如何从AI高级用户成长为AI领导者

现在，任何一个专业群体，无论你怎么划分，总有一部分人成了AI的重度用户。

Every group of professionals organized in any way you can imagine at this point has some subset of people that have become the power AI users.

这些人会尝试每一个新模型和新产品。

These are the folks who try every new model and every new product.

他们会反馈哪些好用、别人应该怎么做，很多时候，这个角色已经被正式化，变成了某种内部大使或推广者计划。

They report back on what works and what others should do, and in many cases, that role has been formalized into some sort of internal ambassadors or champions program.

事实上，在我所有的企业经验中，我看到的最具投资回报率的转型活动，就是公司把这种内部推广者计划组织得很好。

In fact, across all of my experience with enterprises, the single highest ROI transformation activity that I have seen is when companies structure this sort of internal champions program well.

今天的节目就是讲这到底意味着什么。

Today's episode is all about what that actually means.

《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，在我们开始之前，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、Section、Harbor和Blitzy。

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

想要收听无广告版本，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.

如果想了解赞助节目的详情，请发送邮件至 sponsors@aidailybrief.ai。

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

最后，今天是美国的联邦假日，所以这期节目是上周录制的。

Lastly, today, it is a federal holiday in the US today, so this was recorded last week.

这又是我们的一期《操作者剪辑》实用节目，希望你喜欢。

It's another one of our Operators Cut practical episodes, and I hope you enjoy it.

好的，Nufar，欢迎再次来到《操作者剪辑》。

All right, Nufar, welcome back to another Operators Cut.

你好吗？

How are you?

我很好。

I'm good.

谢谢邀请我。

Thanks for having me.

我对这一期非常期待。

Excited about this one very much.

也许我们可以一开始就大胆地说一句，你和我都分别或一起花了很多时间与各种规模的企业和公司打交道，他们都在努力提升AI应用能力。

Maybe just to make a bold claim right up front, you and I have both individually and together spent a lot of time interacting with enterprises and companies of all shapes and sizes who are trying to get better at AI.

在这个过程中，我们和他们尝试了各种各样的策略，来改进他们在特定团队或整个公司范围内使用AI的方式。

And as part of that, we have tried, and they have tried, a huge array of different strategies for improving how they use AI within specific teams or across the board.

而今天我们要讨论的，是唯一一种能够持续、始终成功地改善公司或团队内部员工AI使用体验的方法，对吧？

And what we're going to talk about today is one of the only consistently, always successful ways to make the general experience that people have with AI inside a company or inside a team better, right?

简单来说，这个方法从未失败过。

It never fails, is the short answer.

这确实是人们工作中非常关键的一部分，所以我很期待深入探讨。

This is just such a critical component of what people do, so I'm really excited to dig into it.

先给我们介绍一下你在做什么吧，我想今天我们会聊很多。

Introduce us to what you're doing, and then I think we'll have a lot to talk about today.

太好了。

Perfect.

是的。

Yes.

所以我同意。

So I agree.

我们现在要讲的是在公司里成功运用AI回报率最高的活动之一，那就是建立一个合适的AI冠军网络。

We're now presenting one of the highest ROI activities that you can take in a company in order to be successful with AI, and that is to establish a proper network of AI champions.

今天这个环节我想做的是这样的，因为我们每天都在用AI，我相信大家都见过这种现象：团队里有人找到了一个很好的做事方法。

And what I want to do with the session today is the following, because we all use AI every day, and I'm sure that everybody has seen the same phenomena where someone on the team finds a great way to do something.

通常，这些人都是最优秀的用户。

Often, those are the best users out there.

他们给两个人演示了一下，然后这个方法就再也没有被整个团队采用。

They show two people, and then it never turns into something the whole team uses.

这一集节目要做的部分工作就是帮助解决这个问题，帮助定义这个人。

And part of what this episode is trying to do is to help resolve the problem, to help define the person.

可能是你，也可能是公司里的其他人。

It might be you, it might be others in your company.

他们的工作就是解决这个问题，也就是帮助推广AI的那个人。

Their job is to fix that, the person that helps spread the AI.

我觉得很多公司现在已经在设立AI推广者了，效果各不相同，但在我看来，总体结果都是积极正面的。

And I think that many companies have been establishing AI champions for a while now with varied degrees of success, but always, in my opinion, with a positive outlook and outcome overall.

我的目标是讲三件事，让这个做法变得更好。

My goal is to go over three things to help make it even better.

首先，这个角色到底是什么，怎么定义它，顺便也能帮那些正在听的人解答一个问题：他们自己是不是已经是公司里那个定义模糊的AI推广者，或者他们应该考虑成为这样的人。

First of all, what the role really is and how to define it, and maybe as a by-product, to also help the question that maybe someone listening to this has with their self of whether they are already the under-defined AI champion in their environment, or maybe they should opt to become one.

我还想解释一下，我认为我们应该如何建立一个合适的内部推广者网络。

I also want to explain how I believe we should establish a proper network of internal champions.

最后，我想讨论一下，在一个公司里，最优秀的AI推广者需要做些什么。

And lastly, I want to discuss what the best AI champions in a given company, what they need to do.

在我们深入之前，我想补充三点，我觉得这很重要。

And before we dive in, I want to add three things, I think, to this.

首先，我知道你接下来要给出你对AI推广者的定义，但就命名而言，我想指出，在我们合作过的所有这些不同团队中，你合作过的，我合作过的，这个角色的名称都不一样。

First, I know you're about to give us your definition of an AI champion, but just on a naming convention, I want to point out that across all these different teams that we've worked with, you've worked with, I've worked with, the name is different for this.

有时候叫布道者，有时候叫领导者，有时候叫采用支持者。

Sometimes it's evangelist, sometimes it's leader, sometimes it's adoption support.

冠军是我们听到最多的一个词，或者是一些变体。

Champions is the one I think we've heard the most often, some variation on champions.

所以我们指的就是这个。

So that's what we're referring to.

你可能在团队里听到过不同的叫法，或者有不同的名称。

You might have heard it differently or have it named differently in your team.

第二点是，我觉得对于那些不在大型企业里工作、正在犹豫“这集节目还值不值得我继续花时间听下去”的人来说，

The second thing is, I think that for those of you who are not inside a big enterprise and are wondering, "Is this worth continuing to invest my time in for the rest of this episode?"

我可以很有力地论证，一旦你过了基础或中级水平，进入高级AI领域，你就会被推到这个冠军的角色上，不管以什么形式。

I would make a fairly strong argument that once you get past a basic or intermediate level, once you're in that advanced AI category, you are going to be thrusted to this role of champion in some way, shape, or form.

至少对你的朋友和家人来说是这样，但如果你在团队里和任何人一起工作，这个冠军角色，它不仅仅是使用AI，而是沟通如何用好AI，以及用流程的思维去分享你正在做的事情，而不是仅仅停留在实施层面。

It's going to be for your friends and family at the very least, but it also is likely if you are working with any number of people on a team, this role inhabiting the champion's role, it's a different muscle than just using AI, but communicating how to use it well and thinking in terms where you can share what you're doing, process terms basically rather than just implementation, I guess.

第三点，因为Nufar不会自己夸自己，我觉得没有多少人能像Nufar那样，在这么多不同的公司里做过这么多次这样的分享。

And third thing, because Nufar wouldn't shout this out, I don't think that there are very many people who have done more versions of this with more different companies than Nufar.

所以这背后有很多经验，不只是在家随便捣鼓的经验，而是直接面对面、在几十家不同公司、不同场景下实际做过这类工作的经验。

So this comes out of a lot of experience, and not just hacking at home experience, but direct in-person, dozens and dozens of companies in different contexts in which this type of work has been done.

谢谢。

Thank you.

好，现在轮到我来说了，我得先定义一下它到底是什么。

Okay, I now have a cross to bear, and that is to first define what it really is.

就像纳撒尼尔说的，很多时候，你已经认识这个人了。

As Nathaniel said, in many cases, you already know this person.

就是每当你想知道怎么用AI搞定某件事，或者有什么新功能出来了，你想搞清楚怎么用的时候，你第一个打电话找的那个人。

That's the person that you pick up the phone to whenever you want to know how to get something with AI done, or there is something new that is now accessible, and you want to understand how to work with that.

我的定义是，AI冠军就是团队或环境里负责推动AI采用和价值落地的那个人。

And my definition is that the AI champion is the person in the team or in the environment that is responsible for moving AI adoption and value forward.

我经常叫他们“同事加一”，因为他们就在团队里，做团队的工作，但在AI方面领先一步，因为他们要么是最热衷、最有能力的那个人，要么就是被指派担任这个角色。

Often I call them peer plus one because they sit within the team, they do the team's work, but they're a step ahead in AI because they already are the person that are most psyched and able in AI or they were appointed to this role.

所以不管这个任命是正式的，还是你实际上就是那个人，都不重要。

So it doesn't matter whether this appointment was done officially or that's who you are de facto.

我觉得AI冠军就是那些在同事实际做的工作中展示AI能做什么的人，不是理论上的，而是他们日常环境中的实际用例。

I think that AI champions are people who show what AI can do on the work that the peers actually do, not theoretical, but use cases from the day-to-day of their environment.

他们还会和其他冠军联系，很多时候会和中央AI团队或其他专业人士合作，来构建和使用公司提供的相关资源。

They also connect with other champions and in many cases with the central AI teams or other professionals in order to build and use the relevant things the company offer.

他们还能在没有正式权力的情况下，推动大家去采用AI。

And they also move people, without necessarily having authority on them, forward with the AI adoption.

在我们讨论具体工作之前，这个角色在不同公司里差别挺大的。

And before we go to the actual job, the role does look quite different in different companies.

所以我想给大家一个范围。

So I wanted to give you a spectrum.

这个范围，就像Nathaniel说的，是基于在多家公司积累的大量经验。

The spectrum is, as Nathaniel said, based on a lot of experience across multiple companies.

一方面，有些公司会称他们为“大使”。

On the one hand, we have what sometimes companies will refer to as ambassadors.

这些人是最擅长使用AI的。

These are the best AI users.

他们的主要职责是推动AI的应用，展示实际用例。

Their role is primarily to drive AI adoption, to showcase the use case.

他们会带来真实的例子，同时也会做一部分工作，并且正式或非正式地培训同事。

They will bring their real examples, and they will do part of the work as well as training their peers, officially or unofficially in that case, by the way.

他们还会注意表达团队的需求，包括工具和内部开发，这些都能帮助团队更好地推进AI的应用。

And they also note how to voice what the team needs, the tools and the internal builds that will help the team move forward with AI adoption.

在光谱的另一端，就是我们所说的内部FDE。

At the complete other end of the spectrum, we have what we can now call them an internal FDEs.

FDE指的是全栈部署工程师，很多人可能已经听过这个术语很多年了，所以它有很多不同的解释，因为这个词已经存在一段时间了。

FDEs stands for full deployed engineers, and many of you have probably been hearing this terminology for many years, so it's a bit loaded with many different interpretations because they've been here around for a while.

但现在很多FDE都专注于AI。

But now many of them are very much focusing on AI.

所以在我们这个语境下，我简单定义一下：FDE就是业务团队内部的专业开发者，负责构建团队需要的东西。

So I'm going to define it quite simply in our context as a professional builder who sits inside a business team and builds what the team needs.

从想法到生产，包括我们第二部分会讨论的规划和价值实现，都可以涵盖。

It can be all the way from idea to production, including, as we will discuss in the second part, the planning, the value realization.

有些人有软件和AI背景，有些人，我见过很多这样的，是领域专家，他们真正实现了职业转型，利用现在很多人能用的各种神奇工具为团队构建东西。

Some of them come from software and AI background, some of them, and I've seen many of those, were subject matter experts who made real career pivot into building for their teams with the many amazing tools that now many people can build with.

介于两者之间的是大使型构建者，很多公司通常都会采用这种模式。

In between, where many companies typically land, the ambassador builder.

他们也为自己和直属团队构建东西。

So they also build for themselves and for their immediate team.

他们通常会有明确的职责范围，知道自己能做什么、不能做什么，或者应该有这样一个范围。

They typically will have a well-defined scope of what they can and cannot do, or they should have one.

超出这个范围，他们几乎就像产品经理一样。

And beyond that scope, they will be almost like a product manager.

他们会收集需求，然后交给相关的供应商或内部团队，让他们帮忙正确地构建他们想要的东西。

They will gather the requirements, they will bring them to the relevant either vendors or internal teams that can help them build properly what they want.

定义说清楚了，接下来我们聊聊具体怎么做。

So with the definition out of the way, let's talk about the how-tos.

AI大使的具体做法包括两个部分。

And the how-tos for AI champions, they're comprised of two parts.

第一部分是从公司角度出发，讲如何一步步搭建冠军网络。

We have part one, which is the company side, how you build a champion network stage by stage.

第二部分则是从冠军本人的角度来讲。

And in part two is the champion side.

就像我们说的，不管你是正式的还是非正式的冠军，你都需要知道该做什么，以及怎么把工作做好。

And as we said, whether you're official or unofficial champion, what you need to know and how to do the job well.

我觉得第二部分尤其适合所有超级用户，你们可能也想提升一下自己的工作方式和思维方式。

And I think that the second part especially will be applicable to all of you super users who just want to also progress some of what they do and their thinking.

当然，如果你是赞助方，第一部分就是为你准备的行动指南，我们先从公司角度开始。

So of course, if you are a sponsor, the very first part is the playbook for you, starting with the company side.

我觉得当公司讨论AI冠军时，他们从来不会只问一个问题。

And I think that whenever a company is talking about AI champions, they never just ask a single question.

他们基本上会问四个问题。

They basically ask four questions.

其中很多问题不一定只跟冠军有关。

Many of them are not necessarily just related to champions.

只是当你开始思考如何推动内部更好地采用AI时，这些问题往往会浮现出来。

It's just that when you start thinking about how can I create internal movement towards better adoption, often these questions come to play.

所以他们经常问的第一个问题是，谁有资格去构建？

So the first question they often ask is, who gets to build?

这基本上就是关于民主化的问题。

And that's basically the democratization question.

谁使用别人构建的东西，谁又是既为他人也为自己构建的人？

Who uses what others are building, and who's the one that is building for others and themselves?

第二个问题是，我们对其他人有什么期望？

The second question is, what do we expect from everyone else?

因为冠军有他们的职责，他们周围的每个人，也就是所有其他用户，也都有自己的职责。

Because champions have a job and so does everybody around them, all the other users.

所以这就引发了很多关于普通用户和冠军之间期望差异的问题。

So it creates a lot of questions around expectations from regular users versus champions.

他们还经常问，AI是如何从一线传播开来的？

They also often ask, how does AI spread from the field?

也就是说，一线的人怎么才能用上为他们建的东西，而且用得顺手。

Meaning how people in the ground pick up in a way that works whatever was built for them.

最后，这些骨干怎么跟中央团队或者供应商配合？

And lastly, how do the champion work with the central teams or the vendors?

我指的是那些负责公司AI工具、IT、安全以及所有做决策的人。

And by that I mean the people who own the company's AI tools and the IT and the security and everybody else that makes the decision.

就算你们公司没人有中央团队这个头衔，这些问题也依然存在，因为正是这些核心问题，区分了那些很有主见的公司和还在摸索中的公司。

And even if nobody in your company has the title of central teams, you are living with these questions because these are the core things that differentiate between a company that is very opinionated to the ones that are still kind of asking a bunch of questions.

接下来我们聊聊建立AI网络的过程。

Let's go to process of establishing an AI network.

我觉得我总喜欢问公司的第一个问题是，你们建立AI骨干网络，到底想达成什么目标？

I think the very first question that I always like to ask companies is, what are you planning to accomplish by establishing an AI champion network?

很多时候他们答不上来。

And very often they struggle to answer.

然后我可能会问的第二个问题是，从我们见过的宣传大使，到完全自主开发的工程师，你希望你的骨干处在哪个位置？

And then the second question that maybe I'm asking on the line between the ambassadors that we've seen to the full builders, the four deployed engineers, where do you want your champions to be?

这个问题通常能打开话题，因为现在我们讨论的是更具体的事情，这些人被允许做什么，他们需要做什么，以及哪些事情仍然由中央团队或专业供应商和开发者负责。

This one usually opens the conversation because now we're talking about something more concrete, what these people will be allowed to do, what they'll be expected to do, and what stays within central teams or only professional vendors and builders.

即使你已经定义了一段时间的推广者，定期进行这样的讨论也很重要，因为有时候公司的情况会发生变化，你可能已经根据实际经验对民主化和基层建设有了更深入的理解。

And even if you've been defining champions for a while, that's an important discussion to have periodically because sometimes that changes for companies, and you might have refined your thinking based on actual experience with democratization and building on the ground.

如果你想顺利开始，我希望你明确目标、职责和期望，最好写下来，并得到所有相关方的同意，比如IT、安全等部门。

When you want to start well, I want you to configure the goals, the responsibilities, and the expectations, ideally written down and agreed with everybody else who is involved, like IT, security, and so on.

举个具体的例子，我合作的一家公司，在思考了这些问题一段时间后，他们意识到主要希望推广者成为一个好的渠道，把相关的使用案例从基层带给中央团队，因为中央团队对各个团队的日常实际情况并不了解。

To give a concrete example, one company that I'm working with, after pondering these questions for a while, they realized that primarily what they want is their champions to be a good funnel that brings the relevant use cases from the ground to the central team that is basically blind to the day-to-day reality of the different teams.

因此，对于主要目的是成为团队优秀使用案例渠道的人来说，这是一个完全不同的讨论、角色和培训。

And as such, it's a completely different discussion, and role, and training for people whose main purpose is to basically become a funnel for good use cases coming from the team.

另一方面，我也见过一些公司在一段时间后改变主意，意识到他们希望推广者成为全面的开发者，然后他们培训这些推广者，使他们变得更加专业，几乎就像他们从零培养的AI工程师一样。

On the flip side, I've also seen companies that change their mind after a while and then realize that they want to have champions as full-blown builders, and then they train them to become much more sophisticated, almost like the AI engineers they breed from the ground up.

这就是第一阶段。

So that's the stage one.

第二阶段是提名，我希望你首先在一个好的推广者身上寻找三个信号。

Stage two is the nomination, and there are three signals that I want you to look for very first in a good champion.

一个好的冠军是那种即使没人要求，也会主动使用AI，并且用得最熟练的人。

A good champion is a person that are among the most sophisticated AI users without being told to use AI.

他们就是自然而然会这么做。

They just do.

他们的同事已经会主动找他们请教问题、寻求指导，而且他们的经理也会给他们足够的时间去处理AI相关的工作，甚至是为其他人服务。

Their peers already come to them to ask questions and to get guidance, and their managers will actually give them enough time to work on AI also for others.

这一点很重要，因为很多时候，好的提名就是在这里栽跟头的。

And that's a big one because often that's when great nominations fall short.

然后还有其他一些特质，比如在没有正式权力的情况下也能影响他人，了解业务的来龙去脉，主动性强，还有强烈的好奇心，因为大家都知道，一切都在不断变化。

Then there are other things, like the ability for them to influence without authority, knowing the business ins and outs, taking initiative, and a lot of curiosity because, as we know, everything changes all the time.

我还想提醒你，你提名的那个人，是你希望留住并提拔的，因为成为AI冠军，然后在社交媒体上炫耀等等，这是一个巨大的福利，也会让这个人成为市场上的抢手货。

I want you also to remember that the person that you will nominate is a person that you want to retain and promote because getting to be the AI champion and then boasting on that on social and so on is a huge perk, and also makes this person highly desirable for the market.

所以一定要确保你提名的是真正配得上这个角色的人。

So make sure that you nominate the people that truly deserve the role.

在提名这个环节，我也见过很多公司犯下无数错误，我真的很想指出来，因为这些错误太常见了，如果不指出来，那就是我的失职。

This is also the area, the nomination, where I've seen companies making endless amount of mistakes, and I did want to call them out because they are so common that I will be amiss if I will not call them out.

我看到的第一个问题是，一些非常资深的人只是为了提升自己的AI技能而提名自己，他们说‘我要当AI冠军’，或者因为他们想影响正在建设的内容，所以提名自己。

The first thing that I'm seeing is very senior people nominating themselves just because they want to improve their AI skills, so they say, "I'm going to be the AI champion," or because they want to have influence on what's being built, so they nominate themselves.

资深人士应该坐在AI委员会里，就像一个专门的指导机构，帮助做出战略决策。

Senior people should sit in an AI committee, like a dedicated steering function that helps make the strategic decision.

他们不应该成为负责培训和指导同事的人。

They shouldn't be the people who are chartered to build and train their peers.

我还看到，这种情况很常见，有些候选人在AI能力方面很出色，其他方面也都符合，但他们同时也是团队里最忙的人，根本没有时间和精力去负责培训和指导等工作，所以提名他们并不是明智的选择。

I've also seen, and that's very common, great candidates in terms of their AI ability and everything else falls in place, except they're also the busiest people on the team, and they just don't have the bandwidth to be the person who builds and trains and so on, so they are not a good use of your time.

我还看到有人因为奖励而提名。

I've also seen people nominated as a reward.

他们是优秀的员工，让我们提名他们为AI冠军吧，但他们并不是足够好的AI用户，或者不具备其他一些特质。

They are strong players, let's nominate them as AI champions, but they're just not good enough AI users or doesn't hold some of the other traits.

最后，千万不要提名昨天刚入职的人。

And lastly, never nominate the person who just started working yesterday.

AI冠军最重要的特质之一就是能够识别用例，并且了解业务的来龙去脉。

One of the most important traits that the AI champion has is the ability to identify use case and know the business ins and outs.

如果他们刚加入团队，不了解情况，那他们就不会成为足够好的负责人。

If they don't know they're new to the team, they will not be a good enough champion.

好的，接下来，每个认真对待负责人的公司都需要做的，就是培训。

Okay, the next thing that every company that is serious about champions need to do is the training.

当然，如果你们内部有足够强的人，也可以自己来办。

And of course, you can run it internally if you have strong enough people.

你可以办一个更集中的训练营，这通常是我常做的，或者你也可以循序渐进地培训他们。

You can do a more concentrated boot camp, which is what often I get to do, or you can go gradually and just train them.

但即使他们是优秀的用户，理想情况下，你也应该针对特定内容进行培训，也就是我第二部分要分享的内容。

But even if they are great users, ideally, you should train them on specific things, namely part of what I will share in the second part.

在我看来，无论你选择哪种形式，都要让他们在自己的实际工作中进行训练。

In my point of view, whichever format you choose, train them on their own work.

我最喜欢的测试非常简单。

And my favorite test is very simple.

每个负责人都要从自己的团队中选一个真实的用例，展示他们在培训结束时是如何成功构建或推广这个用例的。

Every champion takes one real use case from their team and shows how they were able to build or promote it by the end of whatever training that you do.

如果他们能做到这一点，那就准备好了。

If they can do that, they're ready.

如果不行，可能还需要再做一些工作。

If not, maybe there is some more work to be done.

通常，这个阶段的培训会以启动会或某种认证来结束。

And usually, this stage of the training ends with a kickoff or some kind of certification.

大家都非常兴奋。

A lot of excitement across everybody.

被选中的人感到非常自豪，而真正的考验才刚刚开始。

People are very proud to have been chosen, and that's where the real test begins.

很多项目就是在这里失败的，因为决定成败的最大因素，是认证某人成为冠军之后的那一天发生了什么。

And that's where many of the initiatives fail because the biggest predictor of whether this succeeds is what happens the day after you certify someone as a champion.

这时我们开始讨论社区或网络，这需要几个条件。

And this is where we start talking about communities or networks, and it needs several things.

我认为，像这样成功的项目，最关键的一点是有一位赞助人。

I think one of the most important thing that a successful endeavor like that needs is a sponsor.

理想情况下，需要两种类型的赞助者。

Ideally, two kinds of sponsors.

一种是更偏向社区型的，来自人力资源或人员团队，或者任何懂得如何在组织内部有效运营社区的人。

One that is more of the community type, coming from HR or the people team, or anyone who knows how to run a community inside an organization effectively.

他们负责保持人员方面的活力。

They keep the people side alive.

但我们也需要一位专业赞助者，如果公司有首席AI官的话，这位赞助者会与首席AI官紧密合作，对AI非常精通，能够引导这些冠军们朝着公司的AI采用战略方向发展。

But we also want a professional sponsor, the one who works closely with the chief AI officer, if you have one, the one that is very savvy about AI, that can also steer the direction of the champions towards the company's AI adoption strategy.

这个人会设定优先级，并连接冠军们所依赖的所有内部团队，确保冠军们专注于重要的事情，同时帮助公司为他们扫清道路，包括审批、访问权限、工具等等。

This person will set the priorities and will bridge to everyone that the champion depends on, namely all of these internal teams, and they keep the champions on what matters and keep the company clearing the path for them, the approvals, and the access, and the tools, and so on.

所以这些赞助者非常重要。举个例子，在我见过的建立内部网络最成功的公司之一，他们有一个由学习与发展专家和AI专家组成的黄金搭档，两人一起合作。

So these sponsors matter a lot, and I think that to name an example, in one of the most successful companies that I've seen in establishing an internal network, they have a dynamic duo of an L&D expert and an AI expert working together.

他们定期保持对情况的敏锐把握。

They regularly keep their hand on the pulse.

他们定期确保了解冠军们的现状、他们的需求、他们在构建什么、行业的最新动态，并始终努力弥合差距。

They regularly make sure that they understand where the champions sit, what they need, what they're building, what the state of the industry is, and always trying to bridge the gap.

反过来看，几乎在每一家我们可能有过很棒的AI冠军课程但从未转化为商业价值的公司里，缺失的部分就是后续跟进。

On the flip side, literally in every company where we might have had an amazing AI champion course that never materialized into business value, the missing piece was the continuation.

没有人负责确保这些冠军被有效利用并持续学习。

Nobody took ownership on making sure that the champions are being used effectively and learning continuously.

这意味着大多数人只会回到以前的工作方式，只是像以前那样使用AI，而不会把它提升到我们期望一个像样的项目能达到的水平。

That means that most will just return to the way they used to work before and just to use AI as they used to work and not take it to the next level that we expect a proper initiative like that will take.

有一件事很有意思，可能对某些人来说显而易见，但并非总是如此，那就是我认为企业决策者面临的一个真正挑战是，既要决定投资多少，也要决定如何正确投资于员工的人工智能培训。

One of the things that is Interesting and maybe obvious to some, but is not necessarily always obvious, is that I think a real challenge for enterprise decision-makers is both how much to invest and the right way to invest in training their people with AI.

这并不完全符合以往企业技能提升的模式，因为它更加全面和包罗万象，但同时，目前还没有大家公认的既定标准或明确的最佳实践。

It doesn't fit comfortably into previous models of corporate upskilling because it's so much more robust and all-encompassing, but at the same time, there's no established standards or clear best practices really yet that everyone agrees upon for how to do this.

正因为如此，这让人力投资在不同类别员工上的问题变得非常棘手。

And because of all that, that makes the question of how much to invest in different categories of people really hard.

这又是一个例外，无论你为这个群体提供什么样的培训，他们之后会传播其中最有用的部分，也就是对业务最相关的部分，这相对于其他领域来说，都是投资回报率极高的培训类别。

This is, once again, one exception, where whatever training you can do for this group who is going to be then disseminating the most useful parts of that, the most contextually relevant for that business parts of that, it's just an extremely high ROI category of training, even relative to other areas.

这使得它成为一个值得投入的切入点。

That makes it a valuable place to intersect.

我敢肯定，还有很多很多公司还没有进行任何正式的培训。

I guarantee that there are still lots and lots of companies out there that have yet to invest in any sort of formal training.

如果你属于这种情况，这里可能是一个很好的切入点和起点。

If you're one of those, this is a place to potentially look as a good intersection and a good starting point.

对，我同意，而且相对少数的人就能对公司产生巨大影响，相比培训所有人，这往往被淹没在太多人和太多变数中。

Yeah, I agree, and a relatively small number of people that has a huge impact on the company versus training everyone, and that often is getting lost in too many people, too many variants.

毕马威和德克萨斯大学奥斯汀分校的一项新研究发现，当人们使用AI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、评估和优化AI的输出，持续放大了AI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.

这些顶尖表现者被称为“AI放大器”，他们的定义不仅仅在于他们知道什么，而在于他们如何与AI合作。

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AI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.

这里有一个残酷的事实。

Here's a harsh truth.

你的公司可能花了几千甚至几百万美元在AI工具上，但这些工具却被大量闲置。

Your company is probably spending thousands or millions of dollars on AI tools that are being massively underutilized.

一半的公司有AI工具，但只有12%的公司用它们来创造商业价值。

Half of companies have AI tools, but only 12% use them for business value.

大多数员工还在用AI来总结会议纪要。

Most employees are still using AI to summarize meeting notes.

如果你负责公司里的AI推广，那你需要Section。

If you're the one responsible for AI adoption at your company, you need Section.

Section是一个平台，帮你管理整个组织的AI转型。

Section is a platform that helps you manage AI transformation across your entire organization.

它指导员工使用真实案例，追踪谁在用AI创造商业影响，并让你清楚看到AI在哪里创造价值，在哪里没有。

It coaches employees on real use cases, tracks who's using AI for business impact, and shows you exactly where AI is and isn't creating value.

结果呢？

The result?

你从推出工具转变为推动可衡量的AI价值。

You go from rolling out tools to driving measurable AI value.

你的员工从写会议纪要转向解决实际业务问题，你也能证明投资回报率。

Your employees move from meeting summaries to solving actual business problems, and you can prove the ROI.

别再猜你的AI投资有没有效果了。

Stop guessing if your AI investment is working.

去sectionai.com看看Section吧。

Check out Section at sectionai.com.

网址是S-E-C-T-I-O-N-A-I.com。

That's S-E-C-T-I-O-N-A-I.com.

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

Every episode, we cover the competition between OpenAI, Anthropic, SpaceX AI, Google, and Meta.

你可能心里已经有数，觉得谁领先了。

Chances are you've already formed an opinion about who's leading.

但每个AI实验室走的路都不一样，开发的技术不同，建立的合作不同，打造的生态也各有特色。

But every AI lab is taking a different approach, building different technologies, forging different partnerships, and developing a unique ecosystem.

Harbor Capital Advisors的AI实验室生态系统ETF系列，让投资者有机会投资他们看好的、最有成功潜力的AI生态系统。

Harbor Capital Advisors AI Lab Ecosystem ETF Suite gives investors a way to gain exposure to the AI ecosystem they believe is best positioned for success.

在您投资的任何地方搜索Harbor AI Lab Ecosystem ETFs，或关注X上的@HarborCapital了解更多信息。

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

请访问harborcapital.com获取包含投资目标、风险、费用、支出及其他重要信息的招股说明书。

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信息。

Read and consider it carefully before investing.

风险包括本金损失及与人工智能相关的风险。

Risks include principal loss and artificial intelligence-related risks.

Harbor ETFs由Foresight Fund Services LLC分销。

Harbor ETFs are distributed by Foresight Fund Services LLC.

Harbor与AI Daily Brief无关，这些基金也与任何AI实验室无关联、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.

市面上所有的AI编程工具，第一步做的都是同一件事。

Every AI coding tool on the market does the same thing first.

那就是开始写代码。

It starts writing code.

但Blitzy反其道而行。

Blitzy does the opposite.

在写任何一行代码之前，Blitzy会花上好几天时间，把你整个代码库逆向分析一遍。

Before writing a single line, Blitzy spends days reverse engineering your entire code base.

成千上万个智能体处理数百万行代码，梳理出每一个依赖关系、每一个没写进文档的约束、以及过去十年里做出的每一个架构决策。

Thousands of agents ingest millions of lines, mapping every dependency, every undocumented constraint, every architectural decision made over the last decade.

最终形成一张动态知识图谱，对软件的了解程度，堪比一位在系统里摸爬滚打了三十年的首席工程师。

The result is a dynamic knowledge graph that understands your software the way a principal engineer would after 30 years in the building.

其他工具靠grep搜索和markdown文件来猜测上下文。

Other tools guess at context with grep searches and markdown files.

而Blitzy从不猜测。

Blitzy never guesses.

它先建立真正的理解，然后自主交付整个软件史诗的80%以上。

It builds true understanding first, then delivers over 80% of entire software epics autonomously.

经过验证、端到端测试、生产级质量的拉取请求。

Validated, end-to-end tested, production-grade pull requests.

这就是为什么财富500强的工程团队把最重要的代码库托付给Blitzy。

That's why Fortune 500 engineering teams trust Blitzy with the code bases that matter most.

你自己去blitzy.com看看吧。

See for yourself at blitzy.com.

就是B-L-I-T-Z-Y.com。

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

我们来聊聊社区，因为我要说的第一点是，任何内部社区适用的规则在这里同样适用。

Let's talk about the community, because first thing that I will say that everything that is true for any internal community is true here.

你需要建立常规，让骨干定期碰面，并让他们轻松分享正在做的事情，一起学习，当他们基于你重视的事情取得有趣成就时，要给予他们庆祝。

You want routines, you want champions to meet regularly and make it easy for them to share the things that they're doing, to learn together, and to get celebrated when they have interesting accomplishments based on what matters to you that they will do.

另一个被低估的做法是跟骨干们的经理合作，因为我经常看到的情况是，优秀的骨干，他们的经理要么没有现实的期望，要么不知道如何用好他们，要么就是骨干和经理之间的知识差距太大。

Another one that is underrated is to work with the champions' managers, because what I often see is great champions whose managers don't have realistic expectations or don't know how to use them, or there's simply too big of a gap in the knowledge between the champion and the person who manages them.

然后大家都会非常沮丧。

And then everybody is very frustrated.

所以，给这些冠军的经理们开一个简短的会议，哪怕很短，告诉他们冠军们知道什么、他们应该做什么、如何有效地利用他们，这会大大提升你建立这些冠军所带来的成果。

So a short session, it can be very short, to the managers of the champions to explain to them what the champions know, what they're expected to do, how to effectively leverage them, creates a big difference in leveraging the output of the fact that you already established these champions.

最后，关于后续，我觉得不是每个被提名的人都会真正成为冠军，要么因为他们或他们的经理有不同的期望，要么因为有人离开、工作太忙，有些人会意识到，我见过很多次，这个角色不适合他们。

And lastly, about the day after, I think that not everybody who was nominated will actually become champion, either because they or their managers have different expectations than what they have, or people might leave, the work gets too busy, and some realize, and I've seen that many times, the role is just not for them.

听起来很好，但他们就是不想做。

It sounded very good on paper, but they just don't want to do that.

所以这个网络必须是一个非常有活力的实体。

So the network has to be a very living entity.

它需要让人们能够轻松地加入和离开，然后他们可以把它交给团队中的其他人。

It needs to enable people to go and leave the network easily, and then they can hand it over to someone else in the team.

我们希望让新人能够以清晰的方式轻松加入，而且教育永远不会停止，所以它一直在进行。

We want to make it easy for new people to join in a clear way, and education never stops, so it's always happening.

如果你已经有一个网络，但它正在失去活力，这就是你开始的地方。

And if you already have a network and it's losing energy, this is where you start.

指定你的两位赞助人，确认每位冠军仍然有一两天或足够的时间来做这项工作，如果需要的话，重新和他们的经理确认承诺。

Name your two sponsors, check that every champion still have the one or two days or sufficient capacity to do the work, and if needed, reinstigate the commitment with their managers.

另外，让那些无法完成工作的人退出这个角色，并感谢他们，同时不断提名新人。

Also, let the people who can't do the work step out of the role and thank them, and nominate new people all the time.

我见过的最疲惫的网络，往往缺少我刚才提到的这些要素之一。

The most tired networks that I see are missing one of these elements that I just called out.

所以这就是公司方面：明确的目标、合适的人选、合适的培训，以及后续的持续运作，让它成为一个有生命力的实体。

So that's the company side, the clear goal, the right people, the right training, and the day after that is also a living, breathing entity.

接下来我们谈谈冠军这一方。

Let's move to the champion side.

这里我也要回到你一开始说的那些话。

And here I'm also reverting to what you said at the beginning.

即使你不是正式的冠军，但你是公司或你所在环境中最好的人工智能用户，我认为你也应该听听这些，因为这些东西正是区分普通高级用户和那些致力于创造更多商业价值的人的关键。

Even if you're not an official champion, but you are the best AI user in the company or in your environment, I think that you should listen to that because these things are what differentiate regular power user from one that is geared towards yielding more business value.

所以，冠军需要知道一些其他员工不一定需要知道的东西。

So a champion needs to know stuff that other employees, not necessarily.

我觉得不管你是纯粹的推广者，还是全面的开发者，都没关系。

I think it doesn't matter if you are a pure ambassador or you are a full builder.

一个推广者需要理解AI用例从想法到生产，也就是人们每天实际使用的东西，所经历的整个路径。

A champion needs to understand the whole road an AI use case travels all the way from idea to production, meaning something that people actually use every day.

我觉得在构建之前有三个阶段，也就是说你要了解自己的基础，知道公司允许做什么，以及用例在哪里。

And I think that there are three stretches before the build, meaning that you know your ground, you know what the company allows to do, and where the use cases reside.

你还需要决定是否要构建、构建什么以及如何构建。

You also need to decide if, what, and how to build.

然后就是从概念验证到生产的构建过程。

Then you have the build itself from POC to production.

你如何创建代理或自动化能力？

How do you create the agent or the automation of the capability?

当然，这里的比例取决于你的技能、你被允许构建的程度，以及你拥有的工具。

And of course, here the proportion is based on the skills and how much you are allowed to build, and what are the tools that you have.

构建之后，还有价值实现、部署，以及在需要时淘汰解决方案的整个部分。

And after the build, there is the entire part of the value realization, deployment, and retiring a solution when needed.

所以我想就每一个环节给你们一些想法，这些是一个优秀的倡导者应该做好的主要环节或主要事情。

So I want to give you a few thoughts on each of them, and these are the main stops or the main things that a good champion should be able to do well.

顺便说一句，一个超级用户也应该做到这些。

So does, by the way, a great power user.

所以，我希望你们首先关注的是用例识别。

So the first thing that I want you to hone in on is the use case identification.

我希望你们找到自己领域中值得解决的问题，并且成本能证明投资是合理的。这里你们要稍微保守一点，假设投资会比预期高，而价值在第一天不一定像预期那么高。

I want you to find the problems in your domain that are worth solving and where the cost is justifying the money or the investment, and you have to be a little bit conservative here in the sense that assume that the investment will be higher than you anticipated, and that the value will not necessarily be as high on day one as you anticipated.

你们能否清楚地说明谁受这个问题困扰、多久一次、以及完成后的样子，这会大大改变局面。

Your ability to articulate who suffers from this problem, how often, and what done look like really changes the picture.

然后，把你们的想法和已有的东西对比一下，这样公司就能只构建一次，而不是重复构建同一个用例一百次。

And then checking your idea that you have against what already exists so that the company build everything just one time and not 100 times of the same use case.

作为倡导者，你们还要注意的另一件事是投资回报率。

Another thing that I want you to pay attention to as a champion is the ROI.

那就是能够清楚地说明它的价值相对于成本是多少。

That's just being able to articulate what it's worth versus what it costs.

这里的关键，就像我说的，是要做一个现实的人，这意味着你需要问很多关于它会被如何使用的问题。

The core knowledge here, as I said, is to be the realistic person, and it means that you need to ask a bunch of questions about how it will be used.

它能不能推动公司像关键绩效指标那样的东西？

Does it move a company like a critical performance indicator?

为了衡量或确保我们达成目标，它需要做什么？

What will it need to do in order to measure or make sure that we achieve the goals?

我还希望作为倡导者的你非常关注部署方面。

I also want you to be very focused as a champion on the deployment side.

我们怎么确保我们建的这个好东西能频繁地送到合适的人手里，并且以正确的使用方式？

How do we make sure that this great thing that we've built is getting into the right people's hands frequently enough and with the right usage mode?

还有一个问题是它放在哪里，谁批准它，谁在第二天支持它，而不仅仅是第一天大家都兴奋的时候。

There is a question of where does it live, who approves it, and who supports it in day two, not just day one, where everybody's excited.

最后，也是很重要的一点，就是评估。

And lastly, and that's a big important thing, the evaluation.

你怎么知道你建的东西真的很好？

How do you know that the stuff that you built is actually good?

我们怎么知道可以信任它？

How do we know that we can trust it?

我们怎么知道什么时候不能信任它？

How do we know when we cannot trust it?

我们怎么理解错误带来的风险？

How do we understand the risks of the mistakes?

而且我觉得，对任何负责人来说，最重要的技能之一就是能拍板：这个东西够不够好，能不能交给用户用？

And I think it's one of the most important skills for any champion to be able to sign off on, is the thing good enough to put in the hands of users?

我回到定义上来，想让它更具体一些，因为即使是同一个使用场景，不同类型的负责人处理方式也可能不一样。

I'm reverting to the definitions, and I want to make it more concrete because for the same use case, the different types of champions might address differently.

比如，可能有一个场景，团队每周要花好几个小时整理客户续约的摘要。

For example, maybe we have a use case where the team spends hours every week assembling a summary of a customer renewal.

对吧？

Okay?

大使的主要任务是发现这个场景，清楚地描述它，并说明理由。

The ambassador is primarily going to spot the use case, describe it clearly, and make the case.

这就是我们目前的成本。

This is what it costs us today.

这就是我认为可以解决的原因。

This is why I think it's solvable.

然后他们会把这个作为一个非常清晰的请求交给中央团队或供应商，让专业人士去执行。

And then they will take it to a central team or a vendor as a very clean request that someone professional can execute.

一旦解决方案准备好了，他们就会确保价值真正实现，确保团队正确使用它。

Once the solution is ready, then they make sure that the value is actually realized, that the team uses that properly.

所以他们在开始和结束时都会参与。

So they're introduced at the beginning and at the end.

大使构建者还会在公司允许的范围内，为自己的团队构建一个可用的版本，然后可能会有第二个团队接手，把它完全产品化。

The ambassador builder also will build the version that works for their own team inside the scope that the company allows them to build, and then perhaps a second team will take it over from them to fully productize.

当然，到了部署的时候，他们还得回来。

And then, of course, they will need to be back for the deployment.

而内部的功能开发者，他们可以一路做到最后，包括构建和其他所有环节。

And the internal FD, they can take it all the way, including the build and everything else.

不管你的负责人角色是怎么配置的，我想给你几个操作指南，确保他们负责的部分能做好，而且这对其他高级用户也有用。

Whichever your champion's role is configured of, I want to give a few playbooks to make sure that the parts that they are doing will be done well, and also this is relevant to other power users.

好，那么在开始构建任何东西之前，你首先应该问的问题是：到底该不该构建它。

Okay, so the first question that you should ask before anything gets built is whether or not you should even build that.

而且我觉得你会惊讶，有多少次人们甚至都不停下来问一句：“我们该不该构建它？”

And I think you'll be surprised how many times people don't even stop to ask, "Should we build it?"

他们就直接去做了。

They just go ahead and build.

你需要检查的是，这种情况多久出现一次？

The thing that you need to check is how often does this come up?

这是一个具体的用例吗？

Is it a concrete use case?

它值多少钱，犯错的代价有多大？

How much does it worth, and how costly the mistake is?

因为如果价值和容错率都合理，那么接下来的决定就是在复用现有的东西、购买、自己构建，或者只是把它加到待办清单里，留到将来再构建之间做选择。

Because if the value and the error tolerance justify it, the decision will then be between reuse something maybe that is already existing, buying something, building, or maybe just adding it to the backlog for building in future state.

但我觉得，个人和公司都特别倾向于自己动手做东西，所以一个好的负责人，关键是要从实际出发，结合公司的整体情况来做决定，而不是把它当成一个个人兴趣项目，而是要确保这个投入是值得的。

But I think there is such a huge bias towards building among individuals and companies that part of being a good champion is making these decisions from the grounded and the company connected point of view, not as a passion project, but something that justifies the investment.

第二点，我想提醒你作为负责人要注意的是，这个东西到底应该做成什么样？

The second thing that I wanted to make sure that as a champion you pay attention to is what should it be?

因为我觉得现在大家都觉得，什么东西都得做成一个智能体，什么都要连起来，什么都能干，等等。

Because I think that nowadays we feel that everything needs to be an agent, fully connected, doing everything, and so on.

而我认为，保持简单，或者做一个能解决问题的最简单的方案，才是正确的做法。

And I believe that keeping it simple or building the simplest solution that solves the problem is the way to go.

所以，一个好的负责人还要有判断力，知道什么样的技术适合这个任务。

So a good champion also has the mental model of what's the right technology for the job.

有时候，一个共享的聊天设置就够了，比如你们团队反复使用的提示指令和文件。

Sometimes it can be a shared chat set up that is enough, like prompt instructions and files that your team reuses.

有时候，用技能（skill）是最好的方式，是创建可按需运行的操作手册的最简洁的方法，而有时候，用自动化（automation）更合适。

Sometimes a skill nowadays is the best way to do it, the most streamlined way to create the playbooks that can run on demand, and sometimes it's an automation.

同样的步骤每次都会执行。

The same steps will be executed every time.

有时候，确实有理由为团队构建一个完整的智能体，但并不是总是这样。

Sometimes there is a good justification for it to be a full-blown agent that you build for the team and so on, but that's not always the case.

我希望你从最简单的能解决问题的方案开始，并且要认识到，从简单到复杂是有工具上的理由的，我们不是要构建我们所能想到的最复杂的东西，结果却留下一个很难构建、维护和测试的东西。

I want you to start with the simplest one that solves the problem and identify that there is a tool justification to go from left to right, and we're not just building the most complex thing that we can think of, only to be left with something very hard to build, maintain, and test.

最后，我希望倡导者们注意的最后一个要点是，在开始构建之前，先定义核心属性。

And lastly, the last playbook that I wanted to make sure that the champions are being mindful about is defining the core properties before you start to build.

所以，这些可能是你需要写下来作为需求或规格说明的东西，每一项都很有分量。

So these are the things that you probably need to write down as requirements or as spec for what's being built, and each of them carries a lot of weight.

这是为谁做的？

Who is it for?

“所有人”不是一个好答案，因为任何做过优秀用户体验设计的人都知道，为所有人设计就是为没有人设计。

Everyone is not a good answer because anyone who has done any good UX knows that designing for everybody is designing for nobody.

所以要具体明确目标受众，以及他们需要知道什么。顺便说一句，这通常是你发现需要做全新项目来整理相关知识的时候，然后才能开始执行你脑海中的想法。

So be specific about the target audience, what it needs to know, and this is, by the way, often when you discover that you have entirely new projects to do just to curate the relevant knowledge before you can even start executing the idea that you have in mind.

我还想让你问问人们会如何使用它，这样它才会真正被使用，而不是一个光鲜亮丽但没人记得网址、也不知道它绝对不能做什么的“氛围感”应用，也就是要有具体而详细的护栏。

I also want you to ask how people will use it so that it will actually be used and not a glorious vibe coded application that nobody remembers where the URL is and what it must never, ever do, AKA concrete and detailed guardrails.

顺便说一句，很多时候，先从“你的解决方案绝对不能做什么”开始思考，反而能更清楚地明确它应该做什么。

Often, by the way, starting the thinking with what your solution must never do creates much more important clarity about what it should do.

然后问问自己：谁来检查输出？

And then ask yourself who checks the output?

在哪里把人的判断纳入流程中？你又如何在业务价值和质量或安全性之间取得平衡？

Where do you enforce human judgment as part of the loop, and how do you balance business value and quality or safety of what you're building?

最后，你怎么知道它有效？

And lastly, how do you know that it works?

你要尽早确定需要测试的用例，因为一旦确定了用例，往往会引发关于边缘情况以及评估它们有多难的关键讨论。

You want to configure the use cases it needs to be tested against as early as possible because when you configure the use cases, often that surfaces critical conversations about edge cases and how difficult it can be to evaluate them.

这也有助于构建正确的东西，同时也有助于在构建之前就将其扼杀，因为无法确保质量，这种情况经常发生。

And this also helps to build the right thing and also helps kill it before it's built because of the inability to ensure quality, which is often the case.

所以这将是冠军们要交出的文件，无论是交给供应商、中央团队，还是交给自己。

So this is going to be the file that champions will hand over, whether they're handing it over to a vendor, to the central team, or even to themselves.

这是之前必须做的一项关键实践。

It's a critical practice to do before.

所以我想给你展示三个问题，我觉得这些问题能区分普通高级用户和冠军用户，或者说那些非常注重以正确方式构建正确东西的用户。

So I wanted to give you a glimpse of three questions that I think differentiate between a regular power user and a champion, or a user that is very mindful about building the right thing in the right way.

我通常把这称为构建者或冠军清单。

I'm calling that often the builders or the champions checklist.

在我们总结的时候，我觉得还有一件事我想提一下，那就是和现在AI领域的很多其他东西不同，这些手册不像其他东西那样容易过时，因为工具每个季度都会变，而成为一个好的冠军或者培养一个好的冠军网络，也包括教他们最新的工具、构建技巧，或者识别用例的技巧。

As we're wrapping things up, I think that one more thing that I wanted to observe is that unlike many things in AI nowadays, these playbooks are not as transient as some of the other things because the tools, they will change every quarter, and part of being a good champion or fostering a good champion network includes also teaching them the latest and greatest in tools and in building techniques or identifying use cases techniques.

但我特意强调了那些即使技术飞速发展也依然有效的东西，因为你仍然需要合适的人，并且有时间去做这些事。

But I deliberately spotlighted the things that remain valid even if the technology races forward, because you will still need the right people with the time to do it.

你仍然需要决定他们在宣传者和构建者之间的位置，并且把它写下来。

You will still need to decide where on the line they sit between ambassador and builder and write it down.

他们仍然需要走完从想法到生产的整个过程，而且仍然需要有人负责上线后的维护。

They will still need to get the whole road from idea to production, and somebody will still need to own the day after.

所以如果你的冠军们把这些都做好了，你就更有条件建立一个成功的流程。

So if all of these are done right by your champions, you are much better positioned to establish a successful process.

如果你自己就是冠军，我可以保证，这是公司里最好的位置之一，因为你能看到整个团队的问题。

And if you are the champion themselves, these are one of the best seats in the company, I can guarantee you, because you see the problems across the team.

你和团队一起工作，或者自己动手构建，你能学到构建的整个流程，从开始到结束，而大多数人只了解中间那部分怎么构建。

You work with the teams, or you build yourself, and you learn the beginning and the end of the road of building while most people only know the middle of how to build the stuff in the middle.

为了帮助大家正确做到这一点，我们正在设立一个AI冠军训练营。

And in order to help people do that properly, we are establishing an AI champions bootcamp.

所以这些都是我们非常明确地认为冠军需要做的事情，而且我们很快就要开始了。

So this is all very opinionated towards what we believe champions need to do, and we're starting quite shortly.

所以，如果你想成为你公司的冠军，或者你被提名为你公司的冠军，又或者你是赞助人，想送一批冠军来我们这里，我们会非常乐意培训他们所有这些内容，甚至更多。

So if you want to be a champion in your company, or you were nominated as a champion in your company, or maybe you are the sponsor who wants to send a bunch of your champions our way, we'll be very happy to train them on all of these and more.

太棒了。

Awesome.

那么回到这个角色如何演变和如何不演变的话题，我越来越多地看到一个非常重要的背景，这个背景其实潜伏了一整年，但现在变得更加紧迫，那就是一种内部全职员工，他们不一定按定义是技术人员，但已经学会了如何使用代码和智能体编码工具来解决部分工作或完成部分工作。

So going back to the idea of how this role evolves and how it doesn't evolve, one really important context that I'm seeing more, that's sort of been lurking for all year but is getting even more urgent now is it's a version of the internal FTE that's not necessarily a technical person by definition, but is someone who's figured out how to start using code and agentic coding tools to solve some parts of their work or to do parts of their work.

我认为，目前对于公司和公司内部的人来说，这是AI的力量和挑战同时存在的最佳例子。

It's the best example, I think, right now for companies and people inside companies of the power and challenge of AI all at once.

显然，对于这些人来说，能够用代码做以前做不到的工作，这是极其强大的。

Obviously, it is incredibly powerful for all these people to be able to use code to do things for their jobs that they weren't able to before.

但这会带来一大堆全新的挑战。

But it creates this massive new set of challenges.

其中一些挑战就是知道该用这些工具做什么、怎么做，甚至是怎么用那种思维方式去思考。

Some of them are just knowing what to use that for and how to do that, and how to even think in those terms.

但另一部分显然是各种安全和防护措施的问题。

But a second part is obviously all the sort of security and guardrails issues that come up.

而且这种情况越来越多，不仅仅是因为人们用错了，而是因为智能体功能强大且执着，往往很快就会突破限制，渗透到各种不同的系统中。

And increasingly, that's not just because people are doing it wrong, but because the agents are powerful and tenacious enough that they tend to escape containment pretty quickly and kind of weave their way into lots of different systems.

所以我确实认为，在这个领域，冠军的角色会变得更加关键，因为这不仅仅是“嘿，我们还没有充分利用这个机会”。

So I actually see that as an area where the role of champions is going to become even more mission-critical because it's not just, "Hey, we're under-optimizing this opportunity."

而是角色本身在发生根本性的变化，带来了所有这些新挑战，我们必须想办法解决这些挑战，才能继续前进。

It's the roles are changing fundamentally and creating all these new challenges, and we simply have to figure out how to solve all those challenges so we can move forward.

所以我认为，这种思维方式和这种角色，对公司来说只会变得越来越重要。

So I think that this type of mindset and this type of role is going to become just nothing but more important for companies going forward.

我同意，与其因为担心非专业开发者搞建设而把他们关掉，不如提供支持和基础设施，甚至培训更多的领域专家，因为他们会专注于正确的用例。

I agree, and instead of shutting them down because you're concerned that less professional builders are building, do the enablement and the infrastructure and even train more of the subject matter experts because they will work on the right use cases.