初创公司尽调
尽调报告 AI / IT Service Management (ITSM) Series B 2026-10-12

Serval

AI 原生 ITSM,搭配确定性工作流自动化和 $1B B 轮融资

Serval 兼具可信的 AI 原生 ITSM 架构和强劲的风险投资支持,但单位经济模型不透明,客户仍集中在技术先行企业;$1B 估值值得跟踪,但押注偏高。

覆盖范围与披露说明

公开证据支持产品架构、融资和早期客户进展,但仍未披露绝对 ARR、利润率、留存率、准确员工数,以及广泛的非科技企业采用情况。

封面要素

最近融资 01
75 USD M [CO007]
估值 02
1000 USD M [CO007]
收入增长 04
500 % [CO008]
员工数 06
sub-30 employees [CV013]

公司概况

Serval 是一家总部位于旧金山的 AI 原生 IT 服务管理公司,由 Jake Stauch 和 Alex McLeod 于 2024 年创立。公司将 Help Desk Agent 与 Catalyst 工作流构建器结合,用于自动化员工服务、权限和跨团队工作流;公司公开披露以 $1B 估值完成 $75M B 轮融资,累计融资 $127M。公开证据仍未揭示绝对 ARR、 准确员工数、客户数和单位经济性。

官网
www.serval.com
创始人
Jake Stauch, Alex McLeod
创立地点
San Francisco, CA
总部
San Francisco, CA
产品
AI 原生 ITSM 与工作流自动化平台,把自然语言请求编译成确定性的 TypeScript 自动化流程,覆盖帮助台、权限配置、资产管理和工作流编排模块。
客户
中端市场至企业级 IT、安全和运营团队
商业模式
定制化企业 SaaS 订阅,价格需通过演示获取,并配合前置部署实施支持
阶段
Series B
融资情况
已融资 $127M;B 轮 $75M,估值 $1B
[CO001, CO002, CO004, CO005, CO006, CO007, CO008, CO009, CO010, CO011, CO012, CO013, CO014, CO015, CO016]

执行摘要

主要优势

  • 确定性的双智能体工作流架构,支持 IT、访问权限和员工工作流的可审计自动化。
  • 强劲的投资者支持和已披露的 $127M 资本,为运营提供了充足的现金跑道。
  • 公开客户证据显示,知名科技客户的工单自动化提升很快。

主要风险

  • 绝对 ARR、毛利率和净留存率仍未披露,削弱了投资判断的把握。
  • 公开客户证据集中在风险投资支持的科技公司,而非广泛的企业垂直行业。
  • 现有 ITSM 厂商可以捆绑 AI 功能,并利用其分销和系统记录优势。

未决问题

  • 绝对 ARR 或 MRR,以支撑 500% 增长声明并校准估值倍数。
  • 各客户队列的毛利率、CAC 回本周期和净收入留存率。
  • 确切员工数和客户数,以及更广泛的非科技企业客户参考。
  • $1B Series B 价格对应的股权结构表、清算优先权和董事会 / 控制权条款。

目录

Chapter 01

01公司概览

1.1 身份、产品与商业模式

Serval 成立于 2024 年,总部位于旧金山。产品把帮助台、权限管理、资产管理和工作流自动化结合起来,服务企业内部运营。Serval 描述了双智能体模式: Catalyst 将自然语言指令转成代码工作流,Help Desk Agent 则针对员工请求运行已获批准的工作流。公开证据支持其企业软件模式,但看不到透明价目表: Multiples 将收入模式归类为订阅,Macha 则未找到公开套餐或起始价格,并称其采用演示驱动、定制合同的流程。因此,外部无法估算合同价值、 部署经济性或经常性收入。[CO001, CO002, CO003, CO014]

快照 KPI 表
指标数值或状态期间 / 日期置信度缺口或限定
成立时间2024公司成立高公司、新闻和市场数据来源相互印证
总部San Francisco, California当前公开资料高未披露完整的办公地点或法人实体所在地清单
阶段B 轮2025-12-11高指定证据中未出现更晚的定价轮次
估值$1 billionSeries B, 2025-12-11高融资估值,并非独立评估
累计融资公司累计 $127 million自 2025 年 8 月起,2025-12-11 报道中Multiples 列出两轮定价融资合计 $122 million;需核对 $5 million 差额
收入 / ARR未披露绝对值2026-10-12 证据审阅低公司报告的收入增长 500% 持续约三个月,并非 ARR
员工人数未披露绝对值2025-12-11低公司称员工人数增长超过三倍,但未给出起始或期末人数
客户未披露总数2026-10-12 证据审阅中已点名的生产环境用户和官网 Logo 不能证明完整的客户分母
自动化公司称为 50–80%;新闻报道六家点名客户超过 50%2025-10-21 至 2026中未披露客户层面的工单构成和测量方法
定价无公开档位或起始价格2026 年审阅中以演示为入口的定制合同,使公开 ACV 或单位经济分析无法进行

指标名称和期间沿用来源;未披露的数字保留为缺口,不作估算。

[CO001, CO002, CO006, CO007, CO008, CO009, CO010, CO011, CO012, CO013, CO014]

1.2 领导层、人才策略与治理缺口

公开确认的领导层仍集中在联合创始人 Jake Stauch(首席执行官)和 Alex McLeod(首席技术官)身上。Serval 还将前 Greylock 合伙人 Christine Kim 列为战略项目负责人及 Serval Start 负责人。这个总部在旧金山、为期两年的项目首期从 500 多名申请者中选出 12 人,旨在持续输送前置部署工程人才。 该项目是可信的招聘信号,但没有解决关键人物依赖或治理问题。现有证据未披露董事会成员、独立董事、创始人投票控制权、继任安排或投资人同意权, 因而无法据此判断董事会独立性或控制权分配。[CO004, CO005, CO016]

管理层与创始人表
人员职务有公开支持的背景职能覆盖依赖项 / 尽调问题
Jake Stauch联合创始人兼 CEOServal 联合创始人;指定证据未提供更完整的履历首席执行管理要求提供继任计划、决策权地图和时间分配
Alex McLeod联合创始人兼 CTOServal 联合创始人;指定证据未提供更完整的履历技术管理要求提供架构所有权地图和技术继任覆盖
Christine Kim战略项目负责人;Serval Start 负责人Greylock 前合伙人战略项目和前置部署人才项目澄清运营职责、汇报关系和该职位的长期性

覆盖信息不完整,因为资料池没有提供完整的高管名单、董事会名单或继任计划。

[CO004, CO005, CO016]

1.3 融资、经营信号与投资判断边界

Serval 的披露融资增长极快。SiliconANGLE 报道称,公司于 2025 年 10 月 21 日完成由 Redpoint 领投的 $47M A 轮,当时累计资本达到 $52M。 Serval 于 2025 年 12 月 11 日宣布由 Sequoia 领投的 $75M B 轮,估值 $1B,并称自 8 月以来累计融资 $127M。Multiples 记录两轮定价融资合计 $122M,因此 Serval 所说总额中多出的 $5M 需要核对,不能直接归入未披露轮次。Serval 还称,A 轮后约三个月内收入增长 500%、员工数增长逾三倍, 但没有给出绝对收入基数或当前员工数。客户证据同样只是方向性信号:公司材料称自动解决 50%–80% 的工单,独立报道则称六家已命名客户将超过 50% 的 IT 工单自动化。两者都不能证明客户总数、ARR、留存率、利润率或现金消耗。[CO006, CO007, CO008, CO009, CO010, CO011, CO012, CO013, CO015]

利益相关方或投资者地图
利益相关方已披露职能经济重要性控制权证据尽调要求
Sequoia CapitalB 轮领投领投 $75 million 轮次未披露董事席位和权利获取董事会、否决权、按比例跟投权和信息权条款
Redpoint VenturesA 轮领投;B 轮参投领投 $47 million 轮次未披露董事席位和权利确认持股、董事会代表权和保护性条款
General CatalystA 轮和 B 轮参投持续参与的机构投资者未披露控制权确认完全摊薄持股和按比例跟投权
First RoundA 轮和 B 轮参投持续参与的机构投资者未披露控制权确认持股和信息权
Meritech Capital PartnersB 轮参投成长期财团成员未披露控制权确认优先级和参与条款
EvanticB 轮参投财团成员;金额未披露未披露控制权确认投资金额和附函权利
Sound VenturesB 轮参投财团成员;金额未披露未披露控制权确认投资金额和附函权利
Radical VenturesB 轮参投财团成员;金额未披露未披露控制权确认投资金额和附函权利

投资者名单不等于股权结构表;持股比例、优先股堆叠、老股交易、债务和治理权利均不可得。

[CO006, CO007, CO008]
FO001: 公司里程碑时间线

时间线按指定证据记录成立、融资、规模、产品定位和人才里程碑。

日期和金额均复述源披露;时间线不推断未披露的种子轮或产品发布日期。

公司成立 [CO001, CO004]

$47M Series A 轮 [CO006]

累计融资 $52M [CO006]

客户自动化信号 [CO013]

$75M Series B 轮 [CO007]

公司累计融资 $127M [CO008, CO009]

公司披露的规模加速 [CO010, CO011]

首届 Serval Start 批次 [CO016]

[CO001, CO004, CO006, CO007, CO008, CO009, CO010, CO011, CO013, CO016]

1.4 展示项

Chapter 02

02市场分析

2.1 市场边界与替代方案

Serval 能够站住脚的核心市场是企业内部服务管理,而不是外部客户支持。其自身定位覆盖员工请求处理、入职与离职、即时权限、工单、资产管理和工作流编排。 因此,最接近的预算边界是 ITSM 软件、身份与权限运营、工作流自动化三者的交集;更广的邻接市场包括 HR、财务、法务和安全服务工作流。 现状替代方案则是人工队列和管理员在旧系统中搭建的工作流。Serval 还提供较低切换成本的切入方式:作为与现有 ITSM 同步的自动化层,客户可以暂缓更换系统记录源。 指定证据无法判断这些预算中哪些是新增预算,哪些只是重新分配。[CM001, CM002, CM003, CM014]

市场定义与支出边界
细分或类别纳入的支出或工作流排除的支出买方或付款方证据与 Serval 的相关性
内部 IT 服务管理员工请求、服务台、工单和请求解决外部客户支持运营IT 和运营团队;付款方未披露核心市场
身份与访问运营访问请求、配置、撤销配置和审批工作流与服务工作流无关的独立身份基础设施IT 和安全团队;预算负责人未披露核心相邻工作流
企业服务管理HR、财务、法务、安全和跨职能员工工作流面向客户的销售和支持自动化职能运营团队;付款方未披露扩展邻接市场
在位 ITSM 的自动化叠加层在现有系统之上进行双向同步和工作流执行立即替换所有在位记录系统现有 ITSM 运营方;付款方未披露更低切换成本的进入路径

定性边界来自 Serval 发布的产品描述和竞争对手发布的评测;公开信息未确定支出分配或预算归属。

[CM001, CM002, CM003, CM014]

2.2 受证据约束的市场测算

没有留存来源提供透明的市场规模方法、地理边界、基准年份、CAGR,或可用于 Serval 的数字化 TAM/SAM/SOM。公开结果证据有助于判断采用情况, 但不能换算成市场规模:Serval 称 Perplexity 的自动化率超过 50%,又单独表示客户通常能将帮助台工单的 50%–80% 自动化。其买方指南还引用 $15–$25 的人工坐席工单成本与 $2–$4 的 AI 解决工单成本,但指定资料池没有底层分析材料。这些指标说明存在潜在效率切口,不等于可触达收入。 因此,估值模型应使用客户具体的工单量、符合条件的工作流占比、定价和留存率,而不是套用泛化的 ITSM 市场估值。[CM004, CM005, CM006]

市场规模与采用视角
视角发布方期间地理范围数值方法置信度局限
数值化 TAM/SAM/SOM指定证据集截至 2026-10-12未保留透明的市场规模研究低无法从产品页面或评测推导市场收入
目标买方边界Macha2026 年审阅定性匹配评估:有融资的中端市场至企业级 IT 或运营中竞争对手发布的评估;无组织数量或支出分母
自动化采用率Serval2026 年买方指南服务台工单的 50%-80%Serval 客户范围内公司报告的区间低没有客户群定义、分母、样本量或独立审计
工单经济性Serval2026 年买方指南$15-$25 人工坐席;$2-$4 AI 解决公司页面引用分析师基准低底层基准来源不在指定资料池内,且不能证明 Serval 的收入

数值反映采用率或单位成本视角,并非 TAM/SAM/SOM;null 表示指定来源未披露可支撑的数值或地理范围。

[CM004, CM005, CM006, CM011]
FM001: 证据约束下的市场规模阶梯

证据将范围从宽泛的服务管理相邻市场收窄到可支撑的初始工作流切入口,但不赋予缺乏依据的市场价值。

各层宽度仅表示层级,不代表数值比例。

这是定性的范围阶梯,刻意不包含任何估算市场价值。

[CM001, CM002, CM004, CM011, CM014, CM015]

2.3 买方、用户、付款方与采用路径

最清晰的买方画像是资金充足的中端市场至企业级 IT 或运营团队;员工则通过聊天、邮件或网页请求渠道成为日常用户。随着工作流延伸到权限、HR、 财务和法务,安全及职能运营团队也会成为新增买方。公开证据没有确认合同付款方或预算负责人,而需演示获取的定价也使自下而上的席位或用量估算无法进行。 采用可以从重复、高频的权限和服务请求开始:AWS Marketplace 评论者称,配置好的权限工作流把完成时间从一两天缩短到几分钟。同一评论者也提到初始学习曲线, 说明工作流配置和运营归属仍是前提,并不会因 AI 而消失。[CM003, CM007, CM008, CM009, CM011, CM012]

细分市场与买方地图
细分市场买方用户付款方工作流预算负责人采用触发因素
IT 服务台IT 或运营管理层员工和服务台人员未披露请求接收和自动解决未披露高量重复工单
身份与访问IT 和安全团队员工和管理员未披露审批、配置和撤销配置未披露手动访问请求缓慢且有审计需求
跨职能 ESMHR、财务、法务或运营团队员工和职能运营人员未披露入职和跨部门工作流未披露系统与部门之间反复交接
在位 ITSM 叠加层现有 ITSM 负责人员工和服务台人员未披露与当前记录系统同步的自动化未披露希望自动化但不立即迁移

买方和用户描述是定性的;指定公开证据未披露签约付款方、预算负责人或采购路径。

[CM002, CM003, CM007, CM008, CM009, CM010, CM011, CM012, CM013]

2.4 增长驱动、约束与估值意义

需求来自工单跟踪与完成解决之间的差距、重复性权限和员工服务工作的广度,以及在替换现有 ITSM 之前先部署自动化的选择。Serval 的确定性工作流模型、 明确的发布控制、受限 API、审批政策和多种部署模式,能够回应企业采用常见的信任与数据驻留顾虑。约束同样明显:买方教育和配置工作真实存在, 定价不透明,没有专职工作流负责人的小团队可能并不适合,现有 ROI 基准也主要来自公司或单个客户,而非独立市场证据。只有在尽调确认合同价值、 实施负担、续约行为,以及自动化收益能否在普通客户而非精选案例中持续之后,估值才应奖励可验证的扩张。[CM002, CM006, CM007, CM009, CM010, CM012, CM013, CM015]

2.5 展示项

Chapter 03

03竞争对手

3.1 竞争格局与架构分层

企业 IT 服务管理市场已经分成两类:在既有系统记录源上加装人工智能的成熟平台,以及围绕智能体自动化原生设计的新兴平台。ServiceNow、Jira Service Management 等现有厂商主导核心企业服务目录,但主要依赖规则路由、管理脚本,以及收购或叠加到传统关系型架构上的对话式分流机器人。相比之下, Atomicwork、Aisera、Serval 等 AI 原生挑战者把服务台重点放在自动执行动作,而不是人工跟踪工单。第三类是 eesel AI 这样的轻量对话层, 可直接接入现有帮助台而不替换底层系统记录源。Serval 试图跨越这些分层:既支持双向同步下的渐进式共存,也支持替换 IT、HR、安全和财务部门的传统 ITSM 基础设施。[CP001, CP002, CP004, CP006]

竞争对手概览表
竞争对手类别规模 / 融资目标细分市场核心差异化核心局限
ServalAI 原生 ITSM已融资 $127M(估值 $1B)科技与中端市场企业 IT双代理架构生成确定性 TypeScript 代码;双向同步定价不透明且需演示;要求前置部署工程试点
AtomicworkAI 原生 ITSM风险投资支持($14M+)替换 / 增强传统 ITSM 的企业面向 IT/ESM 的基于角色 AI Coworkers;公开定价档位和结果积分相比传统在位者,合作伙伴生态和品牌覆盖较小
Moveworks (ServiceNow)被收购的对话式平台被 ServiceNow 收购(企业级规模)全球 2000 强企业员工深度对话式 LLM 分流原生集成于 ServiceNow 工作流后端紧密绑定 ServiceNow 平台;是对话分流而非独立 ESM
AiseraAI 服务台风险投资融资 $100M+中端市场至大型企业服务台GenIQ 多代理工作流声称自动解决率 75%黑箱式专有自动化逻辑;工作流代码可见性有限
eesel AI集成优先的 AI 层种子轮 / 风险投资融资使用现有服务台的支持与 IT 团队轻量自助配置;基于历史工单模拟;按积分计价缺少原生工单、访问治理和资产管理记录系统
ServiceNow在位记录系统上市公司(市值 $160B+)全球企业 IT、HR 和运营全面的企业工作流套件、庞大合作伙伴生态、Moveworks 集成总拥有成本高、管理员脚本复杂、部署周期缓慢
Jira Service Management在位 IT 服务台上市公司(Atlassian)DevOps、工程和 IT 团队原生 Jira 问题跟踪集成、前期成本较低、开发者采用广泛基于规则的自动化脆弱;是附加 AI 功能而非原生确定性代码执行

竞争对手分类、融资阶段和差异化特征综合自厂商文档、比较分析和分析师资料,截至 2026 年。

[CP001, CP002, CP004, CP005, CP006, CP007]

3.2 竞争对手画像与相对定位

对竞争格局的评估显示,自动化执行深度与企业平台广度之间存在明显取舍。ServiceNow 等成熟平台在全球 2000 强企业中拥有无可匹敌的平台广度和合规认证, 最近又通过收购 Moveworks,把多语言 LLM 工单分流直接嵌入生态,增强了对话能力。但传统平台总拥有成本高,专业服务周期复杂。直接竞争者 Atomicwork 以角色化 AI 协作者替代企业服务管理,并提供公开的结果导向定价,给评估提供透明锚点。eesel AI 等点解决方案通过快速自助式上线和历史工单模拟提高采用速度, 却缺少全面企业服务管理所需的原生权限治理、资产跟踪和工单系统记录源。Serval 处于一个独特的高深度象限:把自然语言指令编译为确定性的 TypeScript 代码, 同时横向扩展至企业运营部门。[CP002, CP004, CP005, CP006, CP007, CP015]

FP001: 竞争定位图

按自动化执行深度和企业平台广度,展示 ITSM 与自动化提供商的战略定位。

条目X (2–8.5)Y (2–9.5)说明
1. Serval 8.5 8
2. Atomicwork 7.5 7
3. Moveworks(ServiceNow) 5.5 8.5
4. Aisera 6 5.5
5. eesel AI 4 3.5
6. ServiceNow(传统核心) 2.5 9.5
7. Jira 服务管理 3 7.5
8. 内部脚本 / 现状 2 2

定位基于有记录的执行架构进行定性序数评分(x 轴:对话式拦截与确定性代码生成)以及平台范围(y 轴:点状聊天层与完整 ESM 系统记录源)。

[CP002, CP008, CP015]

3.3 功能与架构能力对比

企业评估 AI 原生 ITSM,核心看架构防御力、代码透明度和治理边界。传统分流机器人使用概率语言模型猜测回复或推荐知识文章,而 Serval 采用双智能体架构: Catalyst Agent 把自然语言管理员简报转成可检查、可测试的 TypeScript 工作流,Help Desk Agent 只执行已批准、已发布的例程。这种分离形成架构隔离, 确保终端用户不能提示系统执行未经审查的代码或访问未授权 API。此外,Serval 原生提供即时权限配置和身份治理,把多日的 SaaS 权限申请压缩成一次点击的审批。 与 Aisera、Moveworks 这类封闭黑箱竞争者不同,Serval 直接暴露工作流代码,支持 Git 版本控制、代码审查和基于分支的变更管理,直面企业安全和可审计性要求。[CP003, CP008, CP009, CP010, CP011]

功能与能力矩阵
采购标准ServalAtomicworkMoveworks / ServiceNoweesel AI传统 ITSM(ServiceNow/JSM)
核心记录系统完整原生套件(工单、JIT 访问、资产)带 ITSM 连接器的完整原生套件ServiceNow 在位记录系统否(需要外部服务台)全面的记录系统
工作流生成自然语言生成确定性 TypeScript 代码基于角色的 AI Coworker 自动化生成对话式编排与目录工作流带选择性自动化规则的提示词编辑器手动配置与管理员脚本
代码可见性与版本控制完整的 Git 托管 TypeScript 代码透明度工作流定义部分开放专有黑盒自动化专有提示词与连接器逻辑平台内可编辑脚本,但受厂商锁定
访问管理与 JIT 配置原生 JIT 配置、SCIM、多审批人链路身份与访问管理连接器ServiceNow Identity Governance(IGA)附加组件API 驱动的 SaaS 访问申请独立 IAM / IGA 模块或人工工作流
共存与同步能力与 ServiceNow、Jira、Zendesk 双向同步连接主流企业服务台原生 ServiceNow 平台环境一键集成 Zendesk、Freshdesk、JSM自定义 API 连接器或第三方中间件
部署架构云 SaaS、混合 worker 或自托管 Kubernetes多租户云 SaaS企业云 SaaS(可提供 FedRAMP)多租户云 SaaS多租户云 SaaS 或自托管

架构和功能标准根据已发布的技术规格、产品文档和厂商比较指南评估。

[CP003, CP008, CP009, CP010, CP011]

3.4 定价模式、产品打包与市场进入摩擦

在竞争领域,市场进入方式和合同透明度构成明显的商业分野。Serval 采用不透明、需演示获取的定价策略,没有公布价位层级、席位费用或用量费率, 而是按组织员工数和自动化范围协商定制订阅合同。每次部署都从专属前置部署工程师主导的四阶段试点开始,并附带至少自动化 50% 新增 IT 工单的明确合同保证。 这种高触达方式能从资金充足的科技公司中最大化提取支付意愿,但相较于 Atomicwork(公开 Professional 计划,年费 $25,000 起并收取结果费用) 和 eesel AI(500 个工单额度、月费 $299 起的透明订阅),也带来明显摩擦。[CP011, CP012, CP013, CP014]

定价与套餐对比
厂商定价 / 合同模式起始价格 / 计价指标试点 / 实施方式定价透明度尽调含义
Serval需定制演示后报价的订阅未披露的定制合同(按员工数与规模)由专属部署工程师参与的四阶段试点不透明(无公开价格档位或最低价)定制定价能最大化提取客户支付意愿,但会造成较高销售摩擦
Atomicwork公开订阅费 + 成果费Professional 版本起价 $25,000/年(25K credits)+ $1-$3/项成果厂商主导实施,并提供演示试用透明(公开基础费率与成果费率)提供清晰预算锚点;在竞争性比价中可能压低定制企业报价
eesel AI按 credits 计费的用量订阅500 credits 起价 $299/月($0.60/credit)自助配置,并模拟历史工单高(公开月度档位)低摩擦入口能吸引不愿承诺数月试点的团队
Moveworks (ServiceNow)企业年度合同 / 平台附加组件未披露的企业平台许可费企业专业服务部署不透明(捆绑在 ServiceNow 协议中)高合同最低额和多年承诺会锁定现有企业客户
传统 ITSM(ServiceNow/JSM)按代理人 / 用户分档授权JSM $20-$40/代理人/月;ServiceNow 企业档位内部管理员配置或 SI 实施分档目录价,折扣可协商成熟的按席位预算会抵制替换,除非 AI 自动化带来已验证的 ROI

根据厂商公开定价页和竞争性买方指南,汇总定价结构、入门单位经济模型和导入方式。

[CP011, CP012, CP013]

3.5 护城河持久性、切换成本与现有厂商反应

Serval 的长期防御力依赖确定性工作流仓库和双向集成层产生的切换成本。IT 组织一旦在 Git 中构建并进行版本控制,把跨部门自动化流程固化下来, 迁移到竞争平台就需要重新设计这些运营例程。与此同时,与 ServiceNow、Jira 和 Zendesk 的双向同步降低了初始采用门槛,允许客户在完全切换前无扰共存。 但竞争性替代风险依然很高。ServiceNow、BMC Helix 和 Atlassian 等平台厂商正把原生 AI 智能体能力直接嵌入现有企业协议,利用深厚客户关系和已经沉没的管理投入。 如果传统套件能在不要求完整数据库迁移的情况下提供可信的自主解决能力,Serval 定制定价、依赖工程师的销售模式可能面临日益加大的获客摩擦和利润率压缩。[CP005, CP007, CP011, CP014, CP015]

3.6 展示项

Chapter 04

04财务

4.1 收入来源、变现架构与定价不透明

Serval 的商业模式围绕企业软件即服务(SaaS)订阅平台展开,先从核心 IT 服务管理(ITSM)工作流切入,再横向扩张。不同于按推理 token、API 调用或工单解决 收取浮动费用的消费型 AI 基础设施厂商,Serval 以可预测的按席位订阅层级打包平台,不另收执行费。平台主要通过员工帮助台自动化、自动权限配置和设备生命周期编排变现, 企业扩张则来自向人力资源、财务、法务和安全团队的跨部门采用。但 Serval 在公开营销渠道上严格保持定价不透明,没有披露标价、最低起订席位或标准层级费用。 公司采用需演示获取的销售漏斗,按企业规模、席位数量和组织支付意愿双边协商合同。[CI003, CI005, CI007]

收入来源表
收入来源变现机制计价单位当前商业化状态收入质量与可预测性投资判断所需尽调
核心企业 ITSM 平台按席位计费的年度或多年 SaaS 合同,覆盖服务台与工单自动化每名员工每月一个席位(按年计费)已在企业及成长型客户中进行生产部署高:订阅收入可重复,年度合同锁定且流失风险低核实平均合同金额(ACV)、席位渗透率和年度合同续约率
部门级 ESM 扩张通过附加模块把工作流代理扩展至 HR、财务和法务按部门或扩展后的席位档位已商业化;已在多个跨职能企业部署中运行高:净扩张驱动因素,推动客户自然扩张并提升 NRR要求提供部门扩张节奏、净收入留存率(NRR)和扩张 CAC
前置部署工程支持将咨询式实施打包进企业试点承诺或企业档位配置专属部署工程师面向大型企业客户的活跃导入机制中:高接触服务投入补贴 SaaS 获客,前期会稀释利润率评估部署工程工时对毛利率的影响及专业服务利润率
Marketplace 与合作伙伴分销通过云 marketplace(AWS Marketplace)和生态合作伙伴(Jamf)间接采购云 marketplace 合并计费已在 AWS Marketplace 上架,并完成企业集成验证高:按企业云承诺标准抵扣 marketplace 消费检查 marketplace 渠道折扣利润率和净收入实现率

收入来源反映 Serval 已商业化的 SaaS 模式和已报道的企业 go-to-market 结构;合同层面的定价拆分及确切的部门费率尚未公开披露。

[CI003, CI007]

4.2 市场进入、企业签约与销售效率动态

Serval 的企业市场进入模式面向首席信息官和 IT 基础设施负责人,采用高触达的前置部署工程模式。为了降低 ServiceNow 等现有厂商带来的部署风险, Serval 为潜在客户配备专属部署工程团队,在试点期间配置定制集成并构建首批 Catalyst 工作流。这项前置服务投入帮助 Serval 在 Spotify、Fox、Live Nation、 GitLab 等头部企业账户,以及 Perplexity、Mercor、Together AI 等高速增长科技独角兽中建立据点。销售策略按部门先落地、再扩展,逐步抬高合同价值: IT 服务台自动化常规工单后,Catalyst 工作流引擎再用于自动化 HR 入职、财务采购审批和法务 NDA 流转。该顾问式模式能锁定更高平均合同价值和高切换成本, 但也带来前端利润摩擦;相比自助平台,采购犹豫还会拉长销售周期。[CI004, CI005, CI006, CI012]

定价 / 变现表
商业档位 / 产品定价与合同单位目录价状态实际定价与前置部署动态折扣、套餐与未知项证据来源与透明度
企业自助服务台每名员工每月一个席位未公开;需销售演示后报价根据公司员工数和自动化范围定制定价企业规模采购适用批量折扣;最低席位数未说明第三方评测(Macha)确认需演示后报价,且无公开档位
Catalyst 工作流构建代理捆绑平台能力或管理员档位未公开;包含在企业平台许可中支持用自然语言生成不限量工作流,并确定性执行并发构建席位或 API 连接器是否需要高级附加费尚不清楚公司公告确认没有隐藏的 API 或执行 credits 附加费
跨部门 ESM 附加模块按部门或席位扩展档位未公开;按扩展的职能团队定价支持 HR 入职、财务审批和法务合同流转多部门捆绑企业协议可协商折扣公司博客和高管访谈确认已实现跨部门部署
专属前置部署工程捆绑式实施试点项目无公开费率表,视为企业销售投入在最初 30 至 90 天部署阶段,工程师驻场或嵌入客户团队持续性高级客户成功服务是否另收年度费用尚不清楚第三方尽调(Macha)识别出前置部署工程试点结构

Serval 未公开透明的定价档位或席位费率;数据综合自公司产品文档和第三方分析师评测资料。

[CI003, CI005, CI006, CI012]

4.3 运营成本结构、计算费用与毛利率机制

Serval 的销货成本(COGS)和服务交付经济性主要由三类支出决定:多云基础设施托管、基础大语言模型推理计算,以及第三方 API 集成维护。运营工单的行业基准显示,自主 AI 解决请求的成本为每单 $2–$4,而人工实时坐席支持为 $15–$25,能为客户 IT 预算带来明显 ROI。但对 Serval 自身而言, 利润率转化高度依赖缓存效率和 token 经济性。Catalyst 将自然语言指令编译为确定性、可复用的 TypeScript 代码,并对企业 API 执行, 因而无需每个重复支持工单都重新运行昂贵的非确定性 LLM 推理。代码生成与确定性工具执行的架构分离限制了边际计算消耗,使公司能在合同规模相对固定集成开销扩张时, 争取软件毛利率。[CI008, CI009, CI010]

FI001: 收入转化与交付毛利流程

展示客户请求接入、工作流执行、按席位 SaaS 变现和软件毛利率实现的流程推进。

节点 / 连接节点 / 起点终点说明
节点 1 1. 服务请求接入员工通过 Slack、Teams、电子邮件或门户提交服务台请求
节点 2 2. 确定性执行Catalyst 生成的 TypeScript 工作流在 RBAC 和 API 上限约束下执行操作
节点 3 3. 按席位 SaaS 订阅每个员工席位收取可预测的核心平台费,不加收执行附加费
节点 4 4. 部门级 ESM 拓展自然扩展超出 IT,进入 HR、财务、法律和安全工作流
节点 5 5. 前置部署工程高接触度客户工程团队构建定制集成并加速部署
节点 6 6. 软件毛利率实现扣除云基础设施、LLM 模型推理和合作伙伴 API 成本后的毛利

流程展示将客户服务需求转化为实际 SaaS 收入和毛利率的架构阶段;绝对美元单位毛利率仍未公开。

[CI003, CI007, CI009, CI010]

4.4 资本充足性、现金消耗与投资判断结论

从资产负债表和融资角度看,Serval 近期资本充足性极强。公司在 2025 年 10 月由 Redpoint 领投完成 $47M A 轮后,于 2025 年 12 月由 Sequoia Capital 领投完成一笔抢先完成的 $75M B 轮,投后估值 $1B,累计股权融资达到 $127M。按报道中不足 100 人的运营员工数,Serval 的流动现金储备在标准企业烧钱轨迹下 可提供多年 runway。然而从机构投资判断角度看,这一估值倍数是未经验证的信念跃迁。Serval 在 2025 年末报告收入增长 500%,但没有 ARR 基线、净收入留存(NRR) 队列或经验证的 GAAP 毛利率作为锚点。在缺少经审计财务报表的情况下,对 $1B 估值的投资判断只能把 Serval 视为高风险的创投资产,而不是可据以承保的经营企业。[CI001, CI002, CI011, CI013, CI014]

单位经济模型表
单位经济维度已披露指标值估算置信度投资判断与尽调相关性关键尽调核实事项
年度经常性收入(ARR)低评估收入质量和估值倍数所需的基础基准要求提供经审计的过去 12 个月 ARR、GAAP 收入拆分和 MRR 增长瀑布
毛利率 %低在扣除 LLM 推理和云托管成本后,评估长期软件盈利能力查看毛利率明细,拆出云托管和第三方模型推理费用
CAC 回收期低衡量销售效率及前置部署工程试点的回收周期按客户群分析获客成本,包括完全计入的工程试点成本
净收入留存率(NRR)低验证跨部门扩张(HR/财务/法务)是否提升净合同金额检查历史客户群留存瀑布和年度扩张率
平均工单解决成本每张自动化工单 $2–$4,对比人工坐席 $15–$25中展示自动化的经济价值主张及客户为自动化付费的意愿获取客户案例遥测数据,验证实际工单分流节省的成本

标为 null 的值代表未在公开申报中披露的私有财务指标;工单经济模型基准来自公开行业买方指南数据。

[CI002, CI008, CI013, CI014]

4.5 展示项

Chapter 05

05产品与技术

5.1 产品定义与工作流自动化引擎

Serval 将自身定位为原生面向人工智能打造的企业服务管理平台,用自主执行请求取代传统工单跟踪队列。它不把用户分流到静态知识库文档,而是接收 Slack、 Microsoft Teams、邮件和网页门户中的员工咨询,直接自主执行后端工作流解决问题。核心用例包括软件权限自动配置、员工入职与离职生命周期管理、密码和 MFA 重置, 以及 HR、法务、财务和安全部门的跨部门管理工作。服务请求与原生权限控制及配置数据结合后,客户部署报告称无需技术人员介入即可解决超过 50% 的新增工单, 将平均解决时间从数小时或数天压缩至几分钟。[CE001, CE004, CE005, CE010]

工作流 / 用例表
用户任务当前工作流公司方案可量化收益局限
软件访问配置在多个控制台中手动提交工单,经理审批要数天采用 JIT 的限时访问工作流,自动通过 SCIM/API 撤销权限几分钟内解决;自动撤权减少长期保留的权限需要深入写入 IdP 和已连接 SaaS 控制台的 API 权限
人事入职与离职HR、IT 和安全团队各自维护割裂的人工清单通过 TypeScript 工作流编排跨系统、跨部门流程消除遗忘撤权任务,压缩处理周期跨部门边界情况仍需人工升级处理
密码与 MFA 重置重复的人工验证,以及服务台技术员工单队列通过 Slack、Teams 或网页门户自动验证并重置技术员零介入;据报道工单自主解决率超过 50%必须严格验证,以降低社会工程风险
资产与许可证治理定期人工审计 SaaS 席位和硬件电子表格定期运行主动代理,扫描 MDM 和 SaaS 日志中的闲置席位无需管理员轮询即可持续回收闲置许可证高接触度的企业旧工具缺乏现代 API 遥测

工作流对比、可量化收益和运营局限基于客户部署案例、用户评价和平台文档。

[CE001, CE004, CE005, CE010]

5.2 双智能体系统架构与确定性执行

Serval 的技术核心是智能体架构,在两个不同 AI 系统之间严格分权:面向员工的 Help Desk Agent,以及面向管理员的 Catalyst Agent。Help Desk Agent 对自然语言请求进行推理并获取组织上下文,但严格限制为通过架构隔离调用预先批准、已经发布的工作流,防止运行时临时编造提示。工作流由面向管理员的 Catalyst 创建:它把自然语言指令转换为可检查、确定性的 TypeScript 代码,而不是脆弱的 LLM 提示链。工作流直接同步至 GitHub 进行企业版本控制,也可通过 Serval CLI 管理,落实基础设施即代码纪律,并设置代码级审批门槛、基于角色的访问控制和严格的 API 范围上限。[CE001, CE002, CE003, CE008, CE014]

技术 / 运营架构表
层 / 组件运营角色系统依赖技术 / 治理风险
交互层通过 Slack、Teams、邮件、语音和网页表单接收请求第三方聊天平台 API 与 webhook 稳定性Chat API 速率限制与对话歧义
推理引擎对意图分类、解析上下文,并匹配预构建工作流基础 LLM 推理与上下文窗口延迟模型幻觉,或误判含糊的用户意图
执行运行时执行确定性的 TypeScript 自动化,并完整记录审计日志Node.js worker 基础设施(云、混合或自托管)运行时执行超时或第三方连接器 API 失败
权威系统同步双向同步工单、状态和 CMDB 对象ServiceNow、Jira Service Management 和 Freshservice REST API双向同步竞态条件及 ITSM 之间的模式漂移

架构组件、运营角色、依赖和风险因素来自技术文档、架构规范和安全披露。

[CE001, CE002, CE007, CE008]
FE001: 产品架构与安全治理堆栈

六层技术架构展示 Serval 的双代理分权、确定性 TypeScript 执行运行时、集成 CMDB 上下文存储和多云部署模式。

第 1 层:全渠道接入与界面层
接收来自 Slack、Microsoft Teams、电子邮件、电话和嵌入式 Web 门户的员工请求及对话上下文。
模块
  • Slack 与 Teams 机器人端点
  • 电子邮件与语音解析网关
  • ServiceNow 嵌入式门户组件
第 2 层:代理推理与意图分流层
评估请求、检索相关知识库文章并分派工单,不在生产环境临时编写代码。
模块
  • 服务台代理运行时
  • 知识检索引擎(Notion/Confluence)
  • 上下文与对话记忆存储
第 3 层:确定性工作流与自动化层
运行 Catalyst 构建的预编译、版本控制 TypeScript 自动化,并严格记录输入 / 输出日志。
模块
  • Catalyst 工作流生成器
  • TypeScript 执行引擎
  • GitHub 同步与 Serval CLI 管理器
第 4 层:访问治理与安全隔离层
在编写与执行之间实施严格的架构隔离,并由 RBAC、API 范围上限和多步骤审批提供支撑。
模块
  • 架构隔离边界
  • API 范围上限执行器
  • JIT 配置与多步骤审批
第 5 层:Serval 数据库与 CMDB 上下文图层
维护硬件、软件许可证、SaaS 订阅和配置项的对账图谱,并设置字段级优先级规则。
模块
  • 多源接入管道
  • 字段优先级对账引擎
  • 主动式诊断代理
第 6 层:部署与企业集成基础设施
连接身份提供商、MDM 和传统 ITSM,覆盖云托管、混合本地工作器和自托管环境。
模块
  • IdP 与 MDM 连接器(Okta、Kandji、Google)
  • 双向 ITSM 同步(ServiceNow、Jira)
  • 云 / 混合 / 自托管基础设施

架构层次与模块边界来自已发布的技术规格、专利披露和安全架构文档。

[CE001, CE002, CE003, CE008, CE014]

5.3 平台模块、资产图谱与治理控制

除了对话式工单,Serval 还把权限管理和资产配置数据库直接整合进核心平台。Serval Databases 充当企业配置管理数据库(CMDB)和上下文图谱, 持续从移动设备管理(MDM)、身份提供商(IdP)和 HRIS 系统接收硬件资产、软件许可证、SaaS 订阅及员工记录。字段级优先级引擎协调冲突数据源, 维护权威的单一事实来源。借助上下文图谱,Serval 部署定时 Proactive Agents,监控工单历史和资产遥测数据,识别安全漏洞、收回闲置软件许可证, 并为管理员审查准备拟议的工作流自动化。权限治理原生嵌入,通过有时间边界的即时(JIT)配置、自动撤销权限和定期访问审查,面向 SOC 2 Type II 和 ISO 27001 合规。[CE002, CE004, CE005, CE006, CE009]

产品模块 / 资产矩阵
模块 / 产品线主要用户状态 / 成熟度技术差异化尽调缺口
服务台代理终端员工与 IT 团队生产环境(GA)全渠道 NLP 分流与确定性工具调用自主解决边界与分流率
Catalyst 工作流构建器IT 与运营管理员生产环境(GA)自然语言生成 TypeScript 代码并与 Git 同步多系统分支逻辑中的提示词到代码准确率
Serval 数据库(CMDB)IT 资产与安全管理者生产环境(GA)多源资产图谱,按字段优先级进行协调大型企业 IdP 间的协调冲突处理
主动代理企业 IT 运营生产环境(GA)定期运行的后台代理,用于回收许可证和检测问题误报率与自动修复安全性

模块状态、差异化和尽调缺口综合自官方文档、CEO 访谈披露和独立 IT 分析师评估。

[CE001, CE002, CE004, CE006]

5.4 部署模式、差异化与技术风险

Serval 根据企业安全态势提供灵活部署选项,包括 99.9% 正常运行时间 SLA 的多租户云服务、在客户网络内执行 worker 的混合模式,以及完全自托管的单租户 Kubernetes 部署。为推动企业采用而不强迫其进行高风险迁移,Serval 支持与 ServiceNow、Jira Service Management 和 Freshservice 等现有 ITSM 双向同步。 但尽调暴露出若干技术和运营风险。首先,客户遥测显示多系统身份集成存在陡峭的运营学习曲线和复杂设置。其次,数十年的传统业务逻辑、自定义表和既有供应商集成, 都使替换根深蒂固的现有厂商更加困难。最后,完全没有公开开发者门户、开源 SDK 或公开开发者社区指标,使平台扩展性依赖供应商私下披露。[CE007, CE008, CE011, CE012, CE013]

5.5 展示项

Chapter 06

06客户

6.1 客户分层表

Serval 的商业进展集中在工程驱动的组织和高速成长的科技扩张企业,这些客户需要超越传统规则引擎的复杂自动化。目标买方主要是希望消除重复员工管理工作的 IT 负责人、首席信息安全官和 IT 运营总监。已披露的生产账户高度集中于前沿人工智能、开发者工具和现代金融科技公司,这些企业的内部 IT 技术栈已经使用 Okta、 JumpCloud、Slack 和 Google Workspace 等现代身份提供商。虽然这种集中带来极强的案例可引用性,并在同类风险投资企业中快速扩散,但 Serval 在传统、 强监管或非技术型企业环境中的渗透仍处于早期;一旦宏观经济逆风压缩风险软件预算,就会形成垂直行业集中风险。[CU001, CU006, CU011, CU013]

客户分群表
分群目标买方与付费方核心用例已观察到的规模与采用情况对 Serval 的战略价值尽调缺口
AI 与前沿科技成长型公司IT 副总裁 / 工程负责人JIT 访问配置、开发者工具访问、快速入职核心客户群(Perplexity、Together AI、Mercor)品牌背书和可供引用的客户案例价值高潜在 AI 市场整合下的预算韧性
高增长科技与金融科技IT 负责人 / IT 运营总监人事生命周期管理、SaaS 应用撤权、Slack 服务台生产部署(Ramp、SeatGeek、Notion)席位数高,具备跨部门扩张潜力合同规模和毛留存数据未披露
企业与中型市场综合行业CIO / 企业架构副总裁横向覆盖 HR/财务/IT 的 ESM,替换传统 ITSM早期试点;为《财富》百强制造商导入年度合同金额大($100k+),平台黏性强企业销售周期慢,传统系统定制障碍重
DevOps 与平台工程团队平台副总裁 / 基础设施负责人基础设施访问审查、Slack 事件分流、CI/CD 权限工程驱动型 IT 团队正在积极采用借助 TypeScript 工作流代码形成深度技术黏性除 chatops 外缺乏正式 SOC/NOC 集成

客户分群综合自已验证的客户披露、AWS Marketplace 评价和竞争对手分析。

[CU001, CU006, CU013]
FU001: 客户旅程图

从前置部署工程试点到横向扩展至多个部门的五阶段采用生命周期。

生命周期阶段与扩展触发因素综合自 Serval 试点合同、AWS Marketplace 评价和 Sacra 分析。

[CU007, CU009, CU010, CU011, CU012]

6.2 采用轨迹与自主解决指标

企业采用速度由快速实现价值支撑;Serval 声称,新上线客户群通常在部署四周内达到 50% 或更高的自主工单解决率。已有客户结果也证明工作量显著下降: Perplexity 将员工入职与离职中的超过一半 IT 请求自动化,SeatGeek 则在三周内从 0% 增至 40%,并在 60 天内超过 50%。权限治理上,Together AI 对即时权限请求实现了 95% 自动化,说明权限配置是 Serval 初期杠杆最高的用例。不过,由于绝对基线工单量、人工升级量和解决复杂度层级仍未公开, 对长期分流持久性的投资判断仍需私下验证。[CU002, CU003, CU004, CU005]

客户增长 / 采用轨迹表
指标已披露值日期来源归属置信度运营含义缺失分母
达到 50%+ 自动化的价值实现时间上线后 4 周2026-02Serval 已发布基准中通过专属前置部署工程试点实现快速部署客户群样本量和纳入标准
SeatGeek 早期工单自动化3 周达到 40%,60 天达到 50%+2026-01通过 Macha 发布的 SeatGeek 案例高证明随着工作流调优,自动解决曲线会迅速变陡每月基线工单总量和复杂度分布
Perplexity 自动处理请求占比超过 50% 的 IT 请求2025-10IT 负责人 Vernon Man 引述高说明无需增加员工,核心 IT 服务台工作量也能减半工单绝对数量及升级给人工工程师的比例
Together AI JIT 访问自动化率95% 的 JIT 访问申请2026-04Serval 工作流平台报告高验证访问治理是杠杆率最高的单一自动化场景每日访问申请总量和例外覆盖率

指标反映公司发布的案例研究和客户高管引述;总体基线工单量仍属私有信息。

[CU002, CU003, CU004, CU005]

6.3 已命名客户部署与案例证据

公开核实的客户部署证明,Serval 在核心账户中已经是实际运行的行动系统,而非实验性试点。Perplexity、Together AI、SeatGeek 和 Ramp 等知名参考客户, 已将 Serval 部署到日常 IT 运营的完整生产环境。在 Ramp,IT 团队利用 Catalyst 工作流智能体从历史工单日志起草并发布定制自动化工作流,速度提升 50%。 AWS Marketplace 的经核实用户反馈确认,Serval 能有效承担一线和二线 IT 团队成员的工作,把权限请求完成时间从数天压缩到数分钟,同时统一政策执行。 但公开案例几乎全部来自云环境整洁、技术导向的软件企业,平台如何处理混乱的传统本地基础设施仍未解决。[CU001, CU002, CU003, CU004, CU012]

具名客户验证表
客户分群部署与主要用例生产环境还是试点已记录成果已观察到的局限
Perplexity AIAI 搜索引擎IT 服务台自动化、入职与离职流程自动化生产环境自动处理超过 50% 的 IT 工单;核心入职流程从耗时数小时变为即时完成技术栈集中,缺少传统企业 ERP 的复杂性
Together AIAI 云基础设施即时访问权限配置、安全审计轨迹生产环境95% 的 JIT 访问请求实现自动化,且运行过程完全可审计主要局限于特权访问工作流
SeatGeek消费科技 / 票务员工 IT 服务台、重复请求拦截生产环境3 周内达到 40% 自动化,60 天内超过 50%首次部署需要 IT 团队定制和打磨工作流
Ramp金融科技 / 企业卡Catalyst 工作流生成、工单历史分析生产环境IT 团队起草并部署新自动化工作流的速度提升 50%高度依赖具备工程能力的 IT 人员审查代码

枚举反映了公开识别且有部署结果归因的生产客户。

[CU001, CU002, CU003, CU004]

6.4 账户扩展循环与持久性风险

客户扩张遵循自然的先落地后扩展轨迹:解决核心 IT 帮助台工作量和权限请求后,企业客户会把 Serval 横向拓展到 HR 入职、安全权限审查、法务合同工作流和工程配置。 定制 TypeScript 工作流带来显著技术切换成本,锁定客户并推动跨部门合同增长,而无需同比增加销售成本。但净收入留存、毛收入留存和年度客户标志流失率等关键留存指标 完全未披露。此外,Serval 采用需演示获取、没有公开价位层级的定制平台费,采购透明度有限,续约涨价上限因此成为潜在机构投资人必须核查的关键事项。[CU007, CU008, CU009, CU010, CU011, CU014]

6.5 展示项

Chapter 07

07风险

7.1 监管合规、隐私治理与数据驻留

对 Serval 监管态势的评估显示,公司高度依赖合同数据保护和云隔离选项来应对企业合规要求。根据主服务协议和数据处理附录,Serval 是客户 IT 请求的数据处理者, 将支持互动路由给获授权的第三方 AI 子处理者。公司虽在合同中保证绝不保留客户材料、输入和输出以训练或微调第三方模型,但如果上游基础模型供应商发生遥测泄露或 未经授权的日志记录事件,企业安全团队仍面临残余暴露。为满足数据主权及医疗或金融合规要求,Serval 提供灵活部署架构,包括单租户、客户自有的 AWS VPC, 以及无需持续外部 IAM 访问的自托管 Kubernetes 集群。[CR006, CR007]

监管 / 法律风险登记表
规则 / 法律框架管辖区状态可能性严重性缓解措施剩余暴露尽调路径
子处理者数据处理与 AI 隐私美国 / 欧盟 / 全球持续合规中严重合同 DPA 条款禁止使用客户输入训练模型;持续维护获授权 AI 子处理者清单若第三方 LLM 提供商发生数据泄露或未经授权的日志记录,剩余责任风险高检查子处理者 BAA/DPA 条款,并与 OpenAI/Anthropic 核实零留存 API 的执行情况
企业访问与凭证治理(SOX / SOC 2)美国 / 全球已认证标准中高即时访问权限配置、限时权限,以及对访问变更自动生成不可篡改的审计日志若客户管理员错误配置特权访问工作流,暴露程度中等查阅 SOC 2 Type II 审计报告,测试访问撤销控制,并检查 SIEM 导出管道
生成代码著作权与知识产权赔偿美国联邦标准条款低高工作流生成确定性的 TypeScript,而非原始模型代码片段;MSA 授予客户对输出的完整所有权AI 生成的自动化脚本可能引发第三方代码相似性法律主张,暴露程度中等检查 MSA 中的知识产权保证条款,并确认供应商专利 / 著作权赔偿的范围
医疗与金融监管要求(HIPAA / GLBA)美国联邦 / 州自证支持中中可选的自托管 Kubernetes 或客户 VPC 部署方案,用于隔离敏感 PHI/PII 数据客户需管理基础设施补丁和 VPC 网络隔离,运营负担中等验证商业伙伴协议(BAA)的签署准备情况,并审查客户 VPC 部署架构

严重性按当前美国及国际隐私框架下潜在的运营和财务影响排序;状态依据公开合同文件确定。

[CR006, CR007]

7.2 运营安全、工作流执行与智能体控制

Serval 的运营风险集中在自主工作流可能对生产企业环境触发意外修改。为缓解对话式 AI 固有的提示注入漏洞,平台在自然语言 Help Desk Agent 与后端工作流执行之间 设置架构隔离。所有 Catalyst 自动化都会编译成确定性 TypeScript 代码,执行前必须经过管理员明确审查、版本控制和发布。但企业客户部署显示,在映射复杂跨部门权限时, 初期存在运营学习曲线和配置负担。此外,主动后台智能体会持续检查已连接的 IT 基础设施,在工单提交前起草修复方案;如果放松管理员审查护栏,可能引入运营扰动。[CR002, CR005, CR010, CR012, CR013]

运营 / 质量 / 安全风险登记表
故障模式可能性严重性缓解措施成熟度剩余暴露未解决缺口
通过遭入侵的工作流执行未经授权的生产操作低严重高中对 Catalyst 工作流执行沙箱进行独立第三方渗透测试
上游前沿 LLM API 中断或延迟激增中高中高明确主模型提供商遭遇限流或服务中断时的备用延迟 SLA
初次部署期间的管理配置错误高高中中标准化客户入职价值实现时间基准,以及自动化配置验证工具
主动式代理漂移或非预期配置变更中中高低细粒度审计日志,衡量企业管理员自动回滚修复的频率

故障模式覆盖平台执行各层,并按其对企业 IT 连续性的潜在严重性排序。

[CR001, CR002, CR005, CR010, CR012, CR013]

7.3 平台依赖、基础 LLM 暴露与竞争动态

Serval 商业模式的根本脆弱性在于完全依赖外部基础模型和既有软件生态。Serval 不构建或维护自有基础模型,而是通过 OpenAI 和 Anthropic 的前沿 API 路由任务。 这使 Serval 暴露于上游突然调价、限速约束和 API 弃用。与此同时,ServiceNow 和 Atlassian 等传统企业服务管理厂商正在把原生智能体能力和对话式构建器嵌入根深蒂固的平台。 虽然 Serval 当前支持与 ServiceNow、Jira Service Management 和 Freshservice 双向同步,但如果现有厂商限制 API 吞吐,或将生成式自动化工具激进地捆绑进现有企业许可协议, 这种中间层定位仍很脆弱。[CR001, CR003, CR011]

合作伙伴 / 依赖风险登记表
依赖项交易对手方角色集中度故障情景严重性缓解措施剩余暴露
基础 LLM 提供商OpenAI / Anthropic核心推理、分流和 TypeScript 工作流生成高模型突然弃用、API 价格飙升或服务条款收紧严重与模型无关的路由框架,支持客户提供 API 密钥高
既有 ITSM 平台ServiceNow / Atlassian传统系统记录源和双向工单同步伙伴高既有平台限制 API 吞吐量、调整合作伙伴费用或复制核心代理功能高提供原生工单、访问和资产管理的全栈替代能力高
身份与目录提供商Okta / Microsoft EntraJIT 访问和员工入职的身份事实源高连接器故障或不兼容的 IDP API 变更阻断自动用户配置高预置连接器、SCIM 合规以及多租户 API 范围上限中
云基础设施托管Amazon Web Services (AWS)主要 SaaS 托管、计算和市场渠道中区域性云中断或 AWS Marketplace 政策修订中多区域架构及兼容自托管 Kubernetes 部署低

依赖项覆盖关键外部基础设施、API 合作伙伴和平台集成对手方,并按故障严重性排序。

[CR001, CR003, CR011]

7.4 客户集中、定价摩擦与财务不透明

Serval 的商业进展带有明显的科技初创客户集中和高度估值不透明特征。早期案例和公开参考客户主要是风险投资支持的 AI 与软件公司,包括 Perplexity、Together AI、 Clay 和 Mercor。虽然这些高速增长的科技客户展示了快速采用,Serval 在制造、医疗和零售等传统强监管企业垂直领域仍基本未经验证。此外,公司采用没有公开层级、 依赖高资源投入前置部署工程试点的不透明、需演示获取的定价模式。尽管公司声称收入增长 500% 并取得 $1B B 轮估值,但尚未公开披露 ARR、毛利率、获客成本或现金消耗, 无法进行标准的单位经济投资判断。[CR004, CR008, CR009]

FR001: 风险热力图

从技术、运营、平台和商业等核心风险维度,系统评估严重程度与发生可能性。

定性风险评级综合自客户案例、技术文档、架构审查和市场比较报告。

[CR001, CR003, CR004, CR008, CR009, CR014]

7.5 风险监控框架与下行论点破坏触发器

以独角兽估值评估 Serval,需要持续监测关键运营、技术和商业里程碑。关键指标包括非科技企业客户获取速度、子处理者 API 毛利率稳定性,以及试点结束后的客户留存指标。 关键下行论点破坏触发器包括:现有厂商限制 API 合作伙伴访问或掠夺式捆绑引发的激烈反击;未经授权的智能体脚本执行造成大范围安全漏洞;以及在风险融资收缩期间, 核心科技初创客户群的客户流失率超过 20%。若这些触发器中的任何一个出现,Serval 作为独立系统记录源的竞争护城河将迅速恶化。[CR001, CR003, CR008, CR014]

7.6 展示项

Chapter 08

08估值

8.1 投资论点、立场与建议逻辑

Serval 的投资论点颇具吸引力:它试图把企业 IT 服务管理从人工、路由工单的文件柜,变成自主执行平台。公司引入双智能体架构,由 Catalyst Agent 构建自然语言工作流, 由 Help Desk Agent 确定性执行,使 Perplexity、Mercor 和 Ramp 等前瞻科技客户能够自动化超过 50% 的日常 IT 支持请求。这种运营效率解决了企业的一大摩擦点, 显著缩短分散技术团队的平均解决时间,并降低支持工单处理成本。 但投资建议严格为“跟踪”,估值立场为“偏高”。技术范式转变可信,也得到一级风险投资支持的验证;然而,公开证据没有提供支撑 $1.0B 投后估值所需的基线数据。 当前客户案例几乎全部集中在风险投资支持的 AI 和软件初创公司,能否扩展到拥有复杂传统基础设施的非科技企业仍有疑问。在 Serval 证明企业扩张可重复、并披露 可验证的 ARR 里程碑之前,应推迟新增资本承诺。[CV001, CV005, CV006, CV013]

建议汇总表
维度评估支持证据决策含义
建议跟踪产品自动化已在科技初创企业中得到验证,但缺少非科技企业证据在 ARR 规模和企业客户可复制性得到证明前,不投入成长股权
信心中已有文件记录 $1B Series B 和 Series A 轮融资,但绝对收入和单位经济效益未披露没有经审计财务数据,不按溢价入场倍数评估投资
风险评级高ServiceNow/Moveworks 的既有平台反击,以及 LLM 自动化可能商品化要求明确的 IP 防御力和多代理工作流粘性证据
估值立场偏高ARR 缺少锚定,估值在 4 个月内上调 4.3 倍至 $1B 投后估值在二级市场争取 40-50% 折价,或等待 Series C 提供估值锚点

汇总评级综合 Series A/B 融资条款、客户自动化指标和既有平台的竞争威胁程度。

[CV001, CV002, CV003, CV013]
FV001: Serval 投资决策逻辑

将企业问题规模、技术架构、客户进展与竞争风险依次串联,得出最终的跟踪建议。

节点 / 连接节点 / 起点终点说明
节点 1 企业 IT 瓶颈 [n1]手工处理工单、重置密码和脆弱的工作流广泛拖累生产力
节点 2 智能代理 AI 架构 [n2]Catalyst Agent 生成 TypeScript 工作流;Help Desk Agent 借助气隙隔离确定性执行
节点 3 早期客户进展 [n3]Perplexity、Mercor 和 Ramp 等知名科技客户的工单自动化率达到 50%+
节点 4 企业壁垒与在位者 [n4]ServiceNow 的规模(季度收入 $3.88B)和模型商品化的可能性威胁扩张
节点 5 估值倍数脱节 [n5]在未披露 ARR、单位经济性或 SEC Form D 文件的情况下,估值跃升 4.3x 至 $1B 投后估值
节点 6 建议:跟踪 [n6]在等待经审计的企业收入披露期间,维持跟踪立场,并将估值评级定为偏高
连接 1 企业 IT 瓶颈 [n1]智能代理 AI 架构 [n2]由……解决
连接 2 智能代理 AI 架构 [n2]早期客户进展 [n3]推动
连接 3 早期客户进展 [n3]企业壁垒与在位者 [n4]面对
连接 4 企业壁垒与在位者 [n4]估值倍数脱节 [n5]拉高
连接 5 估值倍数脱节 [n5]建议:跟踪 [n6]得出结论

该流程将定性尽调发现抽象成一套顺序决策框架。

[CV005, CV006, CV007, CV013]

8.2 融资背景、估值倍数脱节与反向论点

Serval 的资本化速度异常快:从 2025 年 10 月以 $232M 估值完成 $47M A 轮,到 2025 年 12 月由 Sequoia Capital 领投、以 $1.0B 投后估值完成 一笔抢先完成的 $75M B 轮。四个月内估值增长 4.3 倍,支撑因素是公司声称收入增长 500% 和员工数翻三倍。Sequoia 明确将 Serval 与其 16 年前早期投资 ServiceNow 相提并论,强调以 IT 为中心的系统记录源横向拓展至 HR、财务和法务工作流的潜力。 但核心反向论点在于现有厂商根深蒂固,以及缺乏锚点的估值定价。ServiceNow 仍是企业服务管理领域的主导者,2026 年第二季度季度订阅收入达到 $3.88B, 并于 2025 年 12 月以 $2.85B 收购对话式 AI 专家 Moveworks,以强化 AI 原生的前线防御。此外,在没有经审计的绝对收入或净留存指标时,Serval 的 $1.0B 估值明显脱离传统软件基本面。PagerDuty 等中端市场 IT 支持现有厂商的 EV/Revenue 约为 2.5 倍,说明 Serval 按高位投机增长倍数定价,执行失误没有任何安全边际。[CV001, CV002, CV003, CV004, CV007]

正 / 反论点表
视角核心论点证据锚点哪些情况会改变判断
投资正论(市场)原生 AI 的 ITSM 用自主自然语言工作流替代脆弱的手工工单处理Perplexity、Mercor 和 Ramp 的 IT 工单自动化率超过 50%客户流失,或无法拓展到技术导向初创企业之外
投资正论(产品)Catalyst 构建代理与确定性服务台运行时分离,带来可靠的企业级执行生成 TypeScript 工作流,并配有人机协同审计日志和 RBAC客户生产环境出现运行时执行失败或安全漏洞
反论(竞争)ServiceNow $3.88B 的季度收入和对 Moveworks 的收购构成分销护城河,压缩 Serval 的 TAMServiceNow 已部署 Level 1 Service Desk AI Specialist,并拥有既有企业 ERP 覆盖已证明能在 Global 2000 客户中取代 ServiceNow 作为主系统记录源
反论(估值)按估计低于 $25M 的 ARR 和 30 人团队计算,当前 $1B 估值隐含的远期收入倍数高得惊人Series B 以 $1B 投后估值提前锁定,公司声称收入在未披露基数上增长 500%经审计 ARR 超过 $50M,净收入留存率超过 130%,毛利率超过 75%

正论点反映公司声称和投资者口径;反论点反映独立市场可比数据和既有平台的分销规模。

[CV005, CV006, CV007, CV010]

8.3 情景投资判断与可比倍数分析

为评估 $1.0B B 轮基准下的潜在回报,我们围绕 2026 至 2028 年投资期限建立乐观、基准和悲观运营情景。乐观情景(20% 概率)下,Serval 成功把 Catalyst 打造成覆盖 HR、采购和法务部门的横向工作流自动化引擎,ARR 增长超过 $80M,支持以 20 倍至 30 倍倍数实现 $2.5B–$3.5B 退出。基准情景(50% 概率)下, 公司在科技和中端市场账户中稳步扩张,2028 年 ARR 达到 $30M–$45M,估值达到 $750M–$1.2B,相对 B 轮入场价格带来持平至温和回报。 悲观情景(30% 概率)下,Serval 遇到结构性企业摩擦。安全、合规和定制集成要求显著拉长企业 IT 销售周期,基础模型进步也会让基础对话式工单商品化。 随着 ServiceNow 在现有企业合同中推出原生 AI 专家智能体,Serval 的增长放缓至 ARR 低于 $20M,最终以 $200M–$350M 完成低价轮融资或陷入困境的战略收购。 悲观情景被赋予较高概率,凸显进入缺乏锚点的 B 轮独角兽融资所固有的脆弱性。[CV008, CV009, CV010, CV011, CV012]

乐观 / 基准 / 悲观情景表
情景关键运营假设估值 / 回报逻辑关键风险与下行触发因素概率信号
乐观情景ARR 到 2028 年扩大至 $80M+;横向拓展至 HR/财务/法律;NRR >140%;在企业客户中取代传统 ITSM通过 IPO 或 Microsoft/Salesforce 以 ARR 的 20x–30x 战略收购,实现 $2.5B–$3.5B 退出ServiceNow 生态锁定;工作流 LLM 商品化20%(要求企业执行毫无瑕疵)
基准情景ARR 到 2028 年达到 $30M–$45M;在中型科技市场渗透强劲;企业拓展温和;NRR 约 115%按 ARR 的 15x–20x 计算,估值为 $750M–$1.2B;相对 Series B 入场价回报持平至小幅增长吃尽科技初创企业垂直市场后增长放缓;LLM 推理带来利润率压力50%(结合当前客户基础,这是最可能的轨迹)
悲观情景ARR 停滞在 $20M 以下;企业试点未通过安全 / 合规审查;Moveworks/ServiceNow 套装产品压垮独立产品定价按 ARR 的 5x–8x 进行 $200M–$350M 困境收购或下轮重组融资;资本损失超过 65%企业客户流失、代理访问权限引发安全事件、现金消耗高30%($1B 入场估值的高位压力使该概率上升)

情景以 Series B 投后估值为基准,覆盖 2026–2028 年;概率信号反映缺少经审计 ARR 和单位经济效益。

[CV008, CV009, CV010, CV011, CV012]

8.4 退出期限、尽调问题与论点破坏触发器

Serval 的退出准备仍处于早期发展阶段。公司团队规模不足 30 人,成立还不到两年,因此 IPO 至少还要 3—5 年,且前提是经审计的 GAAP 业务规模超过 $100M ARR 并实现正自由现金流。在此之前,被寻求现代智能代理工作流编排层的 Microsoft、Salesforce 或 Atlassian 等企业生态巨头战略收购,是中期最可行的流动性路径。 在参与任何潜在的老股交易或未来 C 轮融资前,机构投资者必须要求四项不可妥协的尽调信息:(1)绝对 ARR 和月度烧钱率,用来核实 500% 的增长说法;(2)计入第三方 LLM 推理费用后的经审计毛利率;(3)客户群留存曲线和净收入留存率(NRR)基准;(4)向非科技行业《财富》1000 强客户做背调访谈。可量化的论点失效触发条件包括:同比 ARR 增长降至 100% 以下、毛利率跌破 65%、关键高管离职,或发生任何涉及自主特权访问配置的重大安全漏洞。[CV013, CV014]

8.5 展品

免责声明

本报告基于公开可得证据,不应视为经审计的财务建议或完整的尽调材料包。

证据索引

结论
编号陈述可信度来源
CO001 Serval was founded in 2024. 高 SO001, SO003, SO004
CO002 Serval is headquartered in San Francisco, California. 高 SO001, SO003
CO003 Serval combines help desk, access management, asset management, and workflow automation in an AI-native enterprise service platform. 高 SO001, SO005, SO007
CO004 Serval was co-founded by CEO Jake Stauch and CTO Alex McLeod. 高 SO001, SO003, SO005
CO005 Christine Kim, a former Greylock partner, leads Serval Start as Serval's Head of Strategic Projects. 中 SO001
CO006 Serval's $47 million Series A was announced on October 21, 2025, led by Redpoint, and brought reported total funding at that date to $52 million. 中 SO003, SO004, SO005
CO007 Serval announced a $75 million Series B led by Sequoia on December 11, 2025 at a $1 billion valuation. 高 SO002, SO003, SO005
CO008 Serval reported $127 million in total capital raised since August 2025 as of its Series B announcement. 中 SO001, SO002
CO009 Multiples records $122 million across Serval's two priced rounds, $5 million below the company's reported cumulative total. 中 SO002, SO003
CO010 Serval claimed revenue grew 500% during the roughly three months after its Series A without disclosing an absolute revenue or ARR base. 中 SO002, SO005
CO011 Serval claimed headcount more than tripled after its Series A without disclosing the starting or ending employee count. 中 SO002, SO005
CO012 Serval's buyer guide claims its customers routinely resolve 50–80% of help desk tickets automatically and names four production users. 中 SO008
CO013 SiliconANGLE reported six named Serval customers using the platform to automate more than 50% of IT tickets. 中 SO004
CO014 Macha found that Serval publishes no pricing tiers or starting figure and instead routes buyers through a demo and custom-contract process. 中 SO003, SO005
CO015 An AWS Marketplace reviewer described Serval as powerful but reported a learning curve and an initially overwhelming configuration surface. 中 SO006
CO016 Serval reported that its first Serval Start cohort selected 12 participants from more than 500 applicants. 中 SO001
CM001 Serval defines its product market as enterprise service management for employee requests, onboarding, offboarding, and just-in-time access. 中 SM002
CM002 Serval says bidirectional synchronization with major ITSM systems allows deployment as an automation layer before a system-of-record migration. 中 SM002
CM003 Serval says customer workflows have expanded from IT into HR, finance, and legal. 中 SM004
CM004 Serval reports that Perplexity reached an automation rate above 50%. 中 SM004
CM005 Serval's 2026 buyer guide says its customers routinely resolve 50%-80% of help-desk tickets automatically. 低 SM005
CM006 Serval's buyer guide cites a $15-$25 live-agent ticket cost and a $2-$4 AI-resolved ticket cost. 低 SM005
CM007 Siit reports that Serval uses demo-gated custom pricing with no public pricing tiers. 低 SM001
CM008 An AWS Marketplace reviewer reports that configured Serval access workflows reduced request completion from one or two days to minutes. 中 SM003
CM009 The AWS Marketplace reviewer reports a learning curve and an initially overwhelming configuration surface. 中 SM003
CM010 Serval says production actions are limited by explicitly published workflows, authorized API scopes, granular access controls, and approval policies. 中 SM007
CM011 Macha characterizes Serval's best-fit buyer as a funded mid-market-to-enterprise IT or operations team. 低 SM008
CM012 Macha characterizes small teams without dedicated IT headcount as a weaker fit because workflow ownership requires ongoing effort. 低 SM008
CM013 Serval says it supports cloud-hosted, hybrid, and fully self-hosted deployments. 中 SM006
CM014 Macha reports that Serval covers ticketing, access management, asset management, and workflow orchestration. 低 SM008
CM015 Serval says generated workflow code is reviewed and published before it executes deterministically at request time. 中 SM006
CP001 The enterprise IT service management market spans legacy systems of record, direct AI-native platforms, and integration-first conversational deflection overlays. 中 SP001, SP004
CP002 Atomicwork is the closest direct AI-native competitor to Serval, deploying role-based AI coworkers to replace or augment legacy ITSM deployments. 中 SP004
CP003 Serval differentiates its architecture through two distinct agents: Catalyst builds deterministic code-based workflows from natural language prompts, while the Help Desk Agent executes them within published guardrails. 高 SP003, SP006
CP004 Direct competitors in AI-native service desks include Aisera with its GenIQ multi-agent engine and Espressive with its Employee Language Cloud. 中 SP009
CP005 ServiceNow acquired Moveworks to integrate conversational AI and multilingual LLM-powered ticket deflection directly into its dominant enterprise platform ecosystem. 中 SP004, SP009
CP006 Integration-first AI vendors such as eesel AI offer lightweight overlays that connect into existing help desks like Zendesk or Jira Service Management without displacing the system of record. 中 SP001
CP007 Platform incumbents including ServiceNow, BMC Helix, and Atlassian are embedding AI agent capabilities and LLM workflows into their existing platforms to prevent customer migration to AI-native alternatives. 中 SP009
CP008 Unlike black-box conversational bots that improvise execution at runtime, Serval compiles workflows into version-controlled TypeScript that runs deterministically. 高 SP006, SP008
CP009 Serval enforces an architectural air gap between conversational request intake and workflow generation, preventing end-user prompts from modifying or creating automation code. 中 SP008
CP010 Enterprise customers deploy Serval for automated just-in-time access provisioning, reducing multi-day SaaS permission approval cycles to minutes. 高 SP002, SP005
CP011 Serval supports bi-directional ticket synchronization with ServiceNow, Jira, and Zendesk, enabling coexistence as an automation layer or progressive cutover as primary system of record. 高 SP001, SP003
CP012 Serval does not publish standardized subscription pricing or public tier rates, selling exclusively through custom enterprise quotes and engineer-led pilots. 中 SP001, SP004
CP013 Competitor Atomicwork publishes a transparent pricing model starting at $25,000 annually with outcome credits, in contrast to Serval's demo-gated sales motion. 中 SP004
CP014 Serval guarantees at least 50% automated resolution of incoming IT service desk tickets by the conclusion of its four-phase guided deployment pilot. 中 SP001, SP004
CP015 Serval positions itself higher on deterministic automation execution depth than conversational deflection bots, while targeting broader enterprise horizontal scope than single-department point solutions. 中 SP001, SP004, SP008
CI001 Serval completed a $75 million Series B financing in December 2025 led by Sequoia Capital at a post-money valuation of $1 billion, following a $47 million Series A in October 2025. 高 SI002, SI003
CI002 Serval reported 500% revenue growth between August and December 2025 alongside a tripling of company headcount during the second half of 2025. 中 SI001, SI003
CI003 Serval commercializes its platform under a seat-based SaaS subscription model without variable execution fees or API consumption surcharges. 中 SI004, SI005
CI004 Serval's customer roster includes enterprise accounts such as Spotify, Fox, and Live Nation alongside venture-backed technology companies like Perplexity, Mercor, and Together AI. 高 SI002, SI005
CI005 Serval does not publish transparent pricing tiers, seat prices, or contract floors, operating instead on a demo-gated, custom-quote enterprise sales motion. 中 SI008
CI006 Serval deploys forward-deployed engineering teams to pilot implementations with prospective enterprise clients, absorbing front-end labor costs to secure larger annualized contract values. 中 SI008
CI007 Serval expands net expansion revenue by transitioning from an initial IT help desk entry point into horizontal departmental workflows across HR, finance, legal, and security. 中 SI001, SI007
CI008 Industry benchmarks indicate that autonomous AI-driven ticket resolution operates at an economic cost of $2 to $4 per ticket compared to $15 to $25 for human live-agent handling. 中 SI006
CI009 Serval's primary cost of goods sold and operating expenses derive from cloud hosting infrastructure, large language model inference costs, and enterprise integration maintenance. 中 SI009
CI010 Serval's gross profit conversion depends on maintaining efficient LLM token consumption and caching across repeated deterministic workflow executions. 中 SI007, SI009
CI011 Serval has raised a total of $127 million in equity capital across its seed, Series A, and Series B rounds, providing an estimated multi-year cash buffer at sub-100 employee scale. 中 SI001, SI003
CI012 Serval's demo-gated pricing strategy enables value-based price discrimination against enterprise buyers but exposes the company to pricing friction and longer sales cycles against transparent competitors. 中 SI008
CI013 Public sources disclose neither absolute Annual Recurring Revenue (ARR) nor baseline gross margin figures, leaving Serval's 500% revenue growth claim unanchored by verified dollar amounts. 中 SI001, SI008
CI014 Underwriting Serval's $1 billion Series B valuation requires verified data room disclosure of GAAP gross margins, net revenue retention (NRR), and customer acquisition cost payback periods. 中 SI003, SI008
CE001 Serval provides an AI-native IT service management platform built around a two-agent architecture that separates request resolution from workflow authoring. 高 SE001, SE006
CE002 The Catalyst agent converts natural language administrative prompts into deterministic, inspectable TypeScript workflows rather than executing probabilistic runtime prompt chains. 高 SE001, SE002
CE003 Serval enforces an architectural air gap where the end-user Help Desk agent is restricted to executing pre-approved, published workflows and cannot author new automation logic. 高 SE001, SE008
CE004 Serval Databases acts as an integrated CMDB and context graph that ingests hardware, software licenses, SaaS subscriptions, and configuration items from MDM, IdP, and HR systems. 中 SE002
CE005 Serval supports just-in-time access provisioning with time-bounded permissions, automated deprovisioning, and scheduled access reviews designed to satisfy SOC 2 and ISO 27001 requirements. 高 SE002, SE007
CE006 Proactive Agents run scheduled sessions across tickets, workflows, and databases to diagnose emerging problems, uncover software license waste, and stage proposed remediations for administrative review. 高 SE001, SE002
CE007 Serval provides two-way data synchronization with legacy ITSM systems including ServiceNow, Jira Service Management, and Freshservice, enabling phased deployments without requiring immediate rip-and-replace migrations. 高 SE001, SE002, SE007
CE008 Deployment options for Serval comprise a fully managed cloud SaaS with a 99.9% uptime SLA, a hybrid model running network workers locally, and a self-hosted single-tenant AWS or Kubernetes deployment. 高 SE003, SE008
CE009 Security controls across the platform incorporate TLS 1.3 encryption in transit, AES-256 encryption at rest, SOC 2 Type II certification, and an explicit API scope ceiling on third-party integrations. 高 SE003, SE008
CE010 Serval claims its platform enables customers such as SeatGeek and Perplexity to automate over 50% of incoming employee service requests without human intervention. 高 SE004, SE006
CE011 User feedback indicates Serval exhibits a steep operational learning curve and significant configuration complexity during initial setup before full automation value is realized. 中 SE005
CE012 Incumbent ITSM providers such as ServiceNow, Freshworks, and Atlassian retain deeply entrenched enterprise customizations and extensive prebuilt ecosystem integrations that make full rip-and-replace displacement difficult. 高 SE001, SE002
CE013 Serval lacks a public developer portal, open-source SDKs, or public developer community metrics, making independent technical verification dependent on private customer references. 中 SE003
CE014 Serval utilizes a six-tier governance model comprising team segregation, builder RBAC, API scope ceilings, code-level execution approvals, deterministic execution, and an architectural air gap. 中 SE008
CU001 Serval has secured production deployments across notable technology scale-ups including Perplexity, Together AI, Ramp, SeatGeek, and Mercor. 高 SU001, SU006
CU002 Perplexity Head of IT Vernon Man reported that Serval automated more than half of their incoming IT tickets, including end-to-end onboarding and offboarding. 高 SU006, SU009
CU003 Together AI deployed Serval's access management architecture to automate 95% of its just-in-time access requests while retaining audit trails. 中 SU001, SU008
CU004 SeatGeek achieved 40% ticket automation within three weeks of deployment and surpassed 50% automated resolution within sixty days. 中 SU003
CU005 Serval published that its customer cohort consistently reaches 50% or higher ticket automation within four weeks of go-live. 中 SU002
CU006 Serval's disclosed customer roster is heavily concentrated in venture-backed technology startups, with limited proven penetration into traditional non-tech enterprises. 中 SU001, SU003
CU007 Serval structures customer onboarding around a dedicated forward-deployed engineering pilot that guarantees 50% ticket automation before full commitment. 中 SU001, SU003
CU008 Serval's pricing model is completely demo-gated with single quoted platform fees rather than public seat or per-resolution tiers. 中 SU001, SU003
CU009 Early enterprise customers expand Serval horizontally from IT service desks into HR onboarding, security access reviews, and engineering operations. 中 SU005, SU010
CU010 Building bespoke TypeScript automations creates technical switching costs that lock in customers while increasing contract values over time. 中 SU001, SU010
CU011 Serval's code-based workflow architecture rewards engineering-led IT teams but creates a steep initial learning curve for organizations seeking a simple no-code service desk. 中 SU001, SU004
CU012 An AWS Marketplace enterprise reviewer reported that Serval operates like a Level 1 and 2 team member, compressing access request fulfillment from days to minutes. 中 SU004
CU013 Customer accounts fall into distinct tiers spanning frontier AI labs, high-growth fintechs, and legacy enterprise pilot accounts. 中 SU001, SU006, SU010
CU014 Net revenue retention, gross renewal rates, and annual logo churn for Serval remain undisclosed in public disclosures and analyst databases. 低
CR001 Serval relies entirely on third-party frontier LLM providers (specifically OpenAI and Anthropic) rather than proprietary foundation models, exposing the platform to upstream API availability, pricing changes, and model deprecation risks. 高 SR002, SR003
CR002 Serval's two-agent architecture enforces an air gap between conversational triage and workflow execution to prevent prompt injection and unauthorized execution of production automations. 高 SR003, SR007
CR003 Legacy ITSM incumbents, including ServiceNow and Atlassian, are embedding generative AI agents and natural language workflow builders directly into their existing platform ecosystems, threatening Serval's architectural differentiation. 高 SR002, SR004
CR004 Serval's commercial pricing is strictly demo-gated with no published tier or starting price, requiring vendor-led pilots with forward-deployed engineering teams that increase sales friction and lengthen procurement cycles. 中 SR004
CR005 Customer deployments reveal an initial configuration learning curve and administrative overhead before organizations can realize autonomous ticket resolution across IT and SaaS applications. 中 SR005
CR006 Under its Master Services Agreement and Data Processing Addendum, Serval acts as a data processor and routes customer support data through authorized AI subprocessors, requiring strict vendor compliance oversight. 高 SR002, SR006
CR007 Serval mitigates data residency and tenant isolation concerns by providing flexible deployment options across multi-tenant cloud, customer-owned AWS VPCs, and self-hosted Kubernetes clusters. 高 SR002, SR006
CR008 Serval's early customer roster is heavily concentrated among venture-backed AI and software companies (such as Perplexity, Together AI, Clay, and Mercor), leaving the company vulnerable to tech sector downturns and unproven in traditional enterprise environments. 中 SR001, SR004
CR009 Serval claims 500% revenue growth between August and December 2025 and a $1 billion Series B valuation, but has not disclosed ARR, gross margins, cash burn rate, or unit retention economics. 中 SR001, SR004
CR010 Catalyst automations execute as deterministic, compiled TypeScript code synced with version control systems rather than dynamic LLM generation at ticket runtime. 高 SR003, SR006
CR011 Serval supports two-way synchronization with ServiceNow, Jira Service Management, and Freshservice, but faces interface fragility and API deprecation risks from incumbent vendors seeking to defend their systems of record. 高 SR003, SR006
CR012 Background proactive agents continuously monitor connected switches, identity providers, and endpoints, introducing risks of configuration churn or unintended policy modifications if admin approval controls are bypassed. 高 SR002, SR008
CR013 Serval enforces role-based access control and maximum API scope ceilings per integration so that generated workflows cannot exceed authorized administrative permissions. 高 SR006, SR007
CR014 A critical thesis-break trigger for investors would be incumbent platform lock-out via restricted API access or an inability to convert tech startup pilot wins into multi-million dollar non-tech enterprise renewals. 中 SR002, SR004
CV001 Serval raised a $75 million Series B round led by Sequoia Capital in December 2025 at a $1 billion post-money valuation. 中 SV004, SV005
CV002 Serval announced a $47 million Series A round led by Redpoint Ventures in October 2025, bringing total disclosed funding to $52 million at that time. 高 SV003, SV002
CV003 Serval post-money valuation stepped up 4.3x from $232 million at its October 2025 Series A to $1 billion at its December 2025 Series B. 中 SV004, SV005
CV004 Sequoia Capital preempted Serval Series B round based on customer feedback and company-reported 500% revenue growth over a four-month period. 中 SV004
CV005 Serval deploys a two-agent architecture comprising the Catalyst Agent that generates TypeScript workflows from natural language and the Help Desk Agent that executes them deterministically. 高 SV006, SV008
CV006 Serval customers including Perplexity, Mercor, and Ramp have reported automating over 50% of routine IT support tickets using the platform. 高 SV002, SV003
CV007 ServiceNow generated $3.88 billion in quarterly subscription revenue in Q2 2026 and acquired conversational AI provider Moveworks for $2.85 billion in December 2025. 中 SV002
CV008 Widespread accessibility of advanced foundation models poses a risk of technical commoditization that could erode pricing power for pure AI ticket-automation platforms. 中 SV005
CV009 Enterprise ITSM automation requires deep change management and custom integrations, creating extended sales cycles and adoption friction beyond tech startups. 中 SV005
CV010 Public IT management and observability comps such as PagerDuty trade at approximately 2.5x EV/Revenue and Datagroup trades at 1.4x EV/Revenue. 中 SV004
CV011 Industry benchmarks estimate AI-driven self-service ticket resolution costs at $2 to $4 per ticket compared to $15 to $25 for live human agent resolution. 中 SV007
CV012 Under a base-case scenario, Serval achieves $30M to $45M in ARR by 2028 and commands a $750M to $1.2B valuation at 15x to 20x ARR. 中 SV004, SV005
CV013 The overall investment recommendation for Serval is Track with a Stretched valuation stance, driven by unanchored revenue growth claims and an absence of disclosed unit economics. 中 SV001, SV004
CV014 Underwriting growth equity in Serval requires verifying audited ARR, gross margins, net revenue retention, and non-tech enterprise customer retention. 中 SV004, SV005
来源
编号出版方标题引文
SO001 Serval Serval Launches Program to Transform Employees Into Future Founders Founded in 2024 by CEO Jake Stauch and CTO Alex McLeod, Serval unifies help desk, access management, asset management, and workflow automation in a single platform powered by AI agents.
SO002 Serval Serval’s Next Chapter: Raising $75M to Build the New Era of Enterprise Automation and Service Management And today, we are proud to share that Serval has raised $75 million in Series B funding, led by Sequoia, bringing us to a $1 billion valuation. Redpoint, Meritech, First Round, General Catalyst, Evantic, Sound Ventures, Radical Ventures, and others also participated. In total, Serval has raised $127 million since August.
SO003 Multiples Serval Valuation, Funding & Investors | Multiples Serval has raised a total of $122M across 2 funding rounds.
SO004 SiliconANGLE Serval raises $47M to bring AI-native automation to IT operations In its one year since being founded, Serval has seen strong growth, with notable customers including Perplexity AI Inc., Clay Technologies Inc., Verkada Inc., Mercor Inc., Cribl Inc. and Together AI Inc. using the company’s platform to automate more than 50% of all IT tickets
SO005 Macha What Is Serval? AI ITSM Features, Pricing & Alternatives (2026) No independent third-party rating we could verify: G2 and TrustRadius both blocked automated access, and the one third-party review we found came from a competing ITSM vendor's site, which we excluded given an apparent factual error in it.
SO006 AWS Marketplace Serval by Serval | AWS Marketplace Serval has a bit of a learning curve. There's a lot under the hood, and it takes some time to really understand everything it can do, so early on it can feel a little overwhelming if you're trying to take it all in at once.
SO007 Serval AI-native IT service management: The guide for modern teams Catalyst Agent: This converts a natural language prompt into a functional workflow. You describe the task in plain English, and the system builds the code.
SO008 Serval 2026 Buyer's Guide: AI ITSM Systems That Deliver Immediate ROI Real customer data: teams using Serval routinely resolve 50–80% of help desk tickets automatically. Perplexity, Verkada, Together AI, and Mercor run Serval in production.
SO009 Startups Union Serval Leads to $1B Valuation - How? Serval operates a SaaS subscription model-straightforward pricing based on scale and usage.
SO010 Sacra Serval funding, news & analysis ServiceNow's acquisition of Moveworks creates a competitor that combines established enterprise relationships with AI-native capabilities. The integrated offering provides both workflow backend and conversational frontend within a single platform ecosystem.
SM001 Siit Serval AI Review: Features and Pricing (2026) Serval AI follows a demo-gated pricing model with custom contracts tailored to organizational requirements. Prospective customers must schedule a demonstration and engage with the sales team to receive pricing information, as no public pricing tiers are available.
SM002 Serval AI-Native Enterprise Service Management and ITSM Let Serval be the automation layer on top of your existing ITSM. Serval supports bi-directional sync with all major ITSMs so you can automate work on day one and cut over the system of record when you’re ready.
SM003 AWS Marketplace reviewer Serval by Serval | AWS Marketplace With Serval, once you build out the right workflows and automations, that whole process collapses down to a single approval or denial click, and most requests now get resolved in minutes instead of days.
SM004 Serval AI-native IT service management: The guide for modern teams For example, Perplexity, the AI answer engine, used Serval to identify high-volume tasks and convert them into automated workflows. It reached an automation rate of over 50% by using these insights to grow over time.
SM005 Serval 2026 Buyer's Guide: AI ITSM Systems That Deliver Immediate ROI The unit economics of AI ITSM are straightforward. A live-agent ticket costs $15–$25 to handle. An AI-resolved ticket costs $2–$4.
SM006 Serval Top AI-native ITSM tools in 2026 Serval supports three deployment models: fully cloud-hosted, hybrid (worker in your network, Serval hosts the product), or fully self-hosted in your own Kubernetes environment.
SM007 Serval Making IT Automation Safe and Secure Our help desk agent can only perform actions (workflows) you explicitly build/install and publish.
SM008 Macha What Is Serval? AI ITSM Features, Pricing & Alternatives (2026) Serval fits a mid-market-to-enterprise IT or operations team, likely at a funded tech company already comfortable evaluating AI-native tools, that wants one platform automating requests across IT, HR, security, legal and finance instead of stitching together department-specific tools.
SM009 SiliconANGLE Serval raises $47M to bring AI-native automation to IT operations In its one year since being founded, Serval has seen strong growth, with notable customers including Perplexity AI Inc., Clay Technologies Inc., Verkada Inc., Mercor Inc., Cribl Inc. and Together AI Inc. using the company’s platform to automate more than 50% of all IT tickets
SM010 Multiples Serval Valuation, Funding & Investors | Multiples Serval has raised a total of $122M across 2 funding rounds.
SP001 eesel AI Serval AI: An overview of the AI-native ITSM platform Serval is an AI-native ITSM platform that brings help desk, access management, ticketing, and workflow automation together in one product. It's pitched at IT teams that want to automate a meaningful share of internal employee requests.
SP002 G2 The G2 on Serval Serval is versatile and powerful. It operates like a genuine Level 1/2 member of the IT team, resolving real support volume on its own thanks to its AI-native design. What sets it apart is that it doesn't stay boxed into ticketing. Once you dig into the configuration, it turns into a general purpose automation layer for IT.
SP003 Serval AI-Native Enterprise Service Management and ITSM Serval can only take action against your systems using deterministic workflows that admins have expressly approved and published.
SP004 Macha What Is Serval? AI ITSM Features, Pricing & Alternatives (2026) Atomicwork is the closest direct comparison: both are AI-native ITSM platforms positioned against legacy ServiceNow-style deployments, both raised venture rounds recently, and both separate building automations from running them. Moveworks, now owned by ServiceNow (its homepage banner reads "ServiceNow officially acquires Moveworks"), is a second comparison point, though it's an older, more established platform.
SP005 AWS Marketplace Serval by Serval | AWS Marketplace With Serval, once you build out the right workflows and automations, that whole process collapses down to a single approval or denial click, and most requests now get resolved in minutes instead of days.
SP006 Serval AI-native IT service management: The guide for modern teams In an AI-native IT service management platform, the entire architecture is built around automation from the start. Workflows rely on real code, often written in TypeScript, rather than simple drag-and-drop rules.
SP007 Serval 2026 Buyer's Guide: AI ITSM Systems That Deliver Immediate ROI Deflection redirects a user to a knowledge base article or FAQ. The help desk agent is off the hook. The user is not... Automation takes the action. The password gets reset. The access gets provisioned. The device gets unlocked.
SP008 Serval Top AI-native ITSM tools in 2026 The Help Desk Agent does not build workflows. It does not generate code. There is no path from the conversational layer to the Catalyst Agent. This separation is the air gap.
SP009 Sacra Serval funding, news & analysis ServiceNow's acquisition of Moveworks creates a competitor that combines established enterprise relationships with AI-native capabilities. The integrated offering provides both workflow backend and conversational frontend within a single platform ecosystem.
SP010 SiliconANGLE Serval raises $47M to bring AI-native automation to IT operations In its one year since being founded, Serval has seen strong growth, with notable customers including Perplexity AI Inc., Clay Technologies Inc., Verkada Inc., Mercor Inc., Cribl Inc. and Together AI Inc. using the company’s platform to automate more than 50% of all IT tickets
SP011 Multiples Serval Valuation, Funding & Investors | Multiples Serval has raised a total of $122M across 2 funding rounds.
SI001 AI Market Watch Serval - AI Startup Profile | AI Market Watch Revenue grew 500% between August and December 2025; customers report 50%+ of IT tickets automated; headcount more than tripled in H2 2025; valuation grew from $232M (Aug 2025) to $1B (Dec 2025)
SI002 Computerworld Q&A: Serval CEO describes building an ‘AI-native' alternative to ServiceNow Serval — whose customers now include Spotify, Fox, and Live Nation — has received significant venture funding from Sequoia and others, and was most recently valued at around $1 billion.
SI003 Multiples Serval Valuation, Funding & Investors | Multiples In December 2025, Serval raised $75 million in a Series B round led by Sequoia Capital at a $1 billion post-money valuation, preempted due to rapid momentum including 500% revenue growth, tripled headcount, and customer expansions beyond IT.
SI004 Startups Union Serval Leads to $1B Valuation - How? Serval operates a SaaS subscription model-straightforward pricing based on scale and usage.
SI005 Serval Automate 80% of IT tickets in 24 hrs Serval uses predictable, seat-based pricing with no hidden API or credit fees so teams know exactly what they’ll pay regardless of automation volume.
SI006 Serval 2026 Buyer's Guide: AI ITSM Systems That Deliver Immediate ROI Analyst benchmarks put AI-driven self-service at $2–$4 per ticket versus $15–$25 for live-agent resolution, and most teams see measurable returns within 90–180 days when they target high-volume requests first.
SI007 Serval AI-native IT service management: The guide for modern teams Serval customers already use these AI agents to manage workflows in HR, finance, and legal departments. Tasks include managing the employee lifecycle and tracking contract approvals.
SI008 Macha What Is Serval? AI ITSM Features, Pricing & Alternatives (2026) A custom-quote model backed by named enterprise logos (Notion, Ramp, GitHub) and a $1 billion valuation lets Serval price each deal against that company's willingness to pay, with no public number for a competitor to undercut.
SI009 Sacra Serval funding, news & analysis Serval's cost structure includes cloud infrastructure, AI model usage, and integration maintenance, but benefits from the scalable nature of software automation.
SI010 SiliconANGLE Serval raises $47M to bring AI-native automation to IT operations In its one year since being founded, Serval has seen strong growth, with notable customers including Perplexity AI Inc., Clay Technologies Inc., Verkada Inc., Mercor Inc., Cribl Inc. and Together AI Inc. using the company’s platform to automate more than 50% of all IT tickets
SE001 JWEasyTech Serval CEO describes building an ‘AI-native’ alternative to ServiceNow – JWEasyTech Serval’s architecture is built around a separation of powers: one agent builds automations, and a separate agent uses them. Catalyst is gated to admins only and writes deterministic, code-based workflows. The help desk agent that end users interact with can call only the workflows that admins have already published; it cannot create new workflows or reach outside that pre-approved scope.
SE002 Expert Insights Serval: In-Depth Review & Analysis Serval Databases covers hardware, software licences, SaaS subscriptions, and configuration items against custom schemas, with configurable fields, statuses, and relationships between assets.
SE003 Serval Welcome to Serval - Serval Documentation Run just the Serval worker in your network (Hybrid), or the full application single-tenant in an AWS account you own.
SE004 Macha What Is Serval? AI ITSM Features, Pricing & Alternatives (2026) The Catalyst/help-desk-agent split separates workflow-building from workflow-execution instead of asking one agent to do both.
SE005 AWS Marketplace Serval by Serval | AWS Marketplace Serval has a bit of a learning curve. There's a lot under the hood, and it takes some time to really understand everything it can do, so early on it can feel a little overwhelming if you're trying to take it all in at once.
SE006 Serval AI-native IT service management: The guide for modern teams Workflows rely on real code, often written in TypeScript, rather than simple drag-and-drop rules. This ensures predictable execution and strong version control.
SE007 Serval 2026 Buyer's Guide: AI ITSM Systems That Deliver Immediate ROI Two-way sync with ServiceNow, Jira Service Management, and FreshService means Serval can run as a standalone ITSM or as an AI automation layer on top of existing infrastructure — which removes the need for a migration to get started.
SE008 Serval Top AI-native ITSM tools in 2026 The scope of what the Help Desk Agent can do is bounded entirely by the set of workflows that have already been built, reviewed, and published. Even if an employee's request is unusual or ambiguous, the Help Desk Agent cannot trigger actions outside that set.
SE009 Startups Union Serval Leads to $1B Valuation - How? Serval operates a SaaS subscription model-straightforward pricing based on scale and usage.
SE010 Sacra Serval funding, news & analysis Serval's cost structure includes cloud infrastructure, AI model usage, and integration maintenance, but benefits from the scalable nature of software automation.
SU001 Rezolve.ai Serval AI Review 2026: Pricing, Features & Catalyst Pricing is a single quoted platform fee... Its published customers are concentrated in technology, including Ramp, Perplexity, Together AI and Mercor.
SU002 Serval Comparing the Top AI-Powered Help Desk Solutions for 2026 Serval customers consistently reach 50%+ ticket automation within four weeks of go-live.
SU003 Macha What Is Serval? AI ITSM Features, Pricing & Alternatives (2026) SeatGeek's Elisa Folden reports going "from 0 to 40% automation within our first three weeks," and Perplexity's Head of IT, Vernon Man, reports completing "over 50% of our incoming requests automatically."
SU004 Amazon Web Services Marketplace Serval by Serval | AWS Marketplace With Serval, once you build out the right workflows and automations, that whole process collapses down to a single approval or denial click, and most requests now get resolved in minutes instead of days.
SU005 Serval AI-native IT service management: The guide for modern teams Perplexity, the AI answer engine, used Serval to identify high-volume tasks and convert them into automated workflows. It reached an automation rate of over 50% by using these insights to grow over time.
SU006 Serval 2026 Buyer's Guide: AI ITSM Systems That Deliver Immediate ROI Perplexity, Verkada, Together AI, and Mercor run Serval in production.
SU007 Serval Top AI-native ITSM tools in 2026 Serval is built on two separate agents: the Catalyst Agent and the Help Desk Agent.
SU008 Serval 11 best IT workflow automation platforms Together AI used this approach to automate 95% of their just-in-time access requests while maintaining complete auditability on every run.
SU009 SiliconANGLE Serval raises $47M to bring AI-native automation to IT operations With Serval, we've automated more than half of our IT tickets, including fully automating onboarding and offboarding across our core tools... What used to take me hours now happens instantly.
SU010 Sacra Serval funding, news & analysis Early customer implementations show HR, Security, and Engineering teams adopting Serval for their own service management needs after seeing IT success. This horizontal expansion within accounts can increase revenue per customer without proportional acquisition costs.
SU011 Multiples Serval Valuation, Funding & Investors | Multiples Serval has raised a total of $122M across 2 funding rounds.
SU012 Startups Union Serval Leads to $1B Valuation - How? Serval operates a SaaS subscription model-straightforward pricing based on scale and usage.
SR001 Silicon Valley Investclub Serval — Funding, Valuation & Leadership The company has grown revenue by 500% since August 2025 and automates over 50% of IT tickets for customers including Perplexity, Together AI, Clay, and Mercor.
SR002 VentureBeat Serval’s super agent Catalyst creates roving background agents to identify and fix IT issues before they’re ticketed Serval is not building its own foundation model. Stauch said in the interview that the company uses models from “frontier labs,” runs evaluations to determine which models work best for particular jobs, and is deliberately model-agnostic.
SR003 Serval AI-native IT service management: The guide for modern teams In an AI-native IT service management platform, the entire architecture is built around automation from the start. Workflows rely on real code, often written in TypeScript, rather than simple drag-and-drop rules.
SR004 Macha What Is Serval? AI ITSM Features, Pricing & Alternatives (2026) No published pricing, not even a starting figure, which costs evaluation time for a team shortlisting several vendors.
SR005 AWS Marketplace Serval by Serval | AWS Marketplace Serval has a bit of a learning curve. There's a lot under the hood, and it takes some time to really understand everything it can do, so early on it can feel a little overwhelming if you're trying to take it all in at once.
SR006 Serval 2026 Buyer's Guide: AI ITSM Systems That Deliver Immediate ROI Access management is built in, not bolted on. Serval handles just-in-time access with time limits, approval requirements, and justification capture.
SR007 Serval Top AI-native ITSM tools in 2026 The scope of what the Help Desk Agent can do is bounded entirely by the set of workflows that have already been built, reviewed, and published.
SR008 Serval AI-Native Enterprise Service Management and ITSM Serval can only take action against your systems using deterministic workflows that admins have expressly approved and published.
SR009 SiliconANGLE Serval raises $47M to bring AI-native automation to IT operations In its one year since being founded, Serval has seen strong growth, with notable customers including Perplexity AI Inc., Clay Technologies Inc., Verkada Inc., Mercor Inc., Cribl Inc. and Together AI Inc. using the company’s platform to automate more than 50% of all IT tickets
SV001 FundedIQ Serval: Funding, Investors & Team (Oct 2026) | FundedIQ Serval has raised $75.0M across 1 funding round on record, the most recent a Series B of $75.0M announced in Dec 2025.
SV002 Yahoo Finance Serval Bets It Can Automate the Work IT Admins Used to Do — and ServiceNow Should Pay Attention ServiceNow is a titan in this space, reporting $3.88 billion in subscription revenue for the second quarter of 2026 alone, with $1 billion in AI contract value. The company is moving aggressively, having closed its $2.85 billion acquisition of Moveworks in December 2025...
SV003 Serval Announcing $52M Total Raised to Deploy AI Agents for IT Today we’re announcing Serval’s $47M Series A, led by Redpoint Ventures with participation from First Round Capital, General Catalyst, Box Group, Bessemer Venture Partner, Chemistry, and others. This brings our total funding to $52M.
SV004 Multiples Serval Valuation, Funding & Investors | Multiples In December 2025, Serval raised $75 million in a Series B round led by Sequoia Capital at a $1 billion post-money valuation, preempted due to rapid momentum including 500% revenue growth, tripled headcount, and customer expansions beyond IT.
SV005 Sacra Serval funding, news & analysis Model commoditization: As large language models become more accessible and capable, technical moats around AI-powered ticket automation may erode. If workflow automation becomes a standard feature rather than a differentiator, competition could shift to pricing and integration breadth over AI sophistication.
SV006 Serval AI-native IT service management: The guide for modern teams In an AI-native IT service management platform, the entire architecture is built around automation from the start. Workflows rely on real code, often written in TypeScript, rather than simple drag-and-drop rules.
SV007 Serval 2026 Buyer's Guide: AI ITSM Systems That Deliver Immediate ROI Analyst benchmarks put AI-driven self-service at $2–$4 per ticket versus $15–$25 for live-agent resolution, and most teams see measurable returns within 90–180 days when they target high-volume requests first.
SV008 Serval Top AI-native ITSM tools in 2026 The Catalyst Agent runs the Workflow Builder. An IT admin describes a workflow in plain language. The Catalyst Agent generates TypeScript, which the admin can review and edit directly in the builder or via the Serval CLI in their local development environment.