高速增长的初创公司如何利用 Claude Code 进行产品交付——从对十几家公司的访谈中提炼出的五项运营原则。How fast-growing startups use Claude Code to ship—five operating principles drawn from interviews with more than a dozen companies.PDF 文件下载https://cdn.prod.website-files.com/6889473510b50328dbb70ae6/6a85f3ce3623355a8a44ca36_Claude_Code_Guide_For_Startups_Final_web.pdfAI natives working at the frontier如果你想一窥工作的未来不妨问问初创公司他们今天的运营方式。我们正是这样做的。我们与十几家高速增长的初创公司进行了交流了解他们如何使用智能编码工具来构建产品并扩展公司规模。这些初创公司正在改变谁可以参与构建、什么应该被舍弃的规则以及如何在构建方式与构建内容之间创造飞轮效应。他们的产品交付速度堪比规模十倍于他们的组织。AI natives working at the frontierIf you want to take a peek at the future of work, ask startups how they are operating today. So we did.We spoke with more than a dozen fast-growing startups about how they use agentic coding tools to build products and scale their companies. These startups are changing the rules of who gets to build, what gets scrapped, and how to create a flywheel between how you build and what you build.And they are shipping like organizations ten times their size.在本指南中我们将深入探讨这些组织的独特部署方式学习他们为快速交付并保持竞争优势所遵循的规则。在此过程中我们还将开始探寻一个问题的答案如果一个组织从零开始用 Claude Code 构建其产品开发生命周期那会是什么样子In this guide, well dive into the unique deployments of these organizations to learn the rules they follow to ship fast and maintain their competitive advantage.In doing so well also start to glean an answer to the question: what would it look like if an organization built their product development lifecycle with Claude Code from the ground up?Everyone ships智能编码降低了准入门槛因此最理解问题的人就能交付第一版修复方案。Everyone shipsAgentic coding lowers the barrier to entry, so the person who understands the problem can ship the first version of the fix.智能编码降低了非技术员工构建产品的门槛。借助 Claude Code即使不精通编程语言也不熟悉 IDE 的使用你也能创建出功能完备的特性。Agentic coding lowers the barrier to entry for non-technical employees to build products. With Claude Code, you can create functional features without being fluent in a coding language or how to use an IDE.对初创公司创始人而言这显然大有裨益。一方面他们不像规模更大的竞争对手那样拥有充足人手因此必须“全员上阵”。但创始人看重的并不仅仅是人力规模——这些非技术团队成员同样带来了深厚的领域专长。For startup founders this has obvious advantages. For one, they dont have the headcount of their larger competitors so its all hands on deck. But its not just raw capacity that founders are after–these non-technical members of the team bring domain expertise as well.说“人人皆可交付”很适合发一条 LinkedIn 动态但在现实中如何落地是市场团队在审批拉取请求吗是法务团队在钻研如何定位那些时好时坏的测试吗我们得到的答案是分工依然存在。市场人员仍专注于市场开发人员仍专注于开发。但最关键的第一步——把一个想法变成可运行的原型也就是从 0 到 1 的过程——对所有人开放。我们还发现最高效的初创公司会建立相应机制让这些贡献系统化地发生而不是依赖偶然或个人的积极性。Saying everyone ships makes for a great LinkedIn post, but how does that work in reality? Is the marketing team approving pull requests? Is the legal team working through the intricacies of bisecting flaky tests?The answer we got is that there is still a division of labor. Marketers still focus on marketing and developers still focus on developing. But the all important first step of getting an idea to working prototype, of going from 0 to 1, is open to everyone.We also saw the most effective startups create mechanisms to make these contributions systemic rather than leaving it to chance or individual ambition.建立连接要求员工使用 AI 是一回事而让他们真正用上 Claude Code 以及所需的工具则是另一回事。Create connectionsIts one thing to create expectations for employees to use AI, its another to give them access to Claude Code and the tools they need.在 Crosby团队并没有把律师带到 Claude Code 面前而是把 Claude Code 带到律师面前——将其接入他们熟悉且每天都在使用的工具和操作系统。At Crosby, the team didnt bring lawyers to Claude Code, they brought Claude Code to the lawyers by connecting it to the tools and operating systems they were familiar with and worked in every day.站会展示在某个节点想法需要有机会被优先考虑以便组织资源能够帮助它们推向市场。这条路对产品经理来说是清晰的——毕竟那是他们的本职工作——但对非技术员工来说就没那么清晰了。Clay 会举办季度评审会会上对原型进行考量并允许其进入正式路线图。正是通过这种方式Clay 的一位市场团队成员构建了一个自主智能体它会访问你的网站、填写你的线索收集表单、统计响应耗时、评估体验并生成一份绩效报告。Omni 为 Claude 生成的原型开设了专门的 Slack 频道包括资深技术人员在内的所有人都可以贡献内容。他们还践行了「人人皆可交付」的推论那就是「人人都与客户交流」。Standup showcasesAt some point, ideas need to be given the opportunity to be prioritized so that organizational resources can help bring them to market. That road is clear for product managers—its their job after all—but not as clear for non-technical employees.Clay creates quarterly reviews where prototypes are considered and can enter the formal roadmap. This is how a go-to-market team member at Clay built an autonomous agent that visits your websites, fills out your lead-capture forms, times how long it takes to respond, rates the experience, and generates a performance report.Omni has a dedicated Slack channel for Claude generated prototypes with contributions from everyone including senior technical staff. They also practice the corollary of everyone ships, which is everyone talks with customers.分享技能「人人皆可交付」与「零敲碎打」之间的界限可能很微妙。无论原型出自谁手它们仍需被整合进一个整体协调的产品中。这正是技能skills发挥作用的地方——这些可复用的指令文件承载着团队的标准与上下文能够确保即便流程日益民主化开发工作依然保持对齐。「团队中的任何人都可以借助 Claude Code以我们的设计系统为参考起草产品组件、营销素材或演示文稿。而涉及产品的 AI 则必须跨越高得多的门槛Claude Code 帮助我们以更高的精度达到这一标准。」Heidi 公司的 Thomas Kelly 博士表示。他们还能让新入职的开发者以及非技术员工快速上手并投入工作。Share skillsThe line between everyone ships and piecemeal can be a thin one. Feature prototypes, whoever they come from, still need to be integrated into a product that feels like a cohesive whole. This is where skills, reusable instruction files that encode your teams standards and context, can help ensure development stays aligned even as the process becomes increasingly democratized.Anyone on the team can draft product components, marketing collateral or deck material from Claude Code using our design system as reference. AI that touches the product must clear a much higher bar, which Claude Code helps us meet with more precision, said Dr. Thomas Kelly, Heidi.They can also get new developers and non-technical employees onboarded and up and running quickly.自动化繁琐工作智能体负责产品生命周期中机械性的那 80% 工作从而让工程师把时间花在真正需要判断力的场景上。Automate the tediumAgents own the mechanical 80% of the lifecycle so engineers spend their time on the cases that actually need judgment.自工业革命以来所有公司都一直在寻求通过技术提升效率但这些初创公司之所以脱颖而出是因为他们采用技术的速度之快、程度之深。这些创始人相信AI 是他们使命中不可或缺的组成部分。许多人明确表示智能体负责机械性的那 80% 工作从而让工程师把时间花在真正需要判断力的场景上。All companies have sought to gain efficiencies through technology since the dawn of the industrial revolution, but these startups separated themselves by the speed and depth of their adoption.These founders believe AI is an essential component of their mission. Many are explicit that agents own the mechanical 80% so engineers spend their time on the cases that actually need judgment.具体来说我们观察到这些初创公司将 AI 更紧密地集成到软件开发生命周期的各个阶段同时构建了更多专用智能体用于端到端地处理重复性任务。下面我们来看这两方面的几个例子。Specifically, we saw AI more tightly integrated across their SDLC stages than others as well as more purpose built agents designed to take recurring tasks end-to-end. Lets look at a couple examples of both.AI-native SDLCs在这些受访初创公司中许多都建立了加速团队融入智能编码流程的机制。例如在 EmergentMukund 告诉我们「新员工入职第一天只需让 Claude 指向正确的 markdown 文件就能完成整个开发环境的搭建。如果 Claude 在入职过程中遇到任何损坏或过时的内容它会直接更新该文件。」AI-native SDLCsMany of these featured startups have implemented means of accelerating their teams onboarding into their agentic coding processes. For example, at Emergent, Mukund told us, on day one, a new hire bootstraps their entire dev setup by pointing Claude at the right markdown file. If Claude hits anything broken or out of date during onboarding, it updates that file.Tip:Code Review (research preview) is a managed multi-agent service in Claude Code. It runs an automated review pass on PRs in the repos you enable. You can manually fix the finding and push, or close the loop by commentingClaudeon the finding (if youve set up and configured GitHub Actions).https://code.claude.com/docs/en/code-review这些工程师需要快速完成入职因为这些团队的交付速度很快。These engineers need to be onboarded quickly because these teams ship fast.在这些组织中Claude Code 不仅帮助生成代码还负责审查代码。Heidi 公司的 Kelly 博士表示「我们会针对经过验证的技术与合规框架运行自动化代码审查在代码交付前标记关键问题并将修改建议分发给相应的审查人员。」其中一些组织还构建了用于代码审查、测试和 CI 的自定义智能体。这些初创公司非常重视构建反馈闭环而不仅仅是部署代码。Translucent 创始人 Jack 表示「我最喜欢的智能体是『Translucent 代码审查器』它会全面铺开审查一项改动从多个角度进行审视并像我们资深工程师那样综合结果但速度比任何单个人都快。」Kareem 表示Clay「构建了一个智能体负责处理……缺陷分类从初步筛选到为修复提出代码修改建议」。At these organizations, Claude Code not only helps generate code, but reviews it too. We run automated code reviews against our vetted technical and compliance frameworks, flagging critical issues and routing suggested changes to the right reviewers before anything ships, said Dr. Kelly of Heidi.Some of these organizations have also built custom agents for code review, testing, and CI. These startups have placed considerable attention on building loops vs just deploying code.My favorite [agent] is the Translucent code reviewer, which fans out across a change, reviews it from multiple angles, and synthesizes the results the way one of our senior engineers would but faster than any one person could, said Translucent founder Jack.Clay ...built an agent that handles…bug triage, from first pass to suggesting code changes for fixes, said Kareem.用智能体加速流程另一个反复出现的模式是这些初创公司不仅利用 Claude Code 中的智能体循环来加速开发工作还创建智能体来加速那些重复且往往繁琐的流程。这些往往是例行工作以便将更多注意力集中在他们的竞争优势、客户关系和营收增长上。我们观察到被 Claude 加速的最常见流程之一是自助式数据分析。这些公司几乎都建立了相应的流程以便利用最新数据包括非结构化数据快速做出决策而这些数据正是初创公司生存中至关重要的快速转向所依赖的燃料。例如Clay 构建了一个内部数据分析智能体而 Heidi 则使用 Claude Code 对客户和临床医生的反馈以及使用数据进行分类以发现对产品洞察至关重要的信号。ClickHouse 和 Omni 都推出了内置此类 AI 数据分析能力的产品且全部由 Claude 提供支持。其他例子还包括使用子智能体汇总数千份法律文件Crosby、扫描理赔数据以标记各站点之间的异常Commure以及持续挖掘医院财务数据捕捉任何分析师团队都无法及时发现的预警信号Translucent。Accelerating processes with agentsAnother consistent pattern was that these startups were not only using agentic loops in Claude Code to accelerate their development efforts, but they were also creating agents to accelerate recurring and often tedious processes.This was often routine work so that more attention could be focused on their competitive advantage, customer relationships, and on top-line growth. One of the most common processes we saw accelerated by Claude was self-service data analytics.Nearly every one of these companies had some process in place so they could make quick decisions with fresh data, including unstructured data, that fuels the pivoting so essential in the life of a startup.For example, Clay built an internal analytics agent and Heidi uses Claude Code to categorize customer and clinician feedback alongside usage data to surface signals that matter for product insights.Both ClickHouse and Omni ship products that package this type of AI data analysis within them, all powered by Claude.Other examples include summarizing thousands of legal documents with subagents (Crosby), sweeping claims data to flag anomalies across sites (Commure), and continuously mining hospital financial data for warning signs no analyst team could catch in time (Translucent).信任但需验证除非你拥有可靠的手段来监控和验证结果否则你无法实现流程自动化。Trust, but verifyYou cant automate a process unless you have a reliable means of monitoring and verifying the outcome.这条规则是第二条原则「自动化繁琐工作」的必要推论。除非你拥有可靠的手段来监控和验证结果否则你无法实现流程自动化。This rule is the necessary corollary to Rule 2: Automate the tedium. You cant automate a process, unless you have a reliable means of monitoring and verifying the outcome.需要说明的是这些初创公司并没有让智能体直接合并到主干分支然后听天由命。其中许多公司身处高度受监管的行业需要强有力的治理框架。Cainex 就是一个极具代表性的例子它将智能体与确定性检查相结合用于读取病历并生成指导医院计费的编码。To be clear, none of these startups are having agents merge to main and hoping for the best. Many of them operate in highly regulated industries and require strong governance frameworks. Cainex is a particularly illustrative example of combining agents with deterministic checks to read medical records and generate codes that direct hospital billing.需要说明的是这些初创公司并没有让智能体直接合并到主干分支然后听天由命。其中许多公司身处高度受监管的行业需要强有力的治理框架。Cainex 就是一个极具代表性的例子它将智能体与确定性检查相结合用于读取病历并生成指导医院计费的编码。To be clear, none of these startups are having agents merge to main and hoping for the best. Many of them operate in highly regulated industries and require strong governance frameworks. Cainex is a particularly illustrative example of combining agents with deterministic checks to read medical records and generate codes that direct hospital billing.「这就是 Claude Code 为我们运行的闭环。我们用智能体处理一批数据然后由审计人员在公司内部应用中审查输出结果。他们看到的不仅仅是编码还有模型的推理过程并会对两者都进行评论……所有内容都有版本记录且全程可审计。」他表示。「随后 Claude Code 接手。它直接从数据库中读取原始预测以及每一条修正和评论。每条修正都会按所涉及的编码类型打上标签因此 Claude Code 能判断出这是诊断问题、手术问题还是其他类别并直接定位到管理该类编码的对应指南。接着它会找出智能体指令中导致错误的部分并加以修订如果遇到真正的新情况则会编写新的指南。每一次修改都基于一套有版本管理的指令并针对此前失败的记录进行测试。我们遵循的原则是修正原则本身而不是修正单个案例。」他继续说道。「然后是回测。一条记录可能存在多个可接受的编码因此不能简单地做字符串匹配。这项检查会将语义匹配与我们的既有标准集相结合并借助一个评判机制来问『这是真正的错误还是只是另一条同样有效的路径』Claude Code 还会在此基础上加入自己的对比分析。它会在一个黄金数据集以及随机样本上运行候选修改并在任何内容上线前暴露潜在的回归问题。最终返回的是一份精简清单建议的修改、无法解决的记录以及它希望得到解答的问题。工程师因此能把时间花在真正棘手的案例上而不是那机械性的 80% 工作。」他说。创始人可以从这个医疗计费领域的特定工作流中提炼出许多具有普适性的经验。例如Cainex 会安排领域专家定期审查并引导 Claude 的推理过程确保这些指导成为自我改进闭环的一部分。不过这些专家并不是逐条修正案例他们的指导被用于一个自我改进的循环。正如 Uriah 所说「修正原则本身而不是修正单个案例。」Heres the loop Claude Code runs for us. We process a batch with an agent, and our auditors review the output in an internal app. They dont just see the codes. They see the models reasoning, and they comment on both….Everything is versioned and auditable, he said.Then Claude Code takes over. It reads the original predictions, along with every correction and comment, straight from the database. Each correction is tagged by the kind of code involved, so Claude Code knows whether its looking at a diagnosis issue, a procedure issue, or another category, and it can go straight to the guidance that governs that specific kind of coding.From there, it finds the part of the agents instructions that produced the mistake and revises it, or writes new guidance when the case is genuinely new. Every change is made against a versioned set of instructions and tested against the records that failed. The rule we enforce: fix the principle, not the example, he continued.Then the back-test. A record can have more than one acceptable coding, so its not a string match. The check combines semantic matching against our accepted sets with a judge that asks, Is this a real error or just a different valid path, and Claude Code adds its own comparisons on top.It runs the candidate change across a golden set plus random samples and surfaces any regressions before anything ships. What comes back is a short list: suggested edits, the records it couldnt resolve, and the questions it wants answered. Engineers spend their time on genuinely hard cases rather than the mechanical 80%, he said.There are many generalized takeaways that founders can glean from this healthcare billing specific workflow.For example, Cainex uses subject matter experts to routinely review and guide Claudes reasoning, and ensure that guidance becomes part of a self-improvement loop. However, those experts arent there to fix example by example, their guidance is used as part of a self-improvement loop. As Uriah puts it fix the principle, not the example.另一个值得借鉴的经验是这些团队在维护一套高质量评估「黄金数据集」上投入了大量心力——也就是一组经过验证的问答对用于检验智能体的准确性。每家初创公司都应当针对自己的关键使用场景维护多套评估集并定期更新从而防止模型漂移也为评估未来的模型做好准备。The other takeaway is the diligence placed on maintaining a strong evaluation golden set, or group of verified question answer pairs the team uses to verify the agents accuracy. Every startup should maintain multiple sets of evals for their key use cases, and update them regularly, so they can prevent drift and evaluate future models.Uriah 最后还指出这一过程本身也需要投入不少功夫。「一开始并没有这么顺利。我们的第一版出现了过拟合——它通过把具体案例编码进去来『修复』问题结果我们只是在不断堆积补丁而不是让系统变得更聪明。后来我们改变了思路强制推行通用原则并严格限制单次改动中能纳入的具体细节数量。」The final point Uriah makes is that this process can take some work. It didnt start this clean. Our first version overfitted. It would fix things by encoding the specific case, and we were accumulating patches instead of getting smarter. We changed the approach to force general principles and to cap how many specifics can enter a change at all.Build for rebuilding模型能力在这些团队脚下不断变化因此几乎没有什么是被当作永久不变的。Build for rebuilding模型能力在这些团队脚下不断变化因此几乎没有什么是被当作永久不变的。Many of these AI-native startups are in a state of constant reinvention.AI is often at the heart of what they are building as well as how they are building it. Since model capability continuously evolves, groundbreaking features and critical scaffolding were discarded the minute they became sunk costs. Many of these organizations saw this constant rebuilding as part of their competitive advantage.What we do at Clay is you build it and then you build it again and then you build it again. And then the fourth time you build it, you know everything thats needed and you get it right. And so we dont necessarily throw away things. We just rebuild it: and this time with more clarity, said Kareem.A rebuild isnt done when the new path ships. Its done when the old path is gone. Teardown always lost the prioritization fight before: its tedious and it ships no features, said Commure co-founder Tanay. Now one of Commures engineers just invokes a Claude skill to the tune of for every feature flag already released to everyone, open a PR removing it and the associated code, then the engineer reviews what comes back. Migrations that used to eat a lot of dev cycles are now a plan and a fan out, done in a couple of hours.Each linked worktree is an ordinary directory with its own checked-out branch; all three share the single .git object store inside acme-web.Kareem also described part of Clays moat as the ability to constantly rebuild, evolve, and create self-improvement loops.I think the moat for any company right now is that it needs to be self-improving. So Clay is a self-learning revenue engine. So the more you use this, the more we know who your best customers are, what should you say, whats worked, what hasnt and thats changing over time, he said. The race is really, whoever can get to the distribution fastest… so you can help each [customer] so that you can self-improve.At a May 2026 Code with Claude event, Niko Grupen, Harveys Head of Applied AI spoke about how each new wave of model capabilities — emergent reasoning, agentic automation, planning and orchestration — required a full re-architecture of the platform.Prototype, dogfood, productionize借助 AI 进行构建帮助这些初创公司打造出颠覆性的 AI 产品——这正是他们整个流程的核心飞轮。许多这样的初创公司其开发流程的核心都有一个关键飞轮借助 AI 进行构建帮助他们打造出颠覆性的 AI 产品。当开发者精进他们的智能编码实践时他们对模型能力会有更深刻的理解也能洞察前沿的 harness 设计如何演进。随后他们可以将这些灵感应用到自己的智能体和产品中。「我们从 Anthropic 的『文件与嵌入』方法中汲取了灵感这让我们在自己的产品中坚持化繁为简避免了 RAG 管线可能带来的大量复杂性。」Omni 的 Chris 表示。「我们还看到 Claude Code 的 harness 如何让用户并行处理事务并将其中一些理念融入到了我们自己的界面设计中。」这也有助于他们时刻关注自身产品的性能表现。「因为我们的应用构建器在后台也使用了 Anthropic 模型所以一旦在产品上发现任何异常行为……我们就能通过 Claude Code 在本地快速调试判断是模型行为问题还是 harness 的问题。这极大地改善了我们的故障排查周期。」Emergent 的 Mukund 表示。我们反复听到的模式是先用 Claude Code 构建一个内部智能体在内部使用dogfood然后根据反馈将其升级为面向客户的产品通常借助 Claude API、SDK 或 Claude Managed Agents 来实现。「我们在产品中构建了自己的 AI 智能体供团队直接使用包括 SQL 控制台中的智能体和 AI SRE。我们使用 Claude Code 来构建并迭代这些智能体本身。为客户 AI 体验提供支持的工具有一部分正是用 AI 构建的。」ClickHouse 的 Alexey 表示。---https://claude.com/blog/claude-code-guide-for-startups