata v. Avianca, Inc. — Opinion and Order on Sanctions (Document 54, S.D.N.Y. 2023) | OPINION AND ORDER ON SANCTIONS: The Court Orders the following sanctions pursuant to Rule 11, or, alternatively, its inherent authority. |
| SR009 | State Bar of Texas | Texas Ethics Resources | |
| SR010 | American Bar Association | Rule 1.1: Competence — Model Rules of Professional Conduct | |
| SR011 | American Bar Association | Rule 1.6: Confidentiality of Information — Model Rules of Professional Conduct | |
| SR012 | American Bar Association | Rule 5.5: Unauthorized Practice of Law — Model Rules of Professional Conduct | |
| SR013 | EvenUp | Data Processing Addendum | Customer Personal Data as used herein does not include, and this DPA does not apply to, any Protected Health Information (PHI) as defined under HIPAA, which, where required or applicable, will instead be governed by a separate Business Associate Agreement (BAA) entered into by the Parties. |
| SR014 | EvenUp | How Accurate Are Legal AI Tools for PI Firms | EvenUp's 2025 benchmarks report found that PI firms are missing an average of five documents and bills for every 10 cases. |
| SR015 | EvenUp | Driving Innovation with Human-Centered AI: Meet EvenUp's Emre Yamangil | We're not just about quick solutions. We moderate AI outputs to avoid errors that could impact clients' lives. |
| SR016 | EvenUp | Disrupting the World of Legal Tech: Fatemeh Torabi Asr | |
| SR017 | RELX Group | Investor Overview | |
| SR018 | RELX Group | Our Business Overview | |
| SR019 | Artificial Lawyer | Artificial Lawyer — Legal AI News | |
| SR020 | National Law Review | Legal News & Business Law News | |
| SR021 | HIPAA Journal | HIPAA Training Requirements — Updated for 2026 | |
| SR022 | HIPAA Journal | What is Protected Health Information? 2026 Update | |
| SR023 | ABA Journal | Generative AI — News by Topic | |
| SR024 | ABA Journal | ABA Journal Homepage | |
| SR025 | Supio | Why AI in Legal Needs a Higher Standard | Courts are grappling with AI in filings, and attorneys are actively being sanctioned for submitting AI-generated content they failed to verify. |
| SR026 | HIPAA Journal | Healthcare Data Breach Statistics — Updated for 2026 | The number of affected individuals soared by 58% to more than 289 million individuals in a single year, almost 85% of the population of the United States. |
| SR027 | LawSites (Robert Ambrogi) | LawSites — Law, Websites, Technology | |
| SR028 | Clio | Clio Legal Trends Report | |
| SR029 | Filevine | Filevine — Legal AI & Operating Intelligence | |
| SR030 | Artificial Lawyer / LawSites | Harvey Partners With DeepJudge and Unveils Command Center | |
| SR031 | LexisNexis | Lexis+ with Protégé — Legal AI Solution | |
| SV001 | EvenUp (official blog) | EvenUp Announces $135 Million Series D | Today, we're thrilled to announce $135 million in new funding, which brings our valuation to over $1 billion. |
| SV002 | GlobeNewswire | EvenUp Raises 135 Million at 1 Billion Valuation to Automate the Drafting of Demand Letters for Personal Injury Law Firms | EvenUp Raises 135 Million at 1 Billion Valuation |
| SV003 | Bloomberg Law | EvenUp Raises $150 Million for AI Legal Startup at $2 Billion Value | |
| SV004 | Bloomberg Law | EvenUp Raises Series D at $1B Valuation for Personal Injury AI | |
| SV005 | Above the Law | From Startup to $2 Billion: EvenUp Is Transforming Personal Injury Practice | EvenUp now has a $2 billion valuation. |
| SV006 | LawNext | EvenUp Extends Beyond Software with Launch of Pre-Litigation as a Service Offering for PI Law Firms | Early testing of the service has already resulted in sales of more than $10 million in PLAAS subscriptions, EvenUp says. |
| SV007 | LawNext | EvenUp Launches Medical Management Tool to Address Treatment Gaps in Personal Injury Cases | |
| SV008 | LawSitesBlog | EvenUp Launches Medical Management Tool | |
| SV009 | Grand View Research | Legal AI Market Size, Share & Trends – Industry Report, 2030 | The global legal AI market size was valued at USD 1.45 billion in 2024 and is projected to reach USD 3.90 billion by 2030, growing at a CAGR of 17.3% from 2025 to 2030. |
| SV010 | MarketsAndMarkets | Legal AI Software Market – Global Forecast to 2030 | The legal AI software market is expected to grow from USD 3.11 billion in 2025 to USD 10.82 billion by the year 2030, at a CAGR of 28.3% during the forecast period. |
| SV011 | IBISWorld | Personal Injury Lawyers & Attorneys in the US Market Size Statistics | The market size of the Personal Injury Lawyers & Attorneys in the US is $61.7bn in 2025. |
| SV012 | IBISWorld | Personal Injury Lawyers & Attorneys in the US Number of Businesses Statistics | |
| SV013 | Lightspeed Venture Partners | EvenUp – Lightspeed Portfolio | We are a rapidly growing startup ($0 to $10M in ARR in <2 years) |
| SV014 | Bloomberg Law | Harvey Raises $300 Million at $3 Billion Valuation for Legal AI | |
| SV015 | Harvey AI (blog) | Harvey AI Blog | |
| SV016 | Thomson Reuters (CoCounsel) | CoCounsel by Thomson Reuters | |
| SV017 | Legal IT Insider (legaltechnology.com) | EvenUp Raises 135m Series D Round | |
| SV018 | Legal IT Insider (legaltechnology.com) | EvenUp Raises 150 Million Series E | |
| SV019 | Clio | Legal Trends Report – Clio | |
| SV020 | US Securities and Exchange Commission | ServiceTitan, Inc. – DRS/A Registration Statement (EvenUp reference) | ServiceTitan, Inc. DRS/A – identifies EvenUp, Inc. as a Delaware-incorporated privately held software company with board representation beginning April 2023. |
| SV021 | US Securities and Exchange Commission (EDGAR full-text search) | EDGAR Full-Text Search: EvenUp, Inc. – Form D and other filings | EDGAR search returns four ServiceTitan DRS/A filings mentioning EvenUp, Inc.; no independent Form D for EvenUp identified. |
| SV022 | SignalFire | EvenUp – SignalFire Portfolio | |
| SV023 | Bessemer Venture Partners | Ten Principles for Building Strong Vertical AI Businesses | |
| SV024 | Supio AI (competitor blog) | Why AI in Legal Needs a Higher Standard | In the legal world, where facts and accuracy are everything, they [hallucinations] have the potential to destroy a case, ruin a firm's reputation, or cost an attorney their license. |
| SV025 | RELX | RELX Investors | |
| SV026 | Claims Journal | EvenUp Raises Series E at Over $2 Billion Valuation | |
| SV027 | Forbes | Forbes Cloud 100 2025 | |
| SV028 | LawNext | How It Works: EvenUp – 15-Minute Demo with Saam Mashhad | |
| SV029 | EvenUp (official) | EvenUp Newsroom | |
| SV030 | Hacker News (Y Combinator) | Hacker News Discussion – EvenUp Series D (October 2024) | |
| SV031 | Statista | Legal AI Market Size Worldwide – Statista | |
| SV032 | Artificial Lawyer | Artificial Lawyer – Legal AI Industry Coverage | |
| SV033 | Bloomberg Law | EvenUp Raises $150 Million for AI That Helps Personal Injury Firms | |
| SV034 | Darrow AI | Darrow – Legal Exposure Management | |
| SV035 | EvenUp (official about page) | About EvenUp – PLAAS and Platform Overview |