如何构建团队智能体

2026年是智能体的一年。

2026 has been the year of agents.

从年初的OpenClot，到现在的Muse、GroqBot和Instinct等平台，这些平台让人们真正利用这些极其强大的自主工具，越来越多的工作由它们完成。

From OpenClot at the beginning of the year, to now platforms like Muse and GroqBot and Instinct that are getting people to actually take advantage of these incredibly powerful autonomous tools that are getting increasingly large portions of their work done for them.

我们确实已经从智能体是下一个大热门，发展到它们已经无处不在。

We really have gone from agents being the next big thing to just being here.

问题是，我们的工作并不是独自完成的。

The problem is, our work isn't just done alone.

我们通常与他人组成团队工作，但到目前为止，大多数智能体都是单打独斗，只覆盖我们独自完成的那部分工作。

We tend to work in teams with other people, and yet up till now, most agents have been solo affairs, only covering the portion of our work that we do on our own.

我认为这种情况正在改变，我称之为多人AI或共享或团队智能体的趋势。

I think that is shifting now, a trend which I've talked about as multiplayer AI or shared or team agents.

但构建一个团队智能体到底意味着什么？

But what does it mean to even build a team agent?

需要考虑哪些方面？

What are the types of considerations that go into it?

那和自己建一个代理到底有多大区别呢？

And how different is it really than just building an agent for yourself?

这些就是我在《AI每日简报》的“操作者剪辑”版块里和Nufar Gaspar讨论的问题。

Those are the questions that I get into with Nufar Gaspar on this Operator's Cut edition of the AI Daily Brief.

《AI每日简报》是一个每日更新的播客和视频节目，内容涵盖AI领域最重要的新闻和讨论。

The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI.

好了，朋友们，在我们深入之前，先快速说几个通知。

All right, friends, quick announcements before we dive in.

首先，感谢今天的赞助商：毕马威、Blitzy、Harbor和Hyperagent。

First of all, thank you to today's sponsors, KPMG, Blitzy, Harbor, and Hyperagent.

想收听无广告版本，请访问patreon.com/aidailybrief，或者你也可以在Apple Podcasts上订阅。

To get an ad-free version of the show, go to patreon.com/aidailybrief, or you can subscribe on Apple Podcasts.

想了解如何赞助这个节目，请发送邮件到sponsors@aidailybrief.ai。

And to learn more about sponsoring the show, send us a note at sponsors@aidailybrief.ai.

在我们开始之前，还有几点要提一下。

Just a couple other notes before we get in.

显然，这是一集预先录制的节目。

Obviously, this is a pre-recorded episode.

今天有很多事情在酝酿中。

There are a bunch of things cooking today.

我们会有很多话题要聊，所以明天我们会恢复正常格式。

We will have a lot to talk about, so we will be back with our normal format tomorrow.

我还想分享一些即将到来的机会。

I also wanted to share a couple of upcoming opportunities.

首先，这个周四，10月1日，我们有一个免费的线上研讨会，主题是如何构建你自己的个人AI基准测试。

First of all, this Thursday, October 1st, we have a free live webinar all about building your personal AI benchmark.

核心思路是，当你拿到一个新模型，比如Opus 5.5、Sonnet 5.5、Gemini 4，或者任何接下来发布的模型，这个研讨会会帮你制定一套自己的标准基准，让你更清楚地了解这个模型在你的工作流程中能发挥什么作用。

The whole idea is that when you get a new model like Opus 5.5 or Sonnet 5.5 or Gemini 4 or whatever model comes next, this will help you put together your own standard benchmark to better understand where that model is going to fit into your own process.

这个活动完全免费，而且只要你注册了，即使无法参加，之后也会收到所有资料。

That is completely free, and if you register, you will get all the materials after, even if you can't attend.

再次提醒，活动就在这个周四，10月1日。

Again, that is coming up this Thursday, October 1st.

说到培训，如果你想更深入一些，我们下一期的超级智能高管AI和智能体培训项目就要开始了。

Now, speaking of training, if you want to go a little bit deeper, the next cohort of our super intelligent executive AI and agent training programs is coming up.

高管智能体领导力项目是教你如何为真实的业务需求构建AI智能体，并制定一套方案，以便在你的组织中安全地扩展它们。

The Executive Agent Leadership Program is where you learn how to build AI agents for real business needs, as well as building a playbook to scale them safely across your organization.

如果你觉得在进入那个项目之前需要一些基础，你也可以参加高管补课项目。

And if you feel you need a little bit more background before you get into that, you can also do the executive catch-up program.

下一期智能体领导力项目将于10月5日开始，而下一期高管补课项目则在一周后的10月12日开始。

The next agent leadership cohort starts on October 5th, while the next executive catch-up program starts a week later on October 12th.

好了，这些事都交代完了，我们来谈谈如何构建团队智能体。

All right, with all that out of the way, let's talk about how to build team agents.

好的，Nufar，欢迎回到节目。

All right, Nufar, welcome back to the show.

今天我们有一个“操作员剪辑”版。

We got an Operator's Cut today.

是的。

Yes.

很高兴再次来到这里。

Happy to be here again.

这个想法的起源，是我们秋天前讨论今年秋季免费自学培训课程时的一些对话。

This one has its genesis in some conversations we were having as we were coming up into the fall around what we wanted to do with this fall's edition of a free self-directed training program.

我们聊了很多关于多人AI的概念，以及从人们独自构建和使用智能体，转向预测我们会开始看到——我想我们已经开始看到一些早期迹象——更多智能体存在于人们的共享工作空间之间。

And we were talking a lot about this idea of multiplayer AI and a shifting pattern from people just building solo agents that they were using themselves to a prediction that we're going to start to see, and I guess we're starting to see early evidence of more agents that live in between people's shared workspace.

这其实只是顺应了人们自然的工作方式。

And this kind of just follows the natural way that people work.

很多工作是你独自完成的，但也有大量工作是在与他人的交汇处完成的，而团队智能体就可以存在于这些交汇处。

A lot of your work is done individually, but then lots and lots is also done at the intersection with other people, and that's where team agents can live.

在我们构建那个课程（现在可以在Multiplayer AI上找到）并更广泛地思考这个概念时，我们反复想到的一点是，这个领域还很新，所以真正构建一个团队智能体的方法并不一定那么显而易见。

And as we were building out that course that's available right now at Multiplayer AI and just thinking about this concept more broadly, one of the things that we kept coming back to was that this is nascent enough that what it means to actually build a team agent won't necessarily be super obvious.

所以今天这期Operator's Cut的目标，就是帮助大家真正思考如何构建团队智能体，理解团队智能体是什么样子，理解它们与人们可能已经构建的智能体类型有何不同，或者我认为我们可能会论证的是，它们与那些智能体有多么相似，以及它们未来可以走向何方。

And so the goal of today's Operator's Cut is to help actually think through how to build team agents, to understand what team agents look like, to understand in what ways they are different from, or I think probably what we'll argue here, similar to the types of agents that people might have already built and where they can go from here.

所以非常高兴你回来，也很期待深入探讨这个话题。

So super excited to have you back and excited to dive in here.

太棒了。

Amazing.

所以我要扩展一下你的定义，把它们叫做团队智能体，这个概念就是整个团队都能一起使用的智能体，不管是因为工作发生在它们之间，还是因为它们可以被团队成员共享。

So I'm going to broaden your definition, and I'm going to call them team agents, and the concept is teams that your entire team can work with, whether it's because work happened between them or just because they are something that can be shared across team members.

简单来说，有些智能体应该保持个人私有，而其他的则应该成为团队级别的智能体。

And kind of the short version is that some agents should stay yours and private, while others should become the teams-level agents.

而那些成为团队级别的智能体，需要有意地做出一些决定。

And the ones that do become the teams need a few decisions made on purpose.

其中一些，就像你说的，和任何好的智能体配置都有重叠，而另一些则更独特，或者至少更有针对性，这就是我们要讨论的内容。

Some of them, as you said, overlap with any good agent configuration, and some of them are more unique or at least more intentional, and that's what we'll walk through.

我想先讲两个故事来激励大家，你们很可能在自己的公司或生态系统中遇到过类似的情况。

And I wanted to start as a means of motivation to give you two stories that you will probably recognize from your company or your ecosystem.

第一个是关于一个人的故事，我合作的每家公司，至少都有一个这样的人。

So the first is about a person that everybody that I work with, every company that I work with has at least one like that.

这个人，他们非常清楚定价例外是怎么运作的，或者数据法案是怎么回事，或者我们最大客户的配置三年前是怎么设置的。

And this person, they really know how their pricing exception work or what the data act is all about, or how our biggest customer setup was configured three years ago.

当他们忙得不可开交时，工作就得等他们，因为只有他们知道怎么做。

And when they're swamped, then work has to wait for them because they're the only one who knows.

他们休假时，还是会有人打电话给他们；他们离开时，公司的一部分也跟着他们走了。

And when they're on vacation, someone still calls them, and when they leave, a piece of the company leaves with them.

所以在我合作的一家公司里，每个领域都有这样一个关键人物，结果没人能安心休假，总会被同事打电话打扰，这显然不是理想的状态。

So in one of the companies that I work with, they had, I think, a person like that for each and every domain, so no one gets to take vacation without getting a call from their peers, and obviously, that's not a desired state.

第二种情况是，有些工作没有人真正负责。

The second scenario is work that nobody fully owns.

比如，客户从销售转到市场，再到客户成功，销售在谈合同时做了承诺，而市场部却用略有不同的信息在推广。

So a customer can move from sales to marketing to customer success, and then sales made them a promise during the deal conversation, and marketing is running a campaign with slightly different messaging.

然后客户成功在续约前几周才发现这些承诺，每个团队都有自己的部分，但没有人掌握全貌，因为工作就卡在团队之间，正如你所说。

And then customer success finds out these promises were made to them a few weeks before the renewal, and each team has their own piece, and nobody has the whole picture because the work sits between them, so to your point.

即使他们在每个环节都用了AI，这些智能体之间也不互相沟通，很多时候反而让问题更严重。

And even if they are using AI in each step of the process, these agents don't talk to each other and only worsen the problem in many cases.

所以，为整个团队打造的智能体可以同时解决这两个问题，当然，这两个问题性质很不一样。

So an agent that is built for the whole team can help with both of these problems, and they are, of course, quite different problems.

所以我觉得，现在这件事之所以重要，还有一个原因，即使你觉得这有点超出你的理解范围，你也应该现在就开始关注，那就是我们在那些最积极采用AI的公司里开始看到的一种模式，基本上分三步走。

So There is another reason, I think, why this matters right now and why you should pay attention now, even if you feel a little bit like this is above your head, and that's the pattern that I think that we're starting to see across the most AI-forward companies, and it goes in basically three steps.

第一步是每个人都构建自己的智能体，并且对生产力的提升感到满意。

The step one is that everybody builds their own agents, and they're happy with the productivity boost.

但当这样的智能体数量足够多时，我们就会陷入一种智能体泛滥的状态。

Only with enough of those running around, we kind of get into an agent sprawl.

很多智能体在做重复的工作，每个都由一个人维护，每个对公司都有略微不同的理解，而且一旦维护者失去兴趣或离开公司，这些智能体就不再有用。

Lots of agent doing overlapping work, each maintained by one person, each with slightly different picture of the company, and each one stops being useful the day that the owner loses interest or leaves the company.

然后，在那些最积极采用AI的公司里，你会看到他们开始把一些智能体合并成团队级别的智能体。

And then what you see in the most AI-forward company, they started to merge some of those agents into team-level agents.

这些通常数量会少很多，但范围更广，每个都有明确的负责人，被很多人使用，并且随着团队了解它们该做什么、不该做什么，会不断优化。

Those will typically be much fewer agents, much broader in their scope, and each with a named owner and used by many people, and ideally refined over time as the team learns what they should and shouldn't do.

而且这种情况正在公开发生。

And this is happening very publicly.

已经有很多公司在谈论这件事了。

There are many companies already talking about it.

我记得你提到过Every公司的经验。

I think you mentioned Every's experience.

他们年初给每个员工都配了一个智能体，但到了五月，他们就转向了共享的团队智能体。

They started by giving every employee an agent early in the year, and then by May, they have moved to shared team agents.

Sierra把他们的多个专业智能体合并成了一个。

Sierra merged many of their agent specialists into one.

Shopify的内部智能体。

Shopify's internal agent.

所以我们在很多公开演讲的公司里都看到了这种情况，到处都是。

So we see a lot of these in very public-speaking companies all over the place.

我看到的多数团队可能还处在第一步或第二步，但我觉得我们所有人都应该看看这些AI前沿的公司，理解怎么进入第三步，并且怎么做好，这就是今天要讲的全部内容。

And most teams that I see are probably either in step one or two, but I think that it's very important for all of us to look at these AI-forward companies and understand how to get to number three and how to do it properly, and that's the entire purpose of today.

为了确保我们说的是同一件事，因为同一个东西有很多不同的叫法。

So to make sure that we are talking about the same thing, because there are multiple names to basically the same thing.

有些人，包括你自己，把它叫做多人AI。

Some people, including yourself, call it Multiplayer AI.

你可能也听说过共享代理。

You probably also heard shared agents.

有些人把它们称为AI队友，甚至有时也会把公司大脑混进来，或者用来指代那些在公司范围内共享知识使用的代理。

Some people refer to them as AI teammates, and even company brain is sometimes thrown into the mix or interchangeably used to mean agents being used with shared knowledge across the company.

在整期节目中，我会把它们称为团队代理，我的意思是，我们有一个代理，很多人可以跟它对话，它拥有共享的知识、共享的记忆，以及统一的配置。

I'm going to refer to them throughout the episode as Team Agents, and what I mean by that is we have one agent that many people talk to with shared knowledge, shared memory, and one configuration.

指令、技能、访问权限和所有者，这些都是共享的。

The instructions, the skills, the access, and the owner, they are all shared.

当你听到我这么说的时候，你可能会想，我们不是已经共享技能了吗？

And you might be thinking when you hear me saying that we already share skills, right?

大多数公司都有一个很棒的技能库，或者正在建设技能库，这很棒，这是公司工作标准化的好方式。

Most companies have an amazing skill library or working on a skill library, and that's awesome, and that's great standardization of how you do the work in the company.

但技能是针对特定任务的操作手册，而团队代理则是你的整个团队用来处理各种不同任务的工具，无论是临时任务还是重复性工作。

But a skill is a playbook for a specific task, where a team agent is something that your whole team works with on diverse set of tasks, ad hoc as well as repeated stuff.

它承载着团队的知识，记住学到的东西，并且如果你们有团队级别的技能，它也会使用这些技能。

It does carry the team knowledge, remembers what it learns, and using the team level skills if you have them.

当然，这些也可以接入技能市场等等。

Those, of course, can also tap into skills marketplaces and so on.

所以，技能库可能只是其中的一个组成部分，但它们并不是一回事。

So a skill library perhaps is one of the ingredients, but they are not one and the same.

今天，我希望大家思考一下，如何以及何时开始构建你的下一个团队智能体。

And I want you to today think about how and when to start building your next team agent.

另外，关于范围再补充一点，因为现在大家对各种个人智能体都非常兴奋。

And one more note on scope, because there is a lot of excitement right now about all of the personal agents.

我指的是Muse、Instinct以及其他同类产品。

I'm talking about Muse and Instinct and some of the other in this category.

那些是用于家庭或私人生活的。

Those are for home or private life.

今天，我们重点讨论工作场景。

Today, the focus is going to be on work.

所以，这就是关于范围需要明确的一点。

So that's one thing to make sure that it's clear about the scope.

我们说的是你为工作而构建的那些智能体。

We're talking about agents that you build for your job.

毕马威和德克萨斯大学奥斯汀分校的一项新研究发现，当人们与人工智能合作时，相似的技能并不能保证相似的结果。

A new study from KPMG and the University of Texas at Austin found that when people work with AI, similar skills don't guarantee similar outcomes.

研究人员研究了500多名早期职业专业人士，发现表现最好的人通过指导、评估和完善人工智能的输出，持续放大了人工智能的价值。

Researchers studied more than 500 early career professionals and found that the best performers consistently amplified the value of AI by guiding, evaluating, and refining its outputs.

这些被称为“人工智能放大器”的顶尖表现者，并不是仅仅由他们知道什么来定义的，而是由他们如何与人工智能合作来定义的。

These top performers, called AI Amplifiers, weren't defined by what they knew alone, but by how they worked with AI.

想了解更多关于人工智能放大器与其他人的区别，请访问kpmg.com/us/aiamplifiers。

Learn more about what separates AI amplifiers from everyone else at kpmg.com/us/aiamplifiers.

Blitzy在编写代码之前会深入了解你的代码库。

Blitzy deeply understands your code base before it writes code.

这是第一个体现价值的地方：人工智能时代的安全。

Here's the first place that pays off, security in the age of AI.

漏洞并不是孤立存在的。

Vulnerabilities don't live in isolation.

它们藏在数百万行相互关联的代码里，修好一个地方，可能悄悄弄坏另外三个。

They live buried inside millions of lines of interconnected code, where patching one thing quietly breaks three others.

所以表面扫描根本没用。

That's why surface level scans fail.

Blitzy从整个应用的知识图谱出发，识别并暴露整个系统中的CVE，主动推荐补丁，还能直接执行PR。

Blitzy starts from its knowledge graph of your entire application, identifies and surfaces CVEs across the full estate, proactively recommends patches, and can execute the PR.

每个修复都基于系统之间的连接关系，并经过验证，确保不会引入新的问题。

Each fix is grounded in how your systems connect and validates so nothing new breaks.

知识图谱还会动态更新，让你始终领先于不断加速的威胁环境。

And the knowledge graph dynamically updates, keeping you ahead of an ever-accelerating threat landscape.

有个Blitzy客户在四天内解决了6个核心微服务中的21个活跃CVE。

One Blitzy customer resolved 21 active CVEs across 6 core microservices in four days.

零编译错误，所有验证扫描都干净，原本计划几个月的活不到一周就搞定了。

Zero compile errors, every validation scan clean, months of planned work fixed in less than a week.

安全修复基于真实的架构上下文，以计算的速度完成。

Security remediation grounded in real architectural context at the speed of compute.

去 blitzy.com 加固你的代码库。

Harden your code base at blitzy.com.

网址是 B-L-I-T-Z-Y 点 com。

That's B-L-I-T-Z-Y dot com.

每期节目，我都会聊 OpenAI、Anthropic、SpaceX AI、谷歌和 Meta 之间的竞争。

Every episode, I talk about the competition between OpenAI, Anthropic, SpaceX AI, Google, and Meta.

如果你听了一段时间，可能已经有了自己偏爱的公司。

And if you've been listening for a while, you might have a favorite.

也许你觉得 OpenAI 和 Anthropic 能保持领先，或者认为 Meta 的开源策略能胜出。

Maybe you think OpenAI and Anthropic can stay ahead, or perhaps Meta's open source strategy can win out.

不管你怎么看，每个 AI 实验室都创造了不同的投资机会。

Whatever your view, every AI lab creates a different investment opportunity.

Harbor Capital Advisors 的 AI 实验室生态系统 ETF 系列，让你能投资于你相信的 AI 实验室背后的生态系统。

Harbor Capital Advisors AI Lab Ecosystem ETF Suite lets you invest in the ecosystem behind the AI lab you believe in.

在你投资的平台上搜索 Harbor AI 实验室生态系统 ETF，或者在 X 上关注 @HarborCapital 了解更多。

Search Harbor AI Lab Ecosystem ETFs wherever you invest or follow @HarborCapital on X to learn more.

请访问harborcapital.com查看招股说明书，其中包含投资目标、风险、费用、开支及其他重要信息。

Visit harborcapital.com for a prospectus containing investment objectives, risks, fees, expenses, and other important information.

投资前请仔细阅读并考虑这些内容。

Read and consider it carefully before investing.

风险包括本金损失以及与人工智能相关的风险。

Risks include principal loss and artificial intelligence-related risks.

Harbor ETF由Foresight Fund Services LLC分销。

Harbor ETFs are distributed by Foresight Fund Services LLC.

Harbor与AI Daily Brief无关，这些基金与任何AI实验室均无关联、未获其赞助或认可。

Harbor is not affiliated with AI Daily Brief, and the funds are not affiliated with, sponsored by, or endorsed by any AI lab.

这是一则付费广告，并非个性化投资建议。

This is a paid advertisement and not personalized investment advice.

投资涉及风险，包括可能损失本金。

Investing involves risk, including possible loss of principal.

本期AI Daily Brief由Hyperagent为您呈现，您可以在其中运行团队可共同管理的智能体集群。

This episode of the AI Daily Brief is brought to you by Hyperagent, where you run fleets of agents your team can manage together.

别再依赖本地代理和聊天工作流，等着你在笔记本电脑上操作了。

Forget local agents and chat workflows waiting on your laptop to be prompted.

Hyperagent 在云端部署常驻代理，在你团队已经使用的工具中完成实际工作。

Hyperagent deploys always-on agents in the cloud, doing real work across the tools your team already uses.

营销代理能把竞争对手的动作变成落地页。

Marketing agents turn competitor moves into landing pages.

销售代理能丰富潜在客户信息、起草邮件，并更新 CRM。

Sales agents enrich leads, draft emails, and updates the CRM.

运营代理负责处理文书工作，跟踪预算。

Ops Agent chases the paperwork and tracks the budget.

每个代理都能访问共享上下文，并遵循你关于范围和审批的规则。

Every agent has access to shared context and follows your rules about scope and approvals.

是时候让代理像队友一样加入你的团队了。

It's time you add agents that feel like teammates.

去 Hyperagent 招聘你的代理吧。

Hire yours at Hyperagent.

在 hyperagent.com/aidailybrief 领取100美元积分。

Get $100 in credits at hyperagent.com/aidailybrief.

好的。

All right.

我想确保我们清楚这期节目是给谁做的。我觉得它是为所有人准备的，不只是那些在搞最前沿技术的专业人士。

I want to make sure that we understand who I build the episode for, and I think it's built for everybody and not just the frontier professionals that are building the absolute cutting edge.

如果你正准备搭建一个智能体，那当然要留意，因为这期节目会给你提供或者验证完整的操作指南。

If you're about to build one, of course, pay attention because it will provide you or verify the full playbook.

同时，它也面向那些还没到那个阶段的人，因为正如你刚才说的，这些决策适用于你构建或使用的任何智能体，而团队智能体只是让其中一些决策变得更加关键。

It's also aimed at people who are not quite there yet, because the decisions, as you rightfully said, do apply to any agent that you build or use, and a team agent just makes some of them even more critical.

即使你是单干，但有一组智能体，或者正在考虑组建一组智能体，你也会面临同样的决策。

And even if you are working solo and you have a team of agents, or you're contemplating building a team of agents, you have the same decisions.

你生态系统中的另一个角色可能不是你的同行，因为你独自工作，但也许你是在为你的客户构建，或者是在为未来的自己构建，确保它足够稳健，能代表多样化的任务。

The other player in your ecosystems are probably not your peers because you work alone, but perhaps you're building it for your customers, or you're building it for a future you to make sure that it's robust enough and representative enough of diverse set of work.

所以这就是这期节目的动机，或者说谁应该关注。

So that's the motivation or who should pay attention.

还有一点我想特别说明的是，并不是每个智能体都应该被共享。

And one thing that I wanted to make sure that it's very clear is that not every agent should be shared.

这里有一个调节旋钮，或者说一个有三档的频谱。

There is a dial here or a spectrum with three settings.

我们有私有智能体，它只属于你，用于你的工作，带有你的品味和你的访问权限。

We have a private agent that's yours for your work with your taste and your access.

比如，我自己的社交媒体智能体就是保持私有的。

For example, my own social media agent stays private.

我可能没法把它分享给任何人，因为只有我想以特定的方式分享或写社交媒体内容，所以没有人应该听起来像我。

I'm probably not going to be able to share it with anybody because I'm the only one that wants to share or write in social in a specific way, so nobody should ever sound like me.

然后我们有共享知识。

And then we have shared knowledge.

这些可以是共享的知识和技能，但仍然是私有智能体，意思是团队维护一个知识库。

Those can be shared knowledge and skills, but still private agents, meaning the team maintains one body of knowledge.

比如，我们卖什么、我们怎么工作、我们的词汇是什么意思，通常还有一个共享的技能库，每个人都可以把自己的智能体指向那里。

For example, what we sell, how we work, what our words mean, often with a shared skill library, and everybody points their own agents at that.

顺便说一下，技能库就在这里，这通常是最合适的答案，也是最容易开始的地方。

But this is where the skill library lives, by the way, and it's often the right answer, and it's the easiest place to start.

再举一个具体的例子，假设每个销售人员都有一个按自己的风格和偏好定制的潜在客户开发代理。

Another concrete example, say every salesperson has a prospecting agent tuned to their own style and their own preferences.

保留这些代理的原样，因为每个销售人员都希望有自己的声音，但给他们所有人一个维护良好的理想客户画像和公司信息。

Keep those agents as they are because every salesperson wants to have their own voice, but give them all the same well-maintained picture of the ideal customer and the messaging for the company.

所以这是一种混合模式，有些公司或某些用例应该保持这种模式。

So that's a hybrid mode that some companies or some use cases should remain.

最后，我们还有团队代理，即一个代理被许多人使用，它有自己的职责和所有者，而且代理可以沿着这个刻度移动。

And lastly, we do have the team agents, where we have one agent that many people work with that has its own job and its own owner, and agents can move along the dial.

如果你的同事总是想借用你的私人代理，这很可能表明你需要把它变成团队代理，或者考虑与他人共享。

And if you have a scenario where your colleagues keep asking to borrow your private agent, that's probably a sign that you need to make it a team agent or to consider sharing it with others.

接下来我想回答的问题是，所有团队代理是否都来自同一原型，或者它们是否都遵循相同类型的用例？

So the next question that I want to answer is, are all team agents from the same archetype, or do they all follow the same type of use cases?

答案是否定的。

And the answer is not.

在我合作的团队中，我大致可以把现有的或将来要建的团队智能体分成四类。了解你属于哪一类，不仅能帮你找到用例，还能帮你优化用例，同时也能告诉你该注意什么才能做好。

Across the teams that I work with, I can roughly categorize the existing or future-built team agents into four kinds, and knowing which your kind is, it helps you not only identify use cases but also refine the use cases, and also it can tell you what to pay attention to in order to get it right.

第一种团队智能体，我称之为专家智能体。

So the first type of team agent, I'm calling it the expert agent.

它承载的是某个人或某个小团队的专业知识。

It's the one person's know-how or one small team's know-how.

所有依赖它的人都能使用。

It's available to everyone who depends on it.

你一眼就能认出来，因为有了它，那个人就能放心休假了，还记得开头提到的吧。

You'll recognize it by the person who can take the vacation, you remember from the beginning.

再举个例子，比如一个数据智能体，能回答公司里跨部门的各种数据问题；或者一个定价和交易台智能体，服务很多市场团队；还有合规智能体等等。

If you want another example, it can be the data agent that can answer any data question across multiple department in the company, or a pricing and deal desk agent that serves a lot of go-to-market organizations, compliance agent, and so on.

要做好这件事，知识必须来自公司里真正掌握这些知识的专家。

In order to get it right, the knowledge has to come from the experts that holds it in the company.

他们需要接受访谈。

They need to be interviewed.

你需要收集他们在多个论坛上已经给出的答案，无论是直接消息、电子邮件还是其他地方。

You need to collect the answers they already gave in multiple forums, whether those are direct messaging or emails in other places.

而且专家必须从第一天就参与进来，因为对他们来说，这最终让他们能休假，但也带来了很多工作不安全感。

And the experts have to be involved from day one, because for them, this is what finally makes the vacation possible, but also a lot of job insecurity.

所以在这个领域工作时要小心谨慎。

So tread carefully when working in this domain.

第二种类型是我所说的通用工作代理。

The second type is what I refer to the common work agent.

这种情况是很多人都在做类似的重复性工作，并且有一种共同的工作方式。

That's the scenario where many people are doing similar recurring work with one shared way to do it.

识别属于这一类的用例的方法是，当三个人各自构建了同一个代理的自己的版本时。

The way to recognize a use case that falls into this category is when three people have each built their own version of the same agent.

你可以想想团队研究和会议准备代理，这是一个非常经典的例子，或者营销团队的内容代理，我相信你还能想到其他的。

You can think about a team research and meeting prep agent, that's a very classical one, or marketing team's content agent, and I'm sure that you can think of others.

要正确实现这种类型，你必须就工作方式达成一致，这说起来容易做起来难，因为你是在融合三个版本的最佳部分，这实际上是一场关于公司标准达成共识的对话。

In order to get this archetype right, you have to agree on how work is done, and it's easier to say than to actually execute, because you're merging the best of three versions, and that's really a conversation about what you agree in terms of the standard for the company.

所以，一旦你开始考虑统一这样一个智能体，至少会是一场有趣的对话。

So an interesting conversation at the very least, once you start contemplating unifying an agent like that.

然后我们有桥梁智能体。

And then we have the bridge agent.

那是角色之间流动的工作，没有人能独自完成。

That's the work that flows between roles where nobody can do it alone.

当交接断裂时，你会意识到这一点，每个阶段都必须重新解释背景，或者智能体必须在不同部门之间协作。

You'll recognize it when the handoffs break, and every stage has to re-explain the context, or the agents have to somehow work together between different departments.

例如，一个客户智能体可能跨越销售解决方案、客户成功和交付。

Example can be a customer agent that spans sales solution, customer success, and delivery.

这是一个非常典型的例子。

That's a very classical one.

为了做好这一点，我们必须让每个职能部门贡献自己的知识片段，输入到这个智能体中，并且不同权限的人最终会使用它。

And to get it right, we have to have each function their own piece of knowledge that is being fed into this agent, and people with different access will be eventually using that.

所以我们也必须非常注意权限问题，并且要在这方面下功夫。

So we have to also pay attention very carefully to permissions, and you'll have to work hard through that.

最后，还有首席参谋这个角色。

And lastly, we have the chief of staff.

这个代理负责团队的日常运营，也就是把日常工作流程化。

That's the agent that own the team operating, like operationalizing of the day-to-day work.

它可能涉及决策、承诺、状态更新、新人入职等。当团队不断重复同样的事情，新员工要花好几周才能上手时，你就会意识到这个角色的重要性。

It can be the decision, the commitment, the status, onboarding, and you will recognize this one when the team keeps repeating itself, and new joiners take weeks to find their footing.

要做好这一点，你需要明确它被允许学习什么，如何学习流程和持续进行的工作，以及如何自动完成这些，这可不是件简单的事。

And in order to get this one right, what you need to do, you need to clearly define what it is allowed to learn, how can it learn the processes and the ongoing, and how does it do so automatically, which is not very trivial.

当然，还有很多关于权限之类的问题。

And then there are also, of course, many questions around permissions and so on.

所以，当你在思考这些原型，看看哪一个可能适合你正在考虑或应该考虑的一些用例时，在我给你讲如何真正构建这些团队代理的实用指南之前，我们先快速绕个弯，因为我想给大家一个现实检验。

So while you're thinking about these archetypes and which one of them might fit some use cases that you are pondering or that you should be thinking about, and before I give you the playbook on how to actually build these team agents, a quick detour, because I do want to give a quick reality check.

有一些迹象表明，团队代理可能不是正确的选择，或者至少现在不适合你。

There are some signs that a team agent is the wrong move, or at least not the right move for you at this moment.

一个你不应该构建团队代理的迹象是，当个人品味比标准更重要的时候。

So one indication where you shouldn't build a team agent, at least yet, is when taste beats standards.

如果不同的人真的需要不同的答案，或者不愿意接受统一标准，而且他们自己的判断或声音才是关键，那么这些应该保持为私人代理，这样人们才能保持真实，不必为基本事实争论。

If different people truly need different answers or are not willing to agree on a standard, and their own judgment or their own voice is the point, those need to remain private agents, so people can remain authentic and not have to fight about the ground truth.

第二个不该构建的迹象是，没有人能拥有知识。

And the second indication not to build is nobody can own the knowledge.

如果团队无法就工作方式达成一致，或者没有人愿意长期拥有和维护共享知识，代理会在几周内，有时甚至几天内偏离轨道。

If the team cannot agree on how the work is done, or nobody is willing to own and maintain the shared knowledge over time, the agent will drift within weeks, sometimes within days.

所以，在构建团队代理之前，先解决所有权问题，否则那将行不通。

So sort out the ownership before you go and build the team agent because that's going to be a no-go.

最后，当你意识到尝试构建一个共享代理或团队代理，只会让事情变得更复杂而不是更简单，因为需求冲突、权限纠缠、协调无休止。

And lastly, whenever you're realizing that trying to build a shared agent, a team agent, only complicates more than it simplifies because you have conflicting needs or tangled permissions, endless coordination.

如果你发现结果带来的工作量比代理能提供的帮助还多，那就是答案。

If you realize that the result is more work than what the agent can provide for you, that's the answer.

不要构建它，至少在你能够理清一些复杂性之前。

Don't build it, at least not until you are able to untangle some of the complexities.

当然，注意列表中缺少了什么：敏感数据和高风险。

And of course, notice what's missing from the list, sensitive data and high stakes.

这些是设计问题，它们决定了你怎么构建它，这也是我们接下来要讲的内容。

Those are design questions, and they shape how you build it, which is where we're going next.

所以我觉得，数据特别敏感并不是反对的理由，只是需要格外小心处理。

So I don't think that when data is overly sensitive is a reason against, it's just something that needs extra careful attention.

为了构建候选用例，希望你已经通过了之前分享的决策清单，接下来就是为你的代理做五个核心设计决策。

In order to build the candidate use case that hopefully you've gone through the decision checklist that I shared before, it comes down to five core design decisions for your agent.

这包括它做什么、放在哪里、知道什么、能接触什么，以及你怎么运行它。

It goes to what it does, where it lives, what it knows, what it can touch, and how can you run it.

这些都是核心问题。

These are the core questions.

为了让内容更具体，我会用一个例子贯穿始终。

In order to make it more concrete, I'll use one example the whole way throughout.

这个例子是一个客户代理，属于桥梁型，也就是说，它掌握公司对每个客户的所有了解以及我们对客户的承诺，可能会被销售、解决方案工程、客户成功、交付等部门使用。

So it's going to be a customer agent, and it's going to be the bridge kind, meaning one that holds everything the company knows about each customer and everything we've promised them, and it's probably going to be used by sales and solution engineering and customer success, delivery, and so on.

他们会用这个示例代理来做各种与客户相关的活动。

They will be using that example agent to do various customer-related activities.

那我们来看看这些决策，让它们更具体可行。

So let's break down some of these decisions to make it actionable.

首先，你需要做的决策是它做什么。

So first, the decision that you have to make is what it does.

这里先提醒一下，任何正经的工作型代理在范围界定上都有很多相似之处。

A quick caveat here, there is a lot of scoping that is very similar for any serious agent at work.

它需要有明确的任务、完成标准，以及它不做什么的清单。

It needs to have a clear job and a definition of done and a list of what it doesn't do.

我这里主要讲的是，为团队构建的代理和为你自己构建的代理有什么不同。

I'm going to stick here to what changes when it's built for a team versus an agent that you build just for yourself.

第一个决策是它服务于哪些角色。

So the first decision is who it serves by role.

比如，销售问它的东西和交付团队问它的东西就不一样。

For example, sales asks it different things than delivery does.

所以，把每个角色以及他们会来找代理做什么都写下来，确保你覆盖了所有范围。

So write down each role and what they'll come to the agent for to make sure that you're covering all the scope.

当然，你也可以瞄准一个大的工作领域，因为我觉得团队智能体应该能力相当强，否则很难证明它们存在的价值。

And of course, you can aim for one broad area of work, because I think the team agent should be quite capable agents, otherwise it's harder to justify their existence.

在我们的例子中，我期望客户智能体能够准备会议答复，比如我们和这个客户的关系进展如何，能够标记出承诺其他团队的承诺，并起草每一次交接。

In our example, I would expect that the customer agent will be able to prepare meeting answers, where do we stand with this customer, will be able to flag promises that commit other teams, and draft every handoff.

比如，范围相当广，但让它保持专注的是工作领域，在我们的案例中，主要是客户。

For example, of quite a broad scope, and what keeps it focused is the area of work, and that is primarily, in our case, customers.

我还想让你注意团队的“不做”清单。

I want you also to pay attention to the team don'ts list.

这部分人们往往容易忽略。

This is the part people often tend to skip.

比如，它永远不会代表别人做承诺，或者永远不会解决人与人之间的分歧。

For example, it never makes commitment on someone's behalf, or it never settles disagreement between people.

这些事应该交给它的负责人。

Those go to its owner.

你可以想想在你的情况中还有哪些类似的例子。

You can think of another such examples in your case.

另外，确保“不做”清单里包含权限和数据处理的注意事项。

Also, make sure that the don't list includes permissions and data handling stuff.

比如，它绝不会把某个私人空间的信息带到共享空间里。

So for example, it never carries information from one private space into a shared one.

所以它不会在一个客户的专属空间里讨论另一个客户，也不会替一个人向另一个人传话。

So it never discuss one customer in another customer's space, and it doesn't speak for one person to another.

当然，和其他智能体一样，最好从较小的范围开始，比如阅读和起草。

And of course, like with any agent, ideally start with narrower scope, reading and drafting.

只有当它赢得了足够的信任，并且经过了充分验证，你才能随着信心的增强逐步扩大它的职责范围。

Only when it earns sufficient trust and was validated enough, then you can increase the scope as you gain more and more confidence.

这就是第一个决策，我们可以进入下一个了。

So that's the first decision, and we can move to the next one.

第二个问题是智能体放在哪里，这个问题被问得最多，所以我们来具体说说。

The second question will be where the agent lives, and this is the question that gets asked most, so let's be a little bit concrete.

基本上，如果我想尽量简化，大致有三种共享智能体的方式，从最简单到最复杂的。

Basically, if I'm trying to make it as simple as possible, there are roughly three ways to share an agent, from the simplest to the most involved.

最简单的方法就是创建一个共享文件夹，里面放上你们公司已有的工具，然后每个人把自己的工具指向这个共享代理，文件夹里包含指令，以及如何存储新信息的方式，这些都作为代理整体行为的一部分。

The simplest method will just to create a shared folder with your own tools, like the tools that your company already owns, and then each person just point their own tool to the shared agent, and the folder will include instructions and potentially how to store new information as part of the way the agent overall behaves.

这是一种非常基础和简单的方法，但作为构建共享代理和团队级代理的起点，它可能是一个很好的开始，尤其是当你的团队成员已经在使用类似的代理工具，或者能够指向文件夹作为信息来源的代理工具时。

That's a very naive and basic way, but as a stepping stone to building shared agents and team-level agents, that can be very good start, especially if all of your team members are already using similar agentic tools or agentic tools that can point to folders as sources for their information.

这就是第一种最简单的方法。

So that's the first and simplest method.

第二种方法是使用供应商现成的代理，我们越来越多地看到这种情况，而且我们相信在接下来的几周和几个月里，我们会持续看到更多这样的产品。

The second way that we can do that is using a vendor ready-made agent, and we're increasingly seeing more and more, and we believe that we will continuously see more and more in the coming weeks and month of the year.

我们稍后会详细讨论。

We'll talk more about it later.

但这种方式的工作原理是，供应商托管它，你来配置它。

But the way this work is that the vendor host it, and you configure it.

在这个类别中，我们有很多最近非常出名的工具，包括Claude Tag，OpenAI有他们的ChatGPT工作区代理，Copilot也有自己的产品可以这样使用，Notion，还有很多其他工具已经提供了共享空间，里面有代理可以一起协作，你只需要配置它们的具体设置。

And in this category, we have many very recently famous tools, including Claude Tag, OpenAI has their ChatGPT Workspace Agent, Copilot has their own offering that can be used like that, Notion, and many others are already offering shared spaces with agents that you can work together on, and it's just a matter of you configuring their specifics.

最后，当然也是最复杂的一种，就是你自己托管的代理，也就是说，这是你自己运行的。

And lastly, and that's of course, the most sophisticated, is an agent that you host, meaning that's something that you run.

比如，可以是像OpenClaw或Hermes这样的开源代理，部署在你们自己的服务器上，或者某个云平台上，有些公司甚至还会专门为这些需求打造自己的定制化框架。

It can be, for example, an open source agent like OpenClaw or Hermes on your own servers or on a list cloud, or some companies are even building their own custom harnesses specifically for these needs.

所以这是最复杂的方式，但显然技术要求也最高，同时自由度也最大，可以围绕它做很多定制。

So that's the most sophisticated, but obviously has the most technically demanding requirements as well as the most freedom to build around that.

所以，这就是团队代理可以存在的三种大致选择。

So that's the three broad strokes options of where these team agents can live.

除了决定如何构建或选择工具之外，还有两个额外的问题需要考虑。

On top of the decisions of how to build or the tools, there are two additional questions that come with them.

第一个问题是，每个人和代理的对话，谁可以看到？

The first question is who can see each person's conversation with the agent?

可能只有正在和代理对话的人才能看到。

Maybe it's only the people who are conversing with the agent.

也可能是整个频道或者所有参与会话的人都能看到。

Maybe it's the entire channel or everyone in the session.

在这方面，不同的工具之间差异很大。

Over here, there is a lot of differences between the different tools.

有些工具里，每个人都能看到所有的聊天记录。

In some tools, everybody can read every chat.

比如在Slack里的Claude，频道里的每个人都能看到代理做了什么、跟别人聊了什么，还能引导它。

For example, in Claude in Slack, everyone in the channel sees what the agent does and what the agent converses with others and can also steer it.

而很多其他代理，你的聊天是私密的，这有时是供应商的设计决定，有时是你可以配置的。

With many other agents, your chat is private, so that's sometimes a design decision by the vendor, or sometimes it's something that you can configure.

第二个问题是，它学到了什么、存储在哪里，以及谁能看到。

And the second question is what it learns stored and who can read it.

所以理想的代理不是那种一成不变的，而是有很多记忆、能不断学习的。

So an ideal agent is not one that is obviously frozen, but one that has a lot of memory and learning on the go.

然后问题就是，这些学习内容存在哪里，又是怎么发生的？

And then the question, where is this learning being stored and how does it happen?

是每个人单独学习，代理只根据和你的互动来进化，还是按频道或团队来？

Is it something that happens per person, and then the agent evolves just from its interaction with you, or is it per channel or team?

或者整个工作区共享学习和记忆，代理在所有人面前公开进化。

Or maybe the entire workspace has a shared learning and memory, and the agent evolves with everybody in public.

当然，还有一点，永远不要混淆客户。

And of course, one thing, never cross customers.

所以，在第一个真正的任务之前，一定要检查、选择并告诉团队你构建的团队代理的状态，因为这是建立或失去信任的关键时刻。

So do check and choose and tell the team before the first real task what's the status with the team agent that you built because this is where a lot of trust can be gained or lost.

我的原则是选择最简单的方案，只要这周有两个人真正会用就行。

And my rule is to pick the simplest option that two people will actually use this week.

对于我们的客户代理，那可能就是一个频道代理或简单的工具代理，因为完整功能需要它，如果我把它做得过于复杂，人们需要理解如何连接，而不是直接进入Slack或Teams频道，那它就行不通。

And for our customer agent, that's probably a channel agent or a simple tool agent because full functions need it, and if I will make it overly complicated and people will need to understand how to connect to that versus just going into a Slack or Teams channel, it's not going to work.

所以，在我们的情况下，那可能就是正确的解决方案。

So in our case, that's probably going to be the right solution.

这个决定，我想把它提升到最重要的决定，那就是它知道什么。

This decision, I want to move arguably to the most important decision, and that's what it knows.

我觉得如果你构建过任何代理，这个配方听起来会非常熟悉。

And I think if you've built any agent, the recipe will sound very familiar.

对团队来说，不同的是，这是团队就基本事实达成一致的时刻，即工作到底是怎么完成的，我们使用哪些定义，哪个版本的价格政策是真实的。

What's different for a team is that this is the moment the team agrees on the ground truth, how the work actually gets done, which definitions we use, which versions of the pricing policy is the real one.

我觉得这个对话值得进行，即使你最终没有推出这个代理，因为在大多数团队里，光是统一知识就已经是件大事了。

And I think that that conversation is worth having, even if you never ship the agent, because in most teams, even just agreeing on the knowledge is a big deal.

这也是团队代理变得更难的地方，因为一旦知识被共享，就会有更多的贡献者、更多的矛盾，也会有更多的地方让重要的事情被遗漏。

And it's also where team agents get harder because the moment knowledge is shared, then you have more contributors and more contradictions and more places for something important to fall through.

所以围绕知识的管理流程比你在个人代理中做的更加重要，你需要在这里格外注意。

So the process around the knowledge matters even more than it ever did in your private agent, so you need to pay careful attention here.

理想情况下，你应该经历这四个阶段。

And ideally, you should go through these four stages.

我希望你先从收集代理开始，通过采访那些掌握知识的人，并收集已经存在于所有渠道和所有信息所在之处的资料。

I want you to start by collecting the agent, by interviewing the people who hold the knowledge, and harvest what's already written in all the channels and all the places where information already resides.

我希望AI在汇总和收集数据方面承担大部分繁重的工作。

And I want AI do a lot of the heavy lifting in terms of aggregating and collecting the data.

所以对于我们的客户代理，我会确保销售、解决方案、成功和交付团队都贡献他们各自的部分。

So for our customer agents, what I would do is I'll make sure that sales and solutions and success and delivery, they all will contribute their own piece.

然后我希望你进行细化。

And then I want you to refine.

我希望你把来自不同来源的信息合并起来，把矛盾的地方暴露出来。

I want you to merge the information coming from different sources, surface the contradictions.

我敢肯定你会发现有五个版本的真相，然后给所有内容标上日期，把那些绝不应该分享的东西剔除掉。

I'm sure that you will find five versions of the truth, and then date everything and keep out what should never be shared.

当然，这包括密码、关于人的笔记、一个客户的详细信息出现在另一个客户的空间里，等等。

Of course, that includes passwords, notes about people, one customer's details in another customer's space, and so on.

然后我们必须批准它。

And then we have to approve it.

每一部分都应该由它的所有者签字确认，而代理的所有者负责把它们整合在一起。

Each piece should be signed off by whoever owns it, and the agent's owner puts it all together.

最后，这些内容很快就会过时，所以你必须持续维护。

And lastly, this can go stale very quickly, so you have to maintain.

你需要决定代理可以根据自己的工作经验往记忆里添加什么，以及当信息进入知识和记忆时，哪些内容需要先由人来审核。

You need to decide what the agent may add to its own memory based on its working experience and what person has to review first when it goes into the knowledge and the memory.

我们希望确保，不会让某个人对去年情况的定义在没有任何共识的情况下悄悄变成所有人的定义。

And we want to make sure that one person's definition of what's last year quietly becomes everyone's without any agreement.

当然，还要把维护工作排上日程，因为你希望代理定期提出更新，而需要有人来审核并批准这些更新。

And of course, put the upkeep on a schedule because you want the agent to propose updates regularly, and a person needs to review and approve the updates.

这确实是一个需要非常非常认真和自律的地方，因为如果你没有建立一个足够好且能自我维持的流程来获取、维护和验证代理所依赖的知识，那真的可能成就你的团队代理，也可能毁了它，因为一旦出了问题，你不再能快速修复一切。

And this is really the place to be very, very diligent and disciplined because it can truly make or break your team agent if you haven't done a good enough and self-sustaining process around acquiring and maintaining and verifying the knowledge that the agent taps into, because it no longer serves you where you can very quickly fix anything that goes wrong.

如果你的团队代理对重要事项了解不够充分，那可能会在公司里造成很大的混乱。

It can create a lot of havoc in your company if your team agent is not well-educated enough on what matters.

第四个决定是它能接触什么，这也是团队代理和私人代理的区别所在，因为你的私人代理代表你个人，而团队代理则代表很多人。

The fourth decision is what it can touch, and this is where team agents differ from the private ones because your own agent acts as you, and a team agent acts for many people.

所以当然，这里需要应用很多基本的安全措施。

So of course, there are many security 101 that you need to apply here.

那些适用于任何代理的规则肯定要遵守，但我这里只讲适用于团队代理的规则。

Those that apply to any agent, definitely stick for the four rules that are specific to-- I'm going to just stick to the rules that apply to team agents.

你首先要决定的是它使用谁的访问权限，你有三个选择。

The first thing that you have to decide is whose access it uses, and you have three options.

你可以使用提问者的访问权限，或者以提问者的身份行事。

You can use the access or to act as whoever is asking.

所以它只能看到提问的那个人能看到的内容。

So it only sees what the person that was asking the question can see.

这可能是最安全的选择，但有时候执行起来最复杂，除非供应商已经帮你做好了。

And that's probably the safest choice, but sometimes the most complicated to execute unless the vendor already did it for you.

当团队成员有不同的访问权限时，这可能是正确的选择。

And probably the right one when people on the team have different access levels.

另一个选择是为它设置一个独立的账户，只给它工作所需的访问权限，这适用于整个团队共享同一份材料的情况。

The other option that you have is to have its own account set up with exactly the access the job needs, and that's right when the whole team works on the same shared material.

最后，它可以使用某个人的登录信息。但只建议用于只读和非敏感的材料，因为每个和它对话的人实际上都会获得那个人的访问权限。

And lastly, it can use one person's login Only ever do that for read-only and non-sensitive material because everyone who talks to the agent effectively gets that person's access.

所以我不建议走这条路。

So I would not recommend to go down that path.

对于我们的客户代理，我可能会让它以提问者的身份来操作，因为销售成功和交付在CRM和其他系统中看到的内容不同。

And what I would probably do for our customer agent is to act as the person asking because sales success and delivery see different things in the CRM and in other systems.

所以我不想让代理向他们回复他们不应该看到的信息。

So I don't want to have the agents responding to them with information that they shouldn't be able to see.

所以这就是关于智能体能看到什么的第一条规则。

So that was the first rule on what the agent can see.

第二条规则是决定谁能向它提问。

The second rule is to decide who can ask it.

当智能体有自己的账户时，所有能跟它对话的人都可以使用这个账户。

And when the agent has its own account, everyone who can talk to it can use the account.

如果你把一个能访问定价表的智能体放在一个40人的频道里，其中还有几个外包人员，那么突然之间，这40个人都能通过提问获取定价信息。

And if you put an agent with access to the pricing sheet in a channel of 40 people, a few contractors among them, then all of a sudden all 40 can now get the pricing by asking.

智能体知道的信息比一些能接触到它的人还要多。

The agent knows more than some of the people who can reach it.

所以，决定谁能跟它对话时，要像决定它能看什么一样谨慎。

So decide who can talk to it with the same care you give to what it can see.

好的，这就是人们不注意时经常发生的事情。

Okay, so that's something that happens very regularly when people don't pay attention.

我还想让你决定答案落在哪里。

I also want you to decide where the answer lands.

这个规则是反过来的。

This one runs the other way.

智能体用的是提问者自己的权限，而提问者完全有权问这个问题。

The agent uses the asker's own access, and the asker has every right to ask the question.

问题在于，答案会出现在一个共享空间里，被那些没有权限的人看到。

The problem is that the answer shows up in a shared space in front of people who don't.

所以这是现在正在发生的事情。

So this is something that is happening right now.

如果你看一下Claude在Slack里的文档，它可以在团队频道里使用提问者自己的连接。

If you will look at the documentation of Claude in Slack, it can use the asker's own connections inside the team channel.

在用户批准之后，Anthropic的文档目前也提到，它不会考虑频道里还有谁。

And after the person approves, and the Anthropic documentation currently notes that it doesn't consider who else is in the channel.

所以如果我批准使用我的连接器，获取所有我有权查看的信息，然后这些信息被扔到频道里，其他人也能看到，这就是目前Claude实现的现实情况。

So if I approve to use my connectors and fetch all the information that I am permitted to see, and now the information is thrown at the channel where others can see that, that's the reality currently with the existing Claude implementation.

所以如果信息敏感，答案必须私下发给提问的人，而不是发到共享频道里。

So if it's sensitive, the answer has to go to the person who is asking privately and not in a shared channel.

最后，要记录谁问了什么。

And lastly, keep record of who asked for what.

这很重要，因为当代理用自己的账号操作时，日志显示的是代理做的，对吧？

And that's an important logging because when an agent works under its own account, the logs say the agent did it, right?

你需要知道是哪个用户提出的请求，这样才能追溯，确保没有意外行为。

And you want to know which person asked, so you can backtrack and make sure that there are no unexpected behaviors.

其他方面，比如从最小权限开始，以及任何不可逆的操作都需要人工批准，这些对所有代理都是一样的。

And everything else, like starting with the least access and having a person approve anything that can't be undone, is the same for any agent.

所以我不是在给你上安全基础课。

So I'm not giving you security 101.

好的。

Okay.

最后，最后一个决定，也是让你的团队代理能撑过第一周或第一个月的关键。

Lastly, last decision and the one that will make your team agent live beyond its first week or the first month.

这就是完整的操作手册。

That's the full operating manual here.

当然，多人协作的冲刺阶段有更全面的思考方式，但实际中，这四点才是决定智能体成败的关键。

Of course, the multiplayer sprint has a much more comprehensive way of thinking about it, but these four points are what makes or break the agents in practice.

所以第一点就是要有唯一负责人。

So the first thing is one owner.

团队里任何人都可以交接工作，但我希望有一个人或一个很小的团队来负责优先级、维护智能体，并在两个人提出相反要求时，决定如何发展智能体的知识或功能集。

Anyone on the team can hand the work, but I want to have one person or a very small group of people who owns the priorities, maintain the agent, and decide when two people ask for opposite things, how to evolve the agent knowledge or feature set.

他们还要决定长期指令等事项。

They also decide on standing instructions and so on.

这就是第一点。

So that's one thing.

第二点我想说的是明确的互动规则。

The second thing that I want to mention is the clear rules of engagement.

我希望你告诉人们如何使用它，它能做什么、不能做什么，它能看什么，谁能看到与它的对话，以及我们之前讨论的所有内容。

I want you to tell people how to work with it, what it does and doesn't do, and what it can see, who can see their conversation with it, and everything that we discussed so far.

我们还需要明确指示，当智能体出错时如何纠正它。

We also want to have clear indications of how to correct the agent when it's wrong.