数万亿美元仅用于客户服务？
5 分•作者: YihaoZhang•1 天前
我最近读了几篇关于人工智能公司盈利机会的文章，发现作为创始人或运营者，大致有六种赚钱方式。我不确定风险投资界的共识是否完全滞后于实际发展，所以提出这个问题。
以下是我最近看到的一些想法：
1. **AI整合收购（AI Roll-ups）**：这是一个非常热门，有时甚至被过度炒作的话题。它涉及收购那些尚未深度整合AI但急需人工服务的公司。例如，小城镇的会计师事务所、IT托管服务（有时是外包的）、法律服务（不一定是顶级律所，而是为现有客户提供服务的本地律所）以及保险业。这个模式之所以病毒式传播，是因为有人可能认为收购和优化公司比销售软件更好。如今的软件可能需要重新定位来建立护城河，而许多可以自动化的任务都围绕着服务业，尤其是在价值链的低端。风险投资公司也在寻找新的资产，因为传统的SaaS模式已不再能带来高投资回报率。
表面上看，这似乎说得通。但随后我想到，如果真是这样，我们还需要风险投资吗？风险投资公司的存在似乎有点滞后于当前的发展，其商业模式可能也无法很好地运作。风险投资的黄金时期是在移动和云计算时代。
2. **AI自动驾驶/AI原生服务公司（AI Autopilot / AI-Native Service Companies）**：这涉及开发AI自动驾驶系统，其中服务被视为软件，或者建立AI原生服务公司。公司正在探索保险经纪、会计或税务审计等领域，构建基于“行动系统”（system of action）的公司。这意味着集成SAP、Salesforce或ServiceNow等产品，让用户无需在20个不同的页面之间切换来管理采购、入职、期末结账、工单升级等。
3. **公司大脑（Company Brains）**：这条路径涉及将Slack、电子邮件、工单、会议和数据库整合到一个代理中，使其成为公司的“大脑”。这可能是组织重塑的一种方式，因为代理将能更好地理解公司。
4. **可验证工作（Verifiable Work）**：大家都知道开发可验证工作的公司，而编码是第一个用例。但自2024年我首次尝试使用Cursor以来，我没有看到比编码更具病毒式传播性的用例。这让我觉得公司和投资者正在试图找到下一个编码用例，但我们尚未找到。我们在合同红线修改、支持问题解决、质量保证或IT事件摘要等领域看到了尝试，公司也正在这些方面进行开发。
我的问题是，我们能否说自2022年以来投入AI的数万亿美元，都是为了提高效率和降低成本这个更大的目标？我知道公司有很多问题需要解决，但如果这是最大的用例，那么风险投资规模的回报在哪里？从我的角度来看，其中许多事情可以由私募股权公司来完成。一家成长型股权或私募股权公司可以利用杠杆收购，投资于这些用例。一家私募股权公司可以利用其投资组合公司收购大量旨在简化工作流程的AI业务。与目前被炒作的估值相比，回报可能会慢得多，也许3倍或4倍就已经是非常好的消息了。
我是否遗漏了什么重要的事情？这就是我在这里提问的原因。
顺便说一下，我不是专业人士，也不住在湾区，我目前在上海，所以可能有些信息我没有掌握。谢谢。
查看原文
I came across a couple of articles discussing the bigger opportunities for AI companies to make money. It turns out there are pretty much six different ways to make money if I'm a founder or an operator. I'm not sure if the consensus from the venture capital world is entirely lagging behind what's actually going on, so that's why I'm asking this question.<p>Here are a couple of ideas I saw recently:<p>1. AI Roll-ups: This is a very hot, sometimes overhyped topic. It involves buying companies that are less integrated into AI but heavily need human services. Examples include accounting firms in small towns, IT managed services (sometimes outsourced), legal services (not necessarily top law firms, but local ones helping existing customers), and insurance. The reason this has gone viral is that some might figure out that buying and streamlining companies can be a better choice than selling software. Software nowadays might need to redirect to build its moat, and many tasks that can be automated surround the service sector, especially at the lower end of the value chain. Venture capital companies are also looking for new assets because traditional SaaS models no longer present high ROI.<p><pre><code> On the surface, this makes sense. But then I thought, if that's the case, why do we need venture capital? The existence of venture capital businesses seems a bit behind what's currently going on, and the business model may not work that well. The best times for venture capital were during the mobile and cloud eras.
</code></pre>
2. AI Autopilot / AI-Native Service Companies: This involves working on AI autopilot, where everyone knows service as software, or building AI-native service companies. Companies are looking into areas like insurance brokerages, accounting, or tax audits, building companies based on a "system of action." This means integrating products like SAP, Salesforce, or ServiceNow so users don't need to use 20 different pages to manage procurement, onboarding, period closing, ticket escalation, etc.<p>3. Company Brains: This path involves integrating Slack, email, tickets, meetings, and databases together into an agent that can become our company's brain. This might be a way for organizations to restructure themselves, as agents would understand companies much better.<p>4. Verifiable Work: Everyone knows about working on companies that do verifiable work, and coding was the first use case. But since 2024, when I first tried using Cursor, I haven't seen another use case as viral as coding. This makes me think that companies and investors are trying to figure out the next coding use case, but we haven't found it yet. We see attempts in areas like contract red lines, support resolutions, QAs, or IT incident summaries, and companies are already working on these.<p>My question is, can we say the trillions of dollars invested into AI since 2022 are aimed at the bigger topic of improving efficiency and saving costs? I know companies have many problems to solve, but if this is the biggest use case, where is the venture-scale return? From my perspective, many of these things can be done by private equity companies. A growth equity or private equity firm could use leveraged buyouts and invest in these use cases. A private equity company could use its portfolio companies to acquire a large number of these AI businesses that aim to streamline workflows. The returns, compared to currently hyped valuations, might be much slower, perhaps 3x or 4x would be very good news.<p>Am I missing something important? That's why I'm asking here.<p>By the way, I'm not a professional and I don't live in the Bay Area; I'm currently based in Shanghai, so there might be information I haven't grasped. Thank you.