了解是什么让 Intercom 抛弃了它的操作手册——保罗·亚当斯（Intercom 首席产品官）

嘿。

Hey.

这里是产品体验节目，我是兰迪·西尔弗。

It's the Product Experience, and I'm Randy Silver.

多年来，Intercom 一直是业界关注的焦点，尤其是在如何正确做产品方面。

For years, Intercom has been the company to watch in terms of how to do product the right way.

也就是说，采取深刻而周到的方法，同时带来巨大的商业价值。

That is to have great thoughtful approaches that also result in lots of business value.

观察他们如何应对人工智能带来的颠覆——或者称之为加速，一直非常有趣。

It's been fascinating to watch how they've responded to the disruption or call it acceleration that's been brought on by AI.

今天，我们邀请到他们的首席产品官保罗·亚当斯，来了解他们如何应对过去几年的挑战，以及 Intercom 的未来规划。

We're joined today by their chief product officer, Paul Adams, to learn about how they've handled the last couple of years and what comes next for Intercom.

本节目《产品体验》由 Mind the Product 提供，属于 Pendo 家族的一部分。

The Product Experience podcast is brought to you by Mind the Product, part of the Pendo family.

每周，我们都会与来自全球的杰出产品人士对话。

Every week, we talk to inspiring product people from around the globe.

访问 mindtheproduct.com 以回顾往期节目，并获取帮助您提升产品实践的免费资源。

Visit mindtheproduct.com to catch up on past episodes and discover free resources to help you with your product practice.

了解 Mind the Product 的会议及其优质的培训机会。

Learn about Mind the Product's conferences and their great training opportunities.

你好，保罗，欢迎来到《产品体验》节目。

Create a free Hi, Paul, welcome to the Product Experience.

今天过得怎么样？

How are you doing today?

很好，谢谢。

Good, thanks.

感谢邀请我。

Thanks for having me.

很高兴能来这里。

Glad to be here.

今天很高兴有你加入，我们马上就开始深入交流。

It's great to have you here today and we're gonna get stuck into the chat in just a second.

但在我们开始之前，如果你能简单介绍一下你自己、你在产品领域的工作以及你是如何走到今天的，那会非常好。

But before we do, it'd be great if you could give us a quick intro to who you are and what you're doing in product and how you got there.

是的。

Yeah.

我在Intercom工作。

I work at Intercom.

我领导我们的产品团队。

I lead our product team.

我从很早以前就在这里了，已经十一年了，超过十年了，这听起来有点让人害怕。

I've been here since very early days over long well, eleven years, over a decade, which sounds kind of terrifying in a way.

但这段经历很棒，如今更是令人惊叹。

But it's been great, and it's it's amazing these days.

我们正身处一场人工智能革命之中，这场革命不仅在我们周围发生，也发生在公司内部。

We've AI revolution happening around us and in inside the building.

所以这非常棒。

So that's been great.

在那之前，我做过各种各样的产品相关角色。

Before that, I was in all sorts of kinda product roles.

我曾在Facebook工作。

I was at Facebook.

再之前，我在Google工作。

Before that, I was at Google.

我曾经住在旧金山，那地方既疯狂又有趣。

I lived over in San Francisco, which is wild and fun.

而且，我算是个万能手。

And yeah, I'm kinda jack of all trades.

我当过产品经理、设计师、研究员，样样都懂，样样都精。

I've been a PM, designer, researcher, jack of all trades, master then.

所以在一家初创公司待了十一年。

So eleven years in in a startup.

我记得你加入Intercom的时候公司只有13个人，所以这期间亲眼见证公司的发展和演变一定非常有趣。

And I think it was there was 13 people in Intercom when you joined, so that must be really interesting seeing the company change and evolve over that period of time.

是的。

Yeah.

真的已经完全认不出来了。

It's unrecognizable, really.

而且，十年了，你知道，当一家公司有了十年历史时，它显然经历了起起落落和不同的发展阶段，比如在目标和追求方面等等。

And, yeah, ten years, you know, when you're when a company is a decade old, it's obviously gone through highs and lows and different kinda journeys, you know, in terms of the goal of what you're trying to achieve and so on.

所以，是的，我们内部已经多次重新调整过。

So, yeah, we've and we've reconfigured internally many times.

你知道，我们改变了软件的开发方式。

You know, we we have changed how we build software.

我们改变了市场推广的方式。

We've changed how we go to market.

就像我说的，过去几年真的非常令人兴奋。

Like I said, it's been like really exciting the last couple of years.

斯科特，过去两年实际上是我在这里期间公司最令人兴奋的两年之一，因为这一切都得益于这一波新技术的兴起。

Scott, last two years actually have been two of the most exciting years in the company in all of my time here because of what's enabled by the start of this new technology wave.

我觉得这非常令人振奋和激动。

I find it incredibly energizing and exciting.

有太多机会，同时也有焦虑和恐惧。

There's so much opportunity and and anxiety and fear too.

这种不确定性带来了这些感受。

The uncertainty brings that.

但我真的很享受这个过程。

But I'm loving it.

只要你享受其中，我相信你带给团队的活力以及对这些变革的热情，也会让其他人感到兴奋。

Well, as long as you're loving it, I'm sure the energy that you're bringing to the rest of the team and the the passion for all of these changes is exciting everyone else as well.

我记得Intercom曾经是运用‘待完成任务’理论进行营销的经典案例，当时你们的着陆页内容让每个人都津津乐道。我很想知道，如今公司发生了怎样的变化，你们又是如何把握过去几年AI浪潮的。

I do remember Intercom was, I was saying to Randy earlier, Intercom was one of the kind of classic examples of the use of jobs to be done in marketing when your landing pages were like, everyone So was talking about yeah, I'm interested to see now kind of how the company's changed and how you've kind of ridden this wave of AI that has happened over the last couple of years.

所以，这正是我们今天要讨论的主题，但你们也对商业模式和产品定价方式做出了巨大调整。

So, and that's kind of the topic of what we're gonna be talking about today, but also you've made a huge change in your business model as well and how you price the product.

在我们深入探讨组织和流程变革之前，先跟我们讲讲这个转变是如何发生的？你们是如何从Intercom传统的按座位收费模式，转向这种基于价值的定价模式的？

So before we get stuck into the sort of organizational changes and the process changes, tell us a little bit about how this change happened, where you went from the sort of typical seats based pricing for Intercom and more to this value based pricing.

是的，挺有意思的，年纪越大，你越会意识到这一点。

Yeah, it's funny, like the older you get, the less you realize, you know.

你知道，关于我们在那里做的用户任务，这事儿挺搞笑的。

You know, like, it was funny to your jobs we've done there.

我明白了，其实并没有唯一正确的方式去开发软件。

I've learned that, like, no, there's no right way to build software.

也没有唯一正确的方式去进入市场。

There's no right way to go to market.

过去，我有点像某种意义上的强势者。

In the past, I'm kind of like some definition of strong.

我相信持有坚定的观点，但要保持开放。

Like, I believe in strong opinions weekly held.

过去，我有时是持有强烈观点，并且非常固执。

Sometimes in the past, I think I've been strong opinions, strongly held.

这就是正确的方式，这就是唯一的方式。

Like, this is the way, this is the right way.

你知道，我们做过的项目比用户画像更好，我通过惨痛的教训意识到这并不一定成立，不同的事情在不同公司、规模和目标下各有优劣，诸如此类。

You know, jobs we done is better than personas and I realized the hard way that that's not necessarily true and different things are better and worse depending on the company, the size, your goals, all sorts of stuff like that.

所以我努力强迫自己对这些事情保持开放的心态，我觉得当ChatGPT出现时，我就开始谈论这些事了。

So, I try and force myself to be very open minded about all of these things and I think when ChatGPT showed up, you know, I I I talk about these things.

很多人把它们看作是单向门时刻。

Many people do as these like one way door moments.

你知道，一旦你穿过那扇门，就回不去了。

You know, you kind of once you go through the door, you're not going back.

iPhone就是这样一个转折点，它开启了移动技术浪潮。

The iPhone was one that kind of kicked off this kind of mobile technology wave.

ChatGPT则是开启了我们如今处于早期阶段的AI浪潮。

ChatGPT is the one that kicked off the AI wave that were in the early years of now.

当你穿过这些门并意识到自己已经跨过去时，一切似乎都变得不确定了，我觉得是这样。

And when you go through these doors and you realize you've gone through the door, everything was up for grabs, I think.

因此，我们在Intercom早期就意识到并真正接受了这一点：这是一场新技术浪潮的开端。

And so, one thing we did very early in Intercom is realize and internalize that this is the beginning of a new technology wave.

这场AI浪潮很可能规模巨大。

This AI wave is probably going to be very big.

尚不确定，会有多大？

TBD, how big?

仍然不确定，会有多大？

And still TBD, how big?

但肯定至少和移动互联网一样大，可能和互联网一样大，甚至更大。

But certainly as big as mobile, maybe as big as the Internet, and maybe bigger.

所以在ChatGPT发布后的一周内，我们就彻底推翻了我们的路线图，改变了战略，重新思考了所有过去我们认为是最佳实践或行业标准的东西。

So within the space of a week of ChatGPT, we ripped up our road map, we changed our strategy, we started to rethink everything that we used to think was the best practice or an industry standard or whatever.

我们这么做是因为非常明确、很早就意识到，客户支持——这是我们产品所服务的领域——将成为首批受到巨大冲击的行业之一。

And we did that because it was very clear, very early that customer service, which is what we build a product for, was going to one of the first industries being hugely, hugely disrupted.

因为大型语言模型在出厂时就已经非常擅长客户支持人员所做的一系列任务。

Because LLMs were already out of the box, very good at a lot of the tasks that customer service people do.

所以所有东西都被重新审视了，你知道，过去几十年来主导SaaS行业的座位模式也面临被颠覆的可能。

So everything got put on the table, you know, and the seats model that has dominated SaaS for the decade or so that SaaS has existed was up for grabs too.

你知道吗，我们当时想，也许这不是最好的定价方式。

You know, we we we were like, maybe that's not the best way to price.

如果我们目标是让尽可能多的人使用这种他们不理解、无法评估、甚至不知道该如何评估的新产品的话。

And if our goal is to get as many people using this new type of product that they don't understand, that they can't evaluate, they don't know how to even think about evaluating it.

我们可以采用一种有利于我们业务增长的定价方式。

We can price it in a way that is optimized for our business growth.

我们必须采用一种有利于他们试用和使用的定价方式，如果他们从中获得了价值，就会继续使用。

We gotta price it in a way that's optimized for them trying it and using it And if they get value from us, they'll keep it.

如果他们没获得价值，那就说明我们的产品不够好。

And if they don't get value from it, we don't have a good product.

所以，是的，我们十八个月前推出了Fin，也就是我们的AI代理，从第一天起就根据解决的问题数量来定价。

And so, yeah, we launched Fin, which is our AI agent, eighteen months ago, and from day one, priced it based on resolutions.

如果客户的查询得到了解决，你就被收费。

So if if a customer query is resolved, you get charged.

如果没有解决，就不收费。

And if it's not resolved, you don't.

我喜欢你采取了这种做法，保罗，因为人们有时会认为他们的商业模式是理所当然的。

I love that you've taken that approach, Paul, because, you know, people sometimes think that their business model, they are kind of they think they're owed their business model.

你知道，音乐行业长期以来是以单曲为基础的，后来转为专辑模式，接着又因为iTunes和流媒体等平台而再次去中介化，现在的收入模式已经完全不同了。

You know, like, the music industry for a long time was singles based, then became album based, and then it became disintermediated again with with iTunes and then streaming and things like that, and it's a totally different revenue model now.

所以当你开始讨论将Fin转向按价值收费时，这在Intercom内部至少让一些人感到有争议。

So when you started having the discussions about going value based with Fin, that had to be a little contentious for at least some people at the table at Intercom.

销售团队并不习惯这种方式，对吧？

Sales isn't used to working that way, are they?

那是什么样的情况？

What was that like?

你是怎么应对的？

How did you approach it?

是的。

Yeah.

我知道你说得对。

I know you're right.

音乐是一个非常好的类比。

Music is a great a great comparison.

就像，很多人会损失很多东西。

Like, people have a lot to lose.

很多人会损失很多东西。

A lot of people have a lot to lose.

比如，你知道，在音乐行业或任何这类颠覆性时期，目前有很多讨论，比如ChatGPT现在可以生成图像。

Like, you know, in the music industry or any of these kinda disruptive periods, like, there's a lot of discussion at the moment about, you know, ChatGPT can now do image creation.

很多人在问，嘿。

A lot of people are asking like, hey.

它在使用我的内容。

It's using my content.

你知道吗？

You know?

这类似于音乐领域的Napster时代，突然之间，人们开始意识到，等等。

And it's similar to kind of the Napster era of music where, you know, suddenly it was like, wait.

谁在拿钱？

Who's getting paid?

我还在拿钱吗？还是谁在拿？你知道，所以这里既有模式，也有价格。

And am I getting paid still or who, you know, so there's kind of the model and then there's the price.

像Spotify仍然会支付音乐艺术家报酬，但很多人会说，钱很少，很多时候都是主要给大艺术家之类的。

Like Spotify still pays music artists but a lot of people would say, not very much a lot of the time or you know, the kind of mostly when it goes to the bigger artists or whatever.

我对这个了解不够深，没法有强烈的意见，但这里既有模式，也有价格，当然，我们有一个销售团队，他们整个职业生涯都在卖座位。

I don't know enough about that to really have a strong opinion but there's there's the model and there's the price and so, certainly here, like we've a sales team that has spent their whole career selling seats.

他们知道怎么做。

They know how to do it.

这很熟悉。

It's familiar.

突然间，他们要面对销售一种前所未有的模式的产品。

And suddenly they're faced with selling a product based on a model that's never existed before.

这不仅新，而且还不熟悉。

Not only is it like new, you know, it's also unfamiliar.

比如，有时候感觉就像，嘿，从单曲到专辑的转变，其实还是差不多的东西。

Like, sometimes it's like, hey, we're going probably, maybe like the move from singles to albums was like, well, it's still kind of the same thing.

十首歌代替一首歌之类的，你知道的。

10 songs instead of one or whatever, you know.

这是一种完全不同的思考价值和价值交换的方式。

This is like an entirely new way to think about value and value exchange.

所以这对团队来说非常困难，我非常理解他们，因为他们不仅要学习如何销售和解释一个新模型，而且支撑产品的技术——

So it was very it's been very difficult for the team and I feel for them because not only are they trying to learn a new model to sell and explain, but the technology underpinning the product, I.

嗯，大语言模型和人工智能——也在迅速变化。

E, LLMs and AI, is also rapidly changing.

所以他们必须同时应对这两方面的挑战。

So they're trying to grapple both both sides at the same time.

这确实非常困难。

So it's been it's been hard for sure.

我们不得不大力加强销售支持，同时在营销方面开发工具，比如投资回报率计算器，或者帮助人们理解他们如何从产品中获得价值的方法等等。

We've had to lean in a lot to like sales enablement and and also build tools on the marketing side like ROI calculators or ways that people could see how they might derive value from the product and so on.

这是一段持续的旅程。

It's an ongoing journey.

我们还没有完全弄清楚这一点。

We have not yet fully figured this out.

我本来想问的是，你觉得你做出了正确的决定吗？

That was gonna be my next question is, you know, do you feel like you made the right decision?

比如，你有没有达到一个阶段，觉得这真的行得通，或者至少现在在起作用，但还是处于早期阶段？

Like, did you get to a point where you were like, actually, this is this is gonna work or this is working, but is it still early days yet?

这绝对是正确的决定。

It's a 100% the right decision.

好吧，我不能说100%。

Well, I can't say a 100%.

我之前说过，我对一切持开放态度。

I said earlier that, like, I'm open till all things are not.

我非常有信心，这是个正确的决定。

I have high confidence it's the it's the right decision.

我觉得我们肯定不会再回到按座位收费的方式了。

Like, I don't think we're going back to seats for sure.

也许我们会转向一种新的模式，但我还不确定具体会是什么样子。

Maybe we'll move to a new like, you know, what where where I'm uncertain is like, maybe a new model will emerge that's a little bit different.

也许它仍然是基于成果的，但运作方式会有些不同，你知道的。

Maybe it's still outcome based, but the ways in which it works are kinda different, you know.

所以我认为事情肯定会发生变化，现在仍然处于早期阶段。

So I think things for for sure are gonna evolve, And so it is still early.

你知道，这 definitely 仍然处于早期阶段。

You know, like, it's definitely still early.

目前大多数销售的软件仍然是按座位收费的 SaaS，所以这 definitely 仍然处于早期阶段。

The majority of software sold is still seat based and SaaS, so it's certainly still early.

但我对 AI 产品会基于它们为使用者创造的价值来销售这一方向充满信心。

But I I'm I'm high confidence that it's the right path that AI products will be sold based on the value that they create for the people who use them.

我只是觉得这是一种更好的方式。

I just think it's a win it's better.

这只是一个双赢的模式。

It's just a win win model.

我认为人工智能是一种非常独特的技术，因为你本质上是在销售工作成果。

And I think that AI is a very, very kind of unique technology in that it, you're basically selling work.

你知道，在基于座位的SaaS产品中，你卖的是座位。

You know, like in a in like a SaaS product with seats, you're kind of selling the seat.

比如说，如果这是一个销售工具，我们有10个销售人员。

So, like, if it's a sales tool, we have 10 salespeople.

我们再增加一个第11个销售人员。

We add an eleventh salesperson.

是的。

Oh, yeah.

我有了第11个座位。

I have an eleventh seat.

所以你实际上是在销售给单个用户的使用许可。

So you're kinda selling a license for a single user.

10个销售人员，10个席位，十一个，十一个。

10 salespeople, 10 seats, eleven, eleven.

而在这里，你卖的是要完成的工作。

Whereas here, you're selling work to be done.

你知道，这可能有点像任务，也存在某种关联，但本质上你卖的是要完成的工作。

You know, maybe it's sort of like jobs you've as well as a connection there, but it's kinda like you're selling work to be done.

这是一种完全不同的模式。

And that's like just a different model.

这是一种完全不同的思维模式。

It's a different mental model entirely.

所以我不确定是否有人已经真正搞明白了，但这一点肯定会不断发展。

And so I don't know that everyone anyone's cracked it yet for sure, and it'll evolve for sure.

但我认为，我们出售的是工作，这一理念是新型AI产品中人们需要真正理解的核心部分。

But I think the idea that we're selling work is a foundational part of new AI products that people kinda need to wrap their head around.

这到底意味着什么？

And what does that actually mean?

人们的工作和角色也在变化。

People's jobs and roles change too.

所以这并不是一对一的对应关系。

So it's not it's not like for like.

我们看到很多客户服务人员，我们的AI代理Fin承担了大量一级支持工作。

You know, we see a lot of people in customer service where Fin, our AI agent, does a lot of the tier one support.

因此，客户代表只需要处理最简单、最基础的问题。

And so the customer rep that's, the easiest kind of frontline support, the easiest questions.

然后，好吧。

And then, okay.

我们不再需要人们来做这些工作了，因为Fin做得更好、更快、更准确。

We we don't need the people to do that anymore because Fin does it better, quicker, faster, more accurate.

人们本来就不喜欢这份工作。

People hated that job anyway.

对大多数客服团队来说，这其实是好消息，因为他们一直被海量工作压得喘不过气。

It's kind of, like, good news for most customer support teams are drowning in volume.

但他们会把员工重新分配到新的工作类型上，比如，有人需要设计Fin的用户体验。

But then they redeploy people to new types of work, like, hey, someone needs to design the Fin experience.

有人需要确保我们的知识库是最新的，因为如果过时了，Fin就会给出过时的答案。

Someone needs to, like, make sure our knowledge is up to date because if it isn't, Fin will give out of date answers.

所以工作内容也在变化。

So the jobs change too.

然后，你知道，像工具的管理员这样的角色，还需要一个席位吗？

And then, you know, genius, genius, does the, like, admin person of the tool need a seat or no.

这是免费的。

It's free.

对他们来说是免费的，但我们只对解决问题收费。

It's free for them, but we just charge for the resolution.

这对公司的增长产生了什么影响？

And how has it impacted the company's growth?

你们是否不得不暂时停顿一下，知道需要一些时间来理清这些事情，或者，你们通过这种方式是否推动了业务增长？

Have you had to kind of take a bit of a standing position for a while, knowing there's gonna be a bit of figuring this stuff out, or, you know, has it resulted in more growth in the business by taking this approach?

是的。

That's yeah.

这是个很好的问题，因为我认为许多企业、许多SaaS产品都面临这样的困境。

That's a great question because I think a lot of businesses, a lot of SaaS products face this dilemma.

我们承担了风险。

We've taken risks.

在产品发布五天后就彻底推翻原有策略，全面押注AI，这在当时确实是个大胆的举动。

Well well, like, ripping up your strategy and going all in on AI five days after the product launch is kinda a risky thing to do at the time.

我们还做了其他事情。

We did other stuff too.

我们彻底重做了网站。

We ripped up our website.

大约一年前，我们上线了一个新网站，按SaaS行业的标准来看，这简直疯狂。

We launched a new website about a year ago that was, like, kind of nuts by SaaS standards.

SaaS网站都长得一模一样。

Like, SaaS websites all look the same.

它们都有一个产品截图，底部有个行动号召，顶部是价值主张。

They have a product screenshot and, you know, call to action at the bottom and value prop at the top.

而我们推出的东西却是以艺术为基础的。

And we launched something that was based on art.

而且在首屏上方根本没有产品截图。

And it was like it had no product screenshot above the fold.

它更像是对一个人凝视未来城市的创意艺术表达。

It was like a creative artistic expression of like a man looking into a futuristic city.

带有一种乌托邦式的乐观和未来主义氛围。

Kinda utopian optimism, futurism vibes.

对于一个SaaS网站来说，你根本不会这么做，这根本不是常规操作。

Which for a SaaS website, you that just you don't that's not done.

你知道，这根本不是人们会做的事。

You know, it's not kind of a thing you do.

所以我们一直在承担巨大风险，因为我们拥有这样的意愿和文化，认为这样做是必要的。

So we've been taking big risks, and we've got kinda appetite and culture because we think it's necessary to do it.

我认为所有SaaS公司都面临这个困境，无论他们是否已经投身其中。

I think all SaaS companies face this dilemma whether they've leaned into it or not.

对我们来说，结果非常好，但其中有一些细微差别。

For us, it's worked out really well, but there's a nuance to that.

结果非常好，他们都说Fin是我们增长最快的产品。

It's worked out really well, and they're like, Fin is our fastest ever growing product.

如果你把Fin单独作为一个独立业务来看，它取得了巨大的成功，增长非常迅速，真的非常出色。

If you take Fin alone as a, like, stand alone business, it's tremendously successful and really fast growing, like, really, really amazing.

我们增长最快的产品，就像我之前说的。

Our fast ever, like I said.

对此我们非常满意，但我们认为市场仍然 largely 未被开发。

We're very happy about that, and yet we think the market is still largely untapped.

所以有着巨大的增长机会。

So there's huge, huge growth opportunities.

但与此同时，我们确实还有一个基于旧有产品的业务。

But meanwhile, we do have this like older seats based business.

还有很多人在使用人工客服。

And lots of people still do human support.

所以它仍然很重要，而且极其关键。

So it still matters and it's still tremendously important.

但它不可能像Fin业务那样快速增长。

But it's just not going to grow nearly as fast as the thin business.

这根本不可能做到。

It it couldn't possibly.

世界不会突然涌现出大量客户代表，但会涌现出大量AI代理。

The world isn't suddenly gonna, like, explode into, like, more customer service reps, but it will explode into more AI agents.

所以我们有两个业务在并行运行。

So we have both businesses running in parallel.

Seats业务增长较慢，虽然发展不错，但速度较慢，而Fin则实现了爆炸式增长。

The seats one is slower growing, so it's growing well, but slower, and Fin has exploded.

所以对我们来说，这进展不错，但总有一天，两者之间可能会出现相互蚕食。

So it's working out well for us, but at some point, there's gonna be cannibalization there probably.

我们目前还看不到太多，但显而易见的是，AI做得越多，能解决的问题就越多，理论上所需的客服人员就越少，尤其是当你不再为人们在Fin中承担的新角色（比如系统设计等）收费时。

We don't see a ton of it yet, but, obviously, like, the more AI does, the more resolutions it can do, the fewer seats you need in theory, especially when you don't charge for the new types of roles that people are doing in with Fin, like designing the system and so on.

我们会看到的。

So we'll see.

到目前为止还不错，但这确实是个冒险的决定。

It's good so far, but it was a risky thing to do.

保罗，你在这件事上遇到了很多挑战。

Paul, you've had a lot of challenges with this.

我们稍后会聊聊现在团队如何用AI进行开发，但我很好奇，你们是如何从一开始就为AI设计的。

We'll talk in a minute or two about how the teams build with AI now, but I'm curious about building for AI first.

Intercom以众多不同的框架而闻名，你们多年来一直谈论这些框架，以及各种成熟的方法论。

Intercom is well known for lots of different frameworks that you guys have talked about over the years and different ways of thinking and mature ways of approaching the practice.

我想你们一定得推翻很多东西，重新学习很多东西。

And I'm wondering, you've had to probably had to rip up a whole bunch of things and unlearn things.

我猜你们为Fin设定的核心指标，和你们基于座位的服务以及其他一些人工服务的指标已经大不相同了。

I'm guessing your North Star metrics are are very different for Fin than they are for your seat based things and some of your other person.

那么，根本方法上有什么变化？

So what what's changed in the just the fundamental approach?

是的。

Yeah.

所有那些事情都发生了。

There are all those things.

你说得对。

You're right.

所有那些事情都变了。

All those things have changed.

我正在想，到底还有什么是一样的，这听起来可能对别人来说很疯狂。

I'm trying to think is anything the same actually, which might again might sound crazy to people.

我不确定。

I'm not sure.

当我们思考这个世界时，我们想到的是过去两年里诞生的新公司。

When we think about this world, we think about the new companies that have been born in the last two years.

我们拿他们来和自己做比较。

That's who we compare ourselves to.

所以，如果你是一家AI初创公司，可能会有很多公司，比如Cursor、Lovable、Bolt，这些全是AI编程工具，还有Replit，过去两年里涌现了大量令人惊叹的公司。

So, like, if you're an AI startup, you know, it could be many companies, Cursor, Lovable, Bolt, all AI coding tools, like this, you know, Replit, like loads of amazing companies have been founded in the last two years.

我们问自己，他们是怎么运作的？

And we ask ourselves, what how did they work?

他们认为什么是重要的？

Like, what do they think is important?

因为我们竞争的对象就是他们，不是那些公司本身，而是客户服务领域、客户体验领域的AI初创公司，AI客户代理是一个新兴的类别。

Because that's who we're competing against, not those companies directly, but, like, AI startups in the customer service space, customer experience space, you know, AI customer agent is a is a a new emerging category.

这领域非常火热。

And it's gonna be and it's and it's really hot.

这显然是一个非常明显的、无需多想的投资方向，如果你是风投，就会投资这些不同类型的公司。

Like, this is a it's a no kind of an obvious no brainer place to invest money if you're a VC and fund all these different types of companies.

可能会有少数几个大赢家，也可能有很多小赢家。

There'll be probably a few big winners, maybe many smaller winners.

所以它充满了机遇，我们把自己与这些公司作比较。

So it's got loads of opportunity and we compare ourselves to those companies.

我认为这真的、真的很重要。

And I I think that's really, really important.

我们不会把自己和Intercom的旧版本比较，因为那现在毫无意义。

We don't compare ourselves to the prior versions of Intercom because it's meaningless now.

顺便说一句，我不得不学会这一点。

You know, I've had to learn this by the way.

我喜欢把好的流程写下来，你知道的。

Like, I love I love a good process written down, you know.

曾经有很多白板、漂亮的图表、箭头和方框，但如今这些几乎都不复存在了。

Many a whiteboard, beautiful diagram, arrows and boxes, and, you know, and yeah, so very little of it exists anymore.

唯一保留下来的是我们的原则。

One thing that's persisted is our principles.

因此，我们有关于如何构建的原则。

So we have principles for how we build.

我非常推崇原则。

I'm a big, big proponent of principles.

我认为，如果你能掌握撰写原则的艺术，原则就是指导你行为的准则。

I think if you can master the art of writing principles, you know, principles are are guidelines for how you behave.

所以，如果你能掌握这些准则的艺术，你就能过上非常高效的生活。

So if you can master the art of these guidelines, yeah, you can kinda live a very productive life.

我认为，原则对你的个人生活和工作生活同样重要。

I think principles are excellent for your personal life just as much as they are for your work life.

我们的原则一直保留了下来。

Our principles have persisted.

例如，我们的研发团队——也就是产品工程团队——有四项原则。

So for example, we've had we have four principles in in our r and d team, which is our product engineering team.

其中一条是：从问题出发。

One of them is start with the problem.

我们最近改成了‘关注问题’，但我不太喜欢这种新说法，我还是会用我们原来的说法：从问题出发。

We've recently changed to sweat the problem, which is I'm not sure I like the new language, but I'll I'll use our old language, start with the problem.

我认为在以人工智能为先的世界里，这仍然非常重要。

I still think that's really important in an AI first world.

事实上，这可能更重要，因为这些新工具让任何人都能非常快速地开发软件。

In fact, it's probably more important because these new tools let let anyone build software very quickly.

我以前不会编程，连HTML都不会，但现在我能搭建网站和应用了。

I built I can't code like HTML back in the day kind of thing, and I can build websites and apps.

我目前同时开着Replit和Lovable，正试着为乐趣搭建一个新的网页。

I have both Replit and Lovable open in front of me at the moment, and I'm trying to build a new website page for fun.

嗯，勉强能用。

Well, it's workable.

它做不到。

It can't do it.

我们走着瞧。

We'll see.

但我可以试试。

But I can do I can try.

所以，真正比什么都重要的是，这件事本身重要吗？

So what actually matters more than anything is, does this thing even matter?

也就是说，这是一个值得解决的问题吗？

Like, is this a problem worth solving?

我认为我们会看到一波公司爆发，真正的大规模爆发，新的公司、新的应用，其中许多会失败，因为它们没有解决真正重要的问题。

So I think we're gonna see an explosion of companies, like a real explosion of companies, new companies, new apps, and many of them will fail because they're not solving a problem worth solving.

你可以把这一点分解成不同的方面，比如这个问题有多大？

And there's like, kinda you can break that down into different things like how big is the problem?

是否存在一个庞大的市场、高能量的问题，等等。

Is there a big market, high energy problem, and so on.

我认为挑选一个好的问题有很多细节，但我觉得它们会因为这个原因而失败。

I think there's like a lot of detail to picking a good problem, but I think they'll fail because of that.

相反，许多以前根本不可能出现的公司，将会解决人们真正关心的真实问题，而以前根本没人有动力去做。

And then conversely, like, many companies that would have never existed before will solve real problems that people really care about and just nobody was motivated to do it before.

也许它有点小众。

Maybe it's bit niche.

而现在，它突然成了一个绝佳的解决方案。

And now suddenly, it's gonna be a great solution for it.

多亏了人工智能和AI工具。

Thanks to AI and AI tools.

所以，是的，这些原则一直延续了下来。

So, yeah, so principles have persisted.

所以这些原则一直延续了下来。

So the principles have persisted.

就像你说的，你已经彻底抛弃了所有流程，以一种完全不同的方式工作。

Like, what you said you've ripped up all of your processes and you work in a completely different way.

你是怎么工作的？

How are you working?

一团混乱。

Chaotity.

好的。

Okay.

所以根本就没有流程。

So there's just no process.

有一点混乱。

A little bit of chaos.

我觉得有点压力其实是好事。

I think a little bit of gas is good, by the way.

我们是怎么工作的？

How are we working?

就像Intercom一样，过去十年里，很多公司都逐渐形成了四种最佳实践。

So, like, Intercom, like many, many companies over the last ten years, we kinda make four the best practice kind of emerged.

我最早是在Facebook看到的。

I first saw it at Facebook.

当我还在谷歌时，工程团队规模非常庞大。

When I was at Google, the engineering teams were gigantic.

比如我在做YouTube和Gmail时，有时候一个下午会聚集二十到三十名工程师。

Like, I worked on YouTube and and Gmail and you'd have, like, 1PM would be over like twenty and thirty engineers at times.

你知道，你可能只有一个设计师，或者幸运的话，半个设计师，这种比例。

You know, you might have one designer or half a designer if you're lucky, that kind of ratio.

当我去Facebook时，情况非常不同，团队要小得多，通常会有一个产品经理、一个设计师、一个工程经理，以及五到八名工程师左右。

And when I moved to Facebook was very different, the the teams are much smaller, so you'd have like a PM, a designer, an engineering manager, and maybe somewhere between five, seven, eight engineers.

我认为这种模式我们后来在Intercom也采用了，并尝试了各种其他方法，但这种模式一直延续了下来。

And I think that pattern we certainly then did that at Intercom and trialed different things, but that that pattern kind of persisted.

所以每个团队大约有五到六名工程师，组成一个三人核心组。

So you have somewhere like five to six engineers per team, team at a triad.

这个三人核心组就像团队的顶层，他们一起选择要解决的正确问题，设计良好的解决方案，构建优质的产品，等等。

The triad are like a team at the top and, you know, together they pick the right problems to solve, design good solutions, build good solutions, and so on.

然后这些团队会被组织成若干小组。

And then the teams were organized into groups.

所以你会有一个产品团队负责三个领域，形成一个小组；而在Intercom，当我们发展到一定规模时，大约有四到五个这样的小组。

So you have like a product team for three areas and there's a group there and then you might have an Intercom we got to the size where we'd have like four groups, four or five groups.

所以你基本上是在看大约20个产品团队。

So you're kinda looking at 20 product teams basically.

你知道，如果你十二个月前加入我们，那时我们的结构就是这样。

You know, if you've came in our doors twelve months ago, that's how we were set up.

这些都不再存在了。

None of those things exist anymore.

我们已经彻底打破了三角团队和固定团队，现在以一种更加灵活、流动的方式工作。

We've blown up triads, blown up teams, and we're now working in a way more fluid, flexible way.

所以我们现在的组织方式几乎是所有人都在一个庞大的池子里，我们有工作流，也就是说，我们会选择产品中需要开发的某个领域。

So the way we've set ourselves up now is almost everyone's kind of in a big giant pool, and we have work streams, so it's kind of like we pick an area of the product that needs to get developed.

而且，很多这类做法其实都很有名。

And, again, like a lot of this is famous.

这个产品还处于早期阶段。

It's in its infancy as a product.

所以有很多探索性的工作要做。

So it's a lot like exploration to do.

在AI产品开发中，我们进行的实验比以往要多得多。

With with AI product development, there's a lot more experimentation than we would have done historically.

这些做法真的有效吗？

Just just do these things work?

我的意思是，谁知道呢？

I you know, who knows?

所以我们有工作流。

So we've have work streams.

每个工作流都有一个直接负责人。

The work stream has a DRI, like a directly responsible individual.

每个工作流都由一个人全权负责。

So there's a single person in charge of a work stream.

然后，这个工作流在需要的时候，可以随时调配所需的任何资源。

And then that work stream has whatever resources they need at whatever given time they need them.

因此，人们可以在不同工作流之间流动。

And so people can move between work streams.

有些工作流可能侧重于设计，因为这属于早期探索阶段。

Some of those work stream could be design heavy because it's like early exploration world.

有时候可能是工程密集型的。

Sometimes it could be engineering heavy.

人们会根据需求流动。

People move around based on the need.

我们现在就是这样安排的，不再有试运行了。

That's how we're set up now, and there's no trials anymore.

我们已经削减了大量繁重的流程。

We've cut out huge overhead.

并不是所有事情都顺利。

It's not all working.

情况变得更好了。

It's working better.

我们最近在管理团队中对此进行了反思。

We're reflecting on this recently in the leadership team.

情况确实更好了，但还不完美。

It's working better, but it's not perfect.

比如，仍然存在一些问题和协调挑战之类的事情。

Like, there's still issues and coordination challenges and things like that.

我们正在以一种完全不同的方式开发软件。

We're building software in an entirely different way.

保罗，你很久以前谈到过，我认为是六周、六个月、六年作为一个组织原则，这是你方法中经典的一点。

One of the the classic things of of your approach is, Paul, you talked a long time ago about, I think it was six weeks, six months, six years as an organizing principle.

我猜，当你考虑这个问题时，六年这个时间框架很可能已经被抛在脑后了，因为这个时间跨度太长，根本无法有任何把握。

I'm guessing six years is probably thrown out the window at the moment when you're thinking about this, because that's way too long a time frame to have any sense of confidence in.

你还在用这类时间尺度来思考这些团队或这些群体吗？

Are you still thinking in terms of those kinds of horizons with the with these teams or with these groups?

是的。

Yeah.

你说得对。

You're right.

六年这个时间框架已经被抛在脑后了。

That's six years out the window.

这毫无疑问。

That's for sure.

不过话虽如此，唯一的例外是，我认为你仍然需要一个产品愿景。

Now have having said that, like, the only caveat there is I do think you still need a product vision.

我们确实有一个清晰的产品愿景。

Like, we do have a clear product vision.

我们认为，大多数客户沟通将由AI代理完成，客户服务领域将扩展为客户体验和客户成功。

We think that the majority of customer communication will be done by AI agents and the world of customer service will expand into customer experience and customer success.

所以我们确实，你知道的，当你去的时候，我可以一溜烟就完成。

And so we do, you know, we I could whiz when you go.

我们确实有一个清晰的产品愿景。

We do have we do have a clear product vision.

所以我们大致朝着某个方向模糊地前进。

So we're kinda like hand wavy heading to a certain place.

但那之后，就是两个月了。

But then after that, it's two months.

你知道，我们计划和思考的范围有多远？

Like, you know, how far are are we planning and thinking?

两个月。

Two months.

我们五月在伦敦有一个活动，届时将发布一些新产品。

We have an event in London in May where we're gonna launch some new product.

我知道那里会有什么。

I know what's gonna be there.

我们在七月赢了。

We've won in July.

他们原本计划在旧金山举行。

They were planning for probably San Francisco.

截至今天，我大致知道在那里会发布什么。

And as of today, I roughly know what we'll announce there.

然后我们在十月赢了，谁知道呢？

And then we've won in October and who knows?