这位25岁青年如何零法律经验打造出价值6.75亿美元的法律AI初创企业

人工智能正以超乎寻常的速度持续发展，这意味着我们也必须跟上步伐。随着我们在整个法律软件栈中不断深入探索，我们发现软件与服务之间的界限正变得模糊。我认为这正是我们公司的一大优势——我们承认无法精准预知未来走向，但你们同样如此。因此，让我们携手合作，确保无论形势如何变化，双方都能成为赢家。

AI is continuously developing super, super quickly, and that means we need to do the same. We're finding that as we go deeper and deeper and deeper in the entire legal software stack, we're also seeing that the line between software and service is blurring. I think that's been one of our strengths as a company to say, we don't know exactly where the future is going, but neither do you. So let's work together to make sure that we're both winners in whatever happens.

今天与我对话的是马克斯·朱斯特兰德，他是Legora公司的首席执行官兼联合创始人。Legora参与了2024年冬季项目，作为领先的人工智能工作平台，致力于帮助律师和法律专业人士提升工作效率。欢迎你，马克斯。

Today, I'm joined by Max Junestrand, and he's the CEO and cofounder of Legora. Legora was in winter twenty four, and they are the leading AI workspace helping lawyers and legal professionals do their work. Welcome, Max.

嗨，谢谢，古斯塔夫。

Hey. Thanks, Gustav.

距离你们完成那批项目已经过去十三个月了。这一年你们确实非常忙碌。确实如此。

It's been thirteen months since you did the batch. It's been a really busy year for you. It has.

感觉那已经是很久以前的事了。过去这一年让我觉得自己老了五岁。

It feels like it was a really long time ago. I feel like I've aged five years in the last one.

为不太了解的听众介绍一下Legora吧。

For those who don't know, tell us about Legora.

好的。

Yeah.

你们具体在打造什么产品？

What are you guys building?

在Legora，我们正在构建面向律师的人工智能工作平台。本质上，我们正在彻底改变他们完成工作的方式——从文件审阅、文书起草到法律研究等全流程。长期以来，法律科技领域存在严重的软件碎片化问题，充斥着大量单一功能解决方案，而人工智能始终无法有效处理非结构化文本、判例和法律文件。直到GPT-3.5问世，这才彻底改变了游戏规则。

At Legora, we're building the AI powered workspace for lawyers. We're essentially transforming the way that they complete their work. Everything from reviewing, drafting, researching. Essentially, within legal, you've had this incredibly fragmented software space where there was a lot of point solutions, and AI was never good enough to actually work with unstructured text, precedent, legal documents. And when GPT 3.5 got out, that just completely changed the game.

因此我们迅速构建了概念验证原型，如今已将其扩展为服务于数万名律师的企业级系统。

So we were quick to, you know, build a POC, and then now we've scaled that all the way to an enterprise grade system serving tens of thousands of lawyers daily.

没错。那些单点解决方案本质上就是工作流工具。那么它们之前是什么呢？因为法律科技行业在这之前已有历史渊源，并非刚刚起步对吧？

Yeah. And those point solutions were basically workflow tools. So what were they before? Because it's been a history of an of a legal technology industry that existed before. This is not started right

现在？不。我是说，法律科技作为一个分类已存在很久，但长期以来我觉得它缺乏吸引力。以前基本上就是各种单点解决方案，从合同模板工具——用来将合同条款标准化——

now. No. I mean, legal tech has been a category for a long time, but it was really unsexy for a long time, I think. And you'd essentially have a broad range of point solutions, everything from templating tools where you would sort of codify a contract

是的。

Yeah.

到专门的翻译工具、修订标记工具或研究工具。所有这些都以某种方式处理文本，而生成式AI的出现彻底颠覆了局面。当尘埃落定后，你就能清晰地看到如何用同一套底层技术解决大量这类用例。

To special translation tools or redline tools or research tools. And all of them work with text somehow, and generative AI came into the game and just kind of threw up everything off the off the table. And then when it landed, you very clearly saw how you could solve a lot for a lot of these use cases with the same underlying tech.

所以ChatGPT大概比你们创立公司早了八个月？描述下那个时刻，这对公司创立是个重要节点吗？

So ChatTPP came maybe eight months prior to Yeah. You guys starting this company. Describe that moment. Was that an important moment for the company's founding?

我们早在ChatGPT之前就在探索AI+法律领域了，当时用的是谷歌推出的早期Bert模型。它们在英语表现尚可，但在瑞典语中简直糟糕透顶。真正促使公司成立的契机，是其中一位联合创始人的律师朋友——

We were playing around in AI and legal way before Chateapiti, and we were using these early models from Bert coming from Google. Yeah. They were decent in English, but they were just horrendously bad in Swedish. And, you know, the first observation that kind of sparked the founding of the company was one of the co founder's friend who was a lawyer

嗯。

Yeah.

曾花了整个夏天为大型律所整理法院案例。当我们看到GPT-3.5向开发者开放后，立即开始构建。我记得第一个产品是股票期权解读器，能解释股票期权合同——

Spent four months during a summer just summarizing court cases for a big law firm. We basically saw that GPT 3.5 was released to developers, started building. I think the first thing that we built was a stock option reader that would explain how a stock option contract

实用吗？

Worked useful?

没错。作为没有法律背景的初创者，这很合理。但很快我们就转向思考：如何打造一个全覆盖的端到端系统，让每个法律从业者都愿意日常使用。首款产品其实很简单——尤其在欧洲开发，你得费尽周折才能符合各种数据处理要求。

Right. You know, as startup founders with no legal background, that was seemed reasonable. And then very quickly, the sort of focus changed to how do we build this more wall to wall or or end to end system that every legal professional wants to work with on a daily basis. And the first product was really quite simple. Especially building for Europe, you gotta go through a lot of hassle to kind of conform with all the data processing requirements.

所以，你知道，所有数据都托管在欧洲境内，不用于训练，不保留，当你看Azure和AWS的架构时，还能豁免人工审查。我们算是跳过了所有这些障碍，直接构建了一个符合律师事务所使用规范的系统。然后很快，随着ChatGPT、Claude、Gemini等通用AI平台的持续发展，我们必须构建的系统标准也在不断提升，要求越来越高。

So, you know, all data hosted within Europe, nothing for training, no retention, exemption from human review when you look at the way Azure and AWS is structured. And we we kind of jumped through all those hoops and just built a system that was compliant for law firms to work with. And then very quickly as the general sort of AI platforms continue to develop with ChatGPT, with Claude, with Gemini, the requirements for what we had to build to be much, much better know, continuously increased.

在某些行业或领域，比如编程或法律，这些模型的表现简直像魔法一样。它们能做到那些行业前辈们想都不敢想的事情。你能描述一下第一次用Legora为客户实现魔法般体验的场景吗？他们当时是如何反应的？

In some industries or some some categories like coding or law, for example, it seems like the models are just magical. Like like, they do things that the people that were in those industries before could not even imagine be possible. Could you describe sort of like the first time you used Legora to do something that was magical for a customer and and and how they experience it?

可以。我想第一次是在我们将Legora部署到北欧最大律师事务所Mannheimer Swartling的时候。他们的管理合伙人曾在报纸上有句名言，说AI‘人工有余而智能不足’，那是早期的事了。

Yes. I think the first time was when we deployed Legora into the biggest law or largest law firm in The Nordics, Mannheimer Swartling. Yeah. They their managing partner had a famous saying in the newspaper that AI was more artificial than intelligent, which was back from the early It's true.

可能吧，我是说，在当时的情况下。

Probably, I mean, at the time.

是啊。很多律所花大价钱买了根本解决不了问题的工具。我带着笔记本走进会议室，直接让他输入一个查询。他输入了一个法律研究问题，而我们已经通过RAG系统将Ligura与瑞典立法完美对接。

Yeah. I mean, a lot of firms burnt themselves buying expensive tools that didn't solve anything. And I came into that meeting, you know, I bring up my laptop and I just ask him, you know, put in a query. And he puts in this legal research query and we've tied Ligura to Swedish legislation with a rag system. Perfectly.

你能从他眼神里看出来——就是那个顿悟时刻。现在我们...

And, you know, you kind of see it on his eyes, like, it's the moment. Now when we're

那也是你的顿悟时刻吗？

Is that your moment as well?

不。我个人顿悟时刻是第一次用ChatGPT的时候。那种感觉就像完全置身科幻世界——你能和电脑对话，它还会回应。作为创业者，你很快就能意识到：好，我们可以这样应用在这个领域，那样应用在那个领域。

No. I think my personal moment was just using chat GPT Yeah. Generally. Right? Like, it was amazing.

就法律领域而言，聊天体验一直很酷。但当我们用相同模型实现不同应用时，最早做的用例之一是尽职调查——你要审阅数百份文件。我们不是逐份查看，而是做了个大表格，每份文件对应一行。

It felt complete sci fi that you could talk with the computer and it talked back. And, you know, as an entrepreneur, you you you kind of quickly, you know, from that you understand that, alright, we can apply it in this space in this way and in that space in this other way. And think I for legal specific, the chat experience, I think was always cool. But when we took the same models and sort of applied them differently, one of the first use cases we did was due diligence where you have hundreds or, you know, a lot of documents that you want to review. And instead of going through them one by one by one, we just made this large grid where essentially every document represented a row.

然后把查询条件放在列里。比如你导入100份雇佣合同，问‘是否都包含保护公司知识产权的IP条款？’系统就会开始扫描并回答：是，是，是...

And then you could put your queries in the columns. Right? And as you then put in, you know, a 100 employment agreements and you ask, does all of them include an IP clause where the company protects its intellectual property? And it just starts to ravel and it goes, yes. Yes.

是。是。不。不。不。

Yes. Yes. No. No. No.

是。是。是。是。而且它总是能链接回引用源。

Yes. Yes. Yes. Yes. And it always links back to the citation.

你会意识到，天啊，这简直是变革性的。它把过去需要数天或数小时完成的任务，缩短到了几分钟。

You realize, like, holy shit. This is transformational. It's taking tasks which used to be, you know, days or hours, and it's turning them into minutes.

等到这个节目播出时，你们应该已经宣布完成了B轮融资。你们融了多少钱？

By the time this is is live, you will have announced that you have raised a a series b. How much did you raise?

我们融资了8000万美元，由Iconic和General Catalyst领投，同时也感谢Weiss的持续参与，以及Benchmark和Redpoint的支持。

We raised $80,000,000 led by Iconic and General Catalyst, and, you know, grateful for Weiss's continued participation as well as Benchmark and Redpoint.

这个软件具体是怎样的？作为一名律师，对，使用Legora，我的日常工作会是什么样子？

What is the the software like? So so as a lawyer Yep. Using Legora, what does my day to day look like?

它主要分为两部分。第一部分是网页应用，第二部分是我们的Word插件。我们直接集成到Microsoft Word里。好，先说网页应用，我们最初做的就是一个简单的聊天功能，可以基于你自己的文档和文件进行对话。

So it's really broken up into two pieces. The first one is the web application, and the second one is our word add in. So we integrate directly into Microsoft Word. Right. So if we start with the web application, the first thing that we had was just a simple chat, a chat over your own documents and files.

是的。这个功能很快发展成了一个自主代理，能够使用应用内的许多其他端点以及外部工具

Yeah. This is quickly developed into its own agent that's able to use a lot of the other end points in the app and also external tools

嗯。

Yeah.

来解决更复杂的、需要分步完成的工作流程。

To solve more complex sort of step by step workflows.

嗯。

Mhmm.

所以你可以想象这样说，嘿，我想写一份备忘录。第一步是出去做些研究。嗯。第二步是将所有研究内容转化为公司标准用语。

So you could imagine saying, hey. I want to write a memo. And the first step of the memo is to go out and do some research. Mhmm. The second step is to take all that research and conform it into the standard language of the firm.

第三步是撰写报告，最终产出就是一份报告。

And the third step is to write the report, and then output is a report.

它能完成所有这些吗？

And does it do all that?

完全可以。哇，没错。我们稍后可以详细讨论，但MCP以及扩展工具使用的方式——

It does all that. Wow. Right. And I think we can talk more about it later, but MCP and the way that you can scale the tool usage of

这些

these

智能代理正是我极其热衷、我们重点投入的领域。因为不同律所在采用工具解决特定工作流程时有不同需求——知识产权、重组、公司业务或争议解决领域的需求各不相同。除了聊天功能外，第二部分就是我之前提到的网格系统，我们称之为表格审查。

agents is something that I'm super, you know, keen on and and that we're leaning very heavily into. Because a lot of firms have different needs in terms of how they want to adopt the tools to solve for their specific workflows. And it's different if you work in intellectual property or if you work in restructuring, or if you work in corporate, or if you work in disputes. The second piece outside of the chat is, well, the grid that I talked about before. We call it tabular review.

对。本质上是输入任意数量文件，再输入任意数量查询，然后进行交叉比对。重大创新不在于如何提示模型，而在于如何实现规模化运行——如何同时并行处理十万条查询并确保零错误、引用准确。

Yep. It's essentially input any number of files, and then input any number of queries, and we sort of cross run that across each other. And the big innovation there does not really come from, you know, how do you prompt and and work with a model, but it's how do you make this run at scale? Right. You know, how do you run a 100,000 queries in parallel at the same time and make sure nothing breaks, all the citations are correct.

嗯。由于文件可能非常冗长，需要进行大量分块和文档内检索。对法律文件而言，必须始终包含定义等细节内容。

Mhmm. There's a lot of chunking, sort of rag searching within the individual documents because sometimes they're very, very long. Yep. And with legal docs, there are certain intricacies where you need to always include things like the definitions. Yep.

条款间可能存在交叉引用。综合这些因素构建了网格系统。看到'add in'这个词，我觉得可以形容为'律师版光标'。

And there might be cross references within each clause to each other. So taking all of that into consideration, that kinda serves the grid. Looking at the word add in, I think you could phrase it as cursor for lawyers.

律师基本上用Word。那就是他们的工具。

Lawyers basically use word. That's that's the They

用Word。

use word.

对吧？这就像是一个长久以来众所周知的事实。

Right? This is a, like, a known fact for a long

是的。我是说，他们用Word或PDF格式起草和审阅合同。我们真正想做的是类似于Cursor那样，如何将生成式AI引入法律专业人士现有的工作环境？这意味着要与Word集成。

time. Yeah. I mean, they draft and they review contracts in word or PDF form. And what we really wanted to do is similar to cursor, how do we bring generative AI into the existing work environment of a legal professional? And that means integrating in Word.

现在的区别在于，你不能分叉Word，也不能占用所有你想要的界面空间。你基本上只能适应这种右侧栏的布局。

Now the difference is you can't fork Word, and you can't take up all the real estate you want. You're basically conformed to this sort of right hand column.

嗯。

Yep.

然后你必须非常有创意。这几乎就像设计一个移动应用，因为这就是你能用的所有空间。我们首先构建的是一个助手或聊天功能，不仅能阅读文档，还能进行编辑。比如你可以说，我希望你为买方重新协商这份MSA，并使用我内部的检查清单或类似的操作手册或先例。

And then you gotta get really creative. It's basically like designing a mobile app almost because that's all the real estate you get. And the first thing that we built there was just, you know, how do we integrate an assistant or a chat that's able to not only read the document, but also create edits. So you might say, I want you to, you know, renegotiate this MSA for the buyer Yep. And do that using this internal checklist that I have or this internal sort of playbook or precedent.

现在我们已经将其扩展到不仅能在聊天基础上工作，还能处理更复杂的工作流程。比如你可以说，这是一份合同，我希望你按照我的操作手册，包含20个不同的步骤，确保我们从起始位置进行谈判，并包含不同的备选方案。

And now we've scaled that to not only work in a chat by chat basis, but also more extensive workflows. So you can say, here's a contract. I want you to take my playbook that consists of 20 different steps and make sure we negotiate from the starting positions and have, you know, different fallbacks included.

你能举一个具体的例子吗？比如几年前律师完全无法做到的事情，现在可以做到了？

Do you have a specific example of something that was impossible a couple years ago for a lawyer? Like, literally, you couldn't do it, and now you can do it.

是的。我觉得有很多例子。早期的机器学习模型在处理法律语言时表现很差，尤其是当不同文档中的语言看起来不一样时，它们表现得更糟糕。

Yeah. I mean, I think there's a lot of it. Right? The early ML models were really bad at legal language. And what they were really bad at was when the language looked different in across documents.

对吧？你可以训练一个系统来查找，比如说，控制权变更条款，如果它在所有文档中看起来都一样的话。是的。但说实话，如果类别看起来不一样，它真的很难找到控制权变更的含义。

Right? You could train a system to find, let's say, a change of control clause if it looked the same way across all the documents. Yep. But it was really, frankly, bad at finding the meaning of a change of control if the class didn't look that way.

嗯。

Mhmm.

所以大型语言模型让我们能够处理的任务，特别是在大型合同和大型文档提取方面。我们如何从中提取见解？另一个例子就是修订标记。在Word中根据先例或操作手册对文件进行修订标记，以前完全不可能。或者进行深入研究，跨越数百或数千份判决，你不仅需要核对判决，还需要将立法和法规等内容整合到同一个地方。

And so what the LLMs have allowed us to do is to just take tasks where especially on, like, large contracting and large document extraction. So how do we pull the insights from this? Another one is just, you know, redlining. So redlining files within word against a precedent or playbook, completely impossible. Or take deep research across, you know, hundreds or thousands of judgments where you need to conform not only the judgments, but also pull in things like legislation and regulation all into the same place.

是的。由于智能成本下降，我们也增加了可以进行的查询数量。对吧。所以一个很酷的事情是，嵌入，对你的文档和文件进行一次搜索，再在网络上进行一次搜索，再对法庭案例、判决和立法进行一次搜索，然后将所有这些结合起来，实际上就像创建一份备忘录那样。

Yeah. Since the cost of intelligence is going down, it also increases the amount of queries we can do. Right. So one pretty cool thing is, you know, embedding, making one search against your own documents and files, making another one on the web, and making another one against court cases and judgments and and legislation and combining all of it to create effectively like a memo that

也许他们过去负担不起这样做。不。当然不。他们只是没做。

Maybe they they couldn't afford to do in the past. No. Course not. They just didn't do it.

不。同样地，对于尽职调查，如果你回溯到很久以前，它曾经是一个物理的数据室。是的。这就是为什么它被称为一个房间。对吧？

No. And similarly with with due diligence, when if you go way back, it used to be a physical data room. Yeah. That's why it's called a room. Right?

你过去会进入那个房间，里面有所有的文档和合同，然后你会坐下来通读所有内容。是的。而且你还得用笔做标记。所以对一家公司进行尽职调查真的很昂贵。嗯。

You used to go into the room, you had all the documents and all the contracts, and then you'd sit down and read through all of them. Yeah. And you had to mark them with a pen. So making and doing a due diligence on a company was really expensive. Mhmm.

而现在它几乎变成了一种商品

And now it's becoming almost a commodity

嗯。

Mhmm.

人们期望你去做，但客户也不太愿意为非常简单的合同审查支付费用。嗯。当他们知道AI可以完成99%的工作时。哇。

Where you're expected to do it, but clients are also not really that excited to pay for very simple contract review Mhmm. When they know that AI can do, you know, 99% of it. Wow.

是的。在我任职YC期间，我们确实资助过一些法律软件公司，但它们面临的最大挑战都是向律师事务所销售产品。

Yeah. So in in the time that I've been at YC, we have funded some legal software companies, but the hardest challenge for all of them was selling to law firms

没错。

Yeah.

向法律行业销售时，大多数公司最终会选择企业客户，因为律师事务所简直无法攻克——但这种情况在两年前发生了根本转变。能否分享一下你认为是什么发生了变化？你们是如何成功向世界顶级律所推销的？

And selling to legal like, most of them would end up selling to companies because law firms were just, like, not possible to sell to. That radically changed just, like, two years ago. Yeah. Can you tell us sort of like what do you think changed, and how do you do it when you go and sell to one of the major law firms in the world?

对于听众朋友们，我记得面试时你也特别强调过这个问题。当时我们持相反观点，坚持说'这次不一样，请相信我们'——现在看来我们是对的。

So for everybody listening, this was also one of the questions that I remember you pushing really hard on during the interview. And I think we were quite contrarian to say, you know, no. It's different this time. Trust us. Yep.

我们的解决思路始终是'你们赢我们才赢'。通过利益绑定来传达：这项技术正在革新行业，律所终将需要采用某种形式的解决方案，而我们希望成为长期合作伙伴。

I'm glad we were right. I think the way that we approached the problem was always with this idea of we win if you win. So let's align our incentives with saying, as a law firm, this technology is revolutionizing. You're going to need to adopt it in some sense, shape, or form. And we want to be that long term partner.

而他们似乎也意识到了这点。

And somehow they know that.

关键在于，很多法律服务的差异化程度很低。无论是X律所还是Y律所做的尽职调查，结果大同小异。当服务同质化达到完美平衡时，若有创新者打破现状，客户会迅速转向——毕竟客户也面临价格压力。

Well, so what happens is a lot of legal work is low differentiation. You know, if you're doing a DD from, you know, law firm x or law firm y, kinda getting the same deal. And so when you have this perfect equilibrium of services and somebody disrupts that by taking a new approach, clients are quick to switch. Yep. I mean, clients are under price pressure.

他们需要高效解决方案，而律师费又极其昂贵。一旦平衡被打破...嗯...你几乎是被迫要采用新技术，而且有充分动力这么做。

They want to be effective. Legal fees are very high. And so if this equilibrium breaks Mhmm. You are almost forced to adopt it. And you you are incentivized.

这就像当年律师们接受电脑的过程。

It's kind of the same as in lawyers adopted computers.

正是。按小时计费时，你可以让人跑去图书馆查案例找判例，也可以直接按Ctrl+F检索——结果显而易见。

Right. If you're billing by the hour, you could say, well, let's have a person walk to the library. You know, find the right book. You know, find the right cases or the right precedent and use that for whatever work we do. Or you press control F, right?

始终存在这样一个困境：你希望以最佳方式服务客户，因为这能长期带来更多收入。对于我们合作的许多律所而言，他们的品牌声誉、信任度——始终将客户放在首位——才是最重要的。因此许多律所也希望成为这个领域的领导者。对吧？你明白吗？

There's always this dilemma of you want to serve your client in the best way possible because that drives you more revenue over time. And for a lot of, a lot of the firms that we work with, their, you know, brand reputation, trust as always putting the client first is what matters the most. And so a lot of the firms also want to be leaders here. Yep. You know?

有些律所想成为快速跟进的第二行动者，但更多希望成为先行者，因为他们明白——如果你处于完美平衡状态，而某项简单业务被颠覆时，通过更快下沉就能获得更多市场份额。但这并非一场竞相压价的比赛，对吧？是的。问题在于，如果我们...

Some of them want to be fast second movers, but many want to be first movers because they're understanding if you have this perfect equilibrium and you take a simple type of work that gets disrupted, you should get more market share by moving down quicker. But then it's not a race to the bottom, right? Yeah. It's a question of, okay, if we

每个国家基本上都有律所排名体系。

take- Every country has a ranking of law firms basically.

没错。而且这也不是价格上的恶性竞争。因为如果你降低了——比方说尽职调查的成本，就能腾出更多时间与董事会商讨复杂的并购交易。通常最终结果是：你在时间压力下本可以处理更多工作，但总有堆积如山的待办事项。

Yeah, right. And it's also not a race to the bottom in terms of pricing. Because if you pull down, let's say the cost of a due diligence, you free up more time to spend with the board on advising them on, you know, a really complex merger or a really complex acquisition. And so what typically ends up happening is you're under time pressure. You could do more work, but you just have all this stuff that needs to get done.

而这正是AI的强项。但它也能以极具创意的方式服务律师。比如我们有用户案例：某人来电说他在角色扮演游戏中与Legora模拟辩论，要求AI扮演对方当事人。很厉害吧？

And that's what AI is really good at. But it's also serving lawyers in very creative ways. I mean, we've had use cases where, we you got a call from somebody and they say, I played a role role playing game with Legora, trying to win this argument, and I'm asking it to act as the other party. Right? Wow.

还有更精彩的——西班牙佩里托尔卡律所的合伙人出庭时，将所有对方证据文档输入Liguora，在听证会上实时查询。当对方律师陈述时，他能立即发现漏洞并打断。他形容得很精妙：'上战场时，拥有Legora就像多穿了一层盔甲'。

There was this amazing situation that one of the Spanish partners at from called Perithiorca had where he went into court. Mhmm. He had put all the evidence and all the documents from the opposing party in Liguora, and he was actively querying it during the hearing and during you know, at the time when the other attorney was speaking because then he could immediately interrupt if he found something that was that was wrong. And he phrased it very nicely. He said, when when he goes into the battlefield, having Legora is like having another piece of armor.

我觉得这个比喻非常富有诗意。

And I thought that was that was very poetically.

能用Legora替你进行谈判吗？

Could you use Legora to do negotiation on your behalf?

可以。我们的设计思路是：目前纯大语言模型还不足以胜任。虽然模型会持续进化，但关键问题在于——功能边界该划在哪里？

Yeah. So the way that we built that is I think the LLMs by themselves are not good enough for that yet. And we can talk about that, but it's it's interesting to to build these products knowing that the models will get better. Yeah. And where do you stop?

没错。每个功能都要权衡取舍。这个功能我们称为'战术手册'——本质上是包含审批规则的集合。

Yeah. Right? On every feature. But so so that feature in the Gore is called playbooks. A playbook is essentially a collection of rules where you either approve or disapprove something.

可以说，在YC这里签署保密协议时，你总希望协议中的定义条款符合特定标准。因此你提供规则和示例文本，然后表示：如果对方不接受这个定义，我们还有备选方案——备选一和备选二。你只需在Agora中打开文档。

So you might say, for the way that you would sign NDAs here at YC, you always want the definition within a confidentiality agreement to look a certain way. So you provide the rule, you provide some example language, And then you say, alright. If the opposing party will not accept this definition, we have some fallbacks. So fallback one and fallback two. And you just open a document in Agora.

你打开操作手册点击播放，系统就会逐条核对合同条款并标注差异。没错。如果内容与手册不符——

You open the playbook, and you say press play. And it goes through every rule and runs it against the contract, and it marks it up. Yeah. So if it does not conform with your playbook

嗯。

Mhmm.

它会提供建议文本使其合规。最妙的是这个系统不仅限于法务部门使用。在Legora，每位销售代表在将NDA提交法务团队前都用它来协商条款。我们刚开始与北欧某大型银行合作——

It gives you the suggested language so that it will. And the really cool thing about this is it scales outside of just legal departments. Mhmm. So at Legora, every sales rep is using Legora to negotiate NDAs before sending it to our legal team. And we just started working with this very large bank in The Nordics.

系统使用范围很快从法务部门扩展到合规、风控，现在连销售团队也在用。因为所有人都能借助这个系统。它不仅更快更准确，还能帮助建立统一标准。

And it's very quickly moved from, you know, the legal team to compliance, to risk, and now to sales. Wow. Because everybody can leverage the system. And the cool thing about it is it's not only faster and and more accurate, but you agree on a standard. Mhmm.

法务团队制定操作手册后，这就成了所有人的执行标准。实际上它持续提升了工作质量和一致性。

Because the legal team then creates the playbook, and that becomes the standard that everybody So it actually increases quality and consistency over time.

YC新一期孵化项目正在招募。心怀创业梦想？请访问ycombinator.com/apply提交申请。行动永远不嫌早，填写申请的过程就能提升你的创意。

YC's next batch is now taking applications. Got a startup in you? Apply at ycombinator.com/apply. It's never too early, and filling out the app will level up your idea. Okay.

回到正片。

Back to the video.

你们初创时都不是律师出身。但现在却打造了全球发展最快的法律AI公司之一。这是如何做到的？

None of you guys, when you started, were lawyers. No. So you still are building one of the largest or fastest growing legal AI company in the world. How do you do that?

现在我觉得自己算是个业余律师了。我们的成功关键在于极度谦逊——承认自己不了解行业。我们快速与早期合作伙伴建立关系，保持每日反馈。这种'我们和你都不知道未来确切方向'的坦诚，反而成为公司的优势所在。

I think at this point, I've become a hobby lawyer. But how we approached it was being incredibly humble. Humble for the fact that we did not know the industry. We were quick to create relationships with our early partners where feedback was, you know, happening daily. And I think that's been one of our strengths as a company to say, we don't know exactly where the future is going, but neither do you.

所以让我们共同努力，确保无论发生什么，我们都是赢家。我认为现在我们有幸聘请了大量律师加入团队，他们直接与产品团队和客户合作，尤其是在这个经历巨大变革的行业中。带着更多天真（如果可以这么说）进入是有益的，比如问：为什么事情要这样做？你知道吗？其实可以换种方式。

So let's work together to make sure that we're both winners in, you know, whatever happens. And I think now we, of course, have the privilege of having hired a ton of lawyers into the team that work directly with the product teams and directly with the customers, especially in an industry that is now going through such big change. It was useful to come in with more naiveness, if you will, saying, why does it work this way? You know? It could work this way instead.

假设你是个正在观看的创始人，你想为物流、保险或金融领域开发AI软件。你的建议基本上是‘你不需要我的任何专业知识’吗？那该如何学习你需要了解的东西呢？

Let's say you're a founder watching this right now. You're like, I wanna build AI software for logistics or for insurance or finance. Is your advice basically you don't need any of my expertise? How do how do how do you I think my How do learn about the things you need to learn, though?

我的建议是，去学习它们。我们进入这个领域时，我做的第一件事就是采访了100名律师。我在LinkedIn上用了个小技巧——发消息问能否共进午餐，我愿意按他们的时薪付费。

I think my advice is, you know, learn about them. Right? Like, we we went into this, and the first thing I did was I interviewed a 100 lawyers. I had this good hack on LinkedIn. I I texted them asking if we could have lunch, and I would pay their hourly rate.

其实我根本负担不起。但没人会真的收钱，他们都说‘太棒了，我很乐意免费和你吃饭’。

And I I could def you know, definitely not afford it. Yeah. And none of them would, you know, impose that. They would just say, oh, that's amazing. Like, I'll have the lunch with you anyways.

我职业生涯中一个非常有益的特质是：我总能让别人愿意帮助我。嗯，我觉得这是被严重低估的

One of the attributes that have been very helpful in in my career has been that I'm I'm I'm somebody people want to help. Mhmm. I think that's a very underrated

确实。

Yep.

没错，这是种能力。有些方法可以培养这种特质：你可以更无畏地接触他人，同时保持极度感恩的心态。

Yep. Skill. I think there are there are things you can do to be more like that. You can be a bit fearless in in your approach to people, and you can also be very, very Grateful. Grateful.

心怀感恩

And grateful

完全同意。

Absolutely.

并且要感激他人给予的帮助。是的，如果没有这样做，我们不会有今天的成就。

And appreciative of the work that other people help you with. Yeah. If we hadn't done that, we would not be where we are today.

那么当你初创公司时，如何与律师共进午餐呢？毕竟你对法律了解不多对吧？

And then how do you conduct a lunch with a lawyer when you're starting a startup you know not that

你会这样安排：选个像样的地方，毕竟他们收入很高。我花了不少时间才明白不同部门的运作方式天差地别——比如交易律师的工作模式与公司法务部门的律师完全不同。

much about law? So you'd sit down like this. You'd go to somewhere decently nice because, again, they make a lot of money. And it took me some time to even understand that the way that departments work are fundamentally different. Like a transactional lawyer works nothing the way a lawyer within the corporate department works.

你只需抛出大量问题。我觉得也要适当回馈——比如我会主动联系，他们看到我的技术背景后，我会分享些让他们觉得'这个很酷'的见解，再问'你觉得这个怎么样？'

You just ask them a ton of questions. And I think also giving them something back. So, you know, I'd reach out. They see my, you know, tech background, and you try to be you know, give them nuggets of, oh, that's really cool. What do you think about this?

其实就是提供创意点子，让他们产生给你建议的欲望。

Like, you give them ideas. You make them engaged in wanting to give you advice.

嗯。确实，人们通常很乐意给创业者建议。这本来就是该善加利用的资源。

Mhmm. And yeah. And people generally feel good giving giving founders advice. Of course. Like, it it's like something that you should take advantage of.

没错。现在能从我们所处的位置做这些事，我真的很高兴。

Yeah. Yeah. And something that I'm, you know, really happy to do now from the position where we're at.

法律科技领域有些大公司，你们是要挑战所有对手吗？如何看待现有法律科技市场？

There are some large companies in in in legal technology. Are you going up against all of them? Or what how do how do think about the existing market of legal tech?

这个领域长期存在大型并购机构和老牌企业，但终端用户并不喜欢它们。虽然它们根基深厚...

Right. So there's been a lot of sort of large m and a machines and incumbents in this space for a long time. They're not very popular with the end users. Mhmm. I think they have very kind of far far reaching roots.

确实存在数据护城河等优势。但AI彻底改变了游戏规则——开发速度被重塑，催生了新品类。许多原本属于并购套件的单点解决方案正迅速过时。

There's some advantages and, you know, data moats and so on that that come into play. But effectively, what AI has done is really changed the game in terms of how quickly you can ship something, and it's created a new category. So a lot of, again, this existing point solutions were in maybe suites of of these M and A machines. And now a lot of it is becoming irrelevant very quickly. Mhmm.

计费软件成本也在急剧下降。我们30人团队（录制时约100人）的交付能力能碾压数千人的工程团队，发展速度远超规模百倍于我们的公司。

And the cost of billing software is also going down very, very rapidly. So our ability to out ship or, you know, out deliver these teams of thousands of engineers Yeah. With just 30 Yeah. Is insane. And so we have instead managed to build a company with I I think at the at the time of recording, it's about a 100, where, like, our velocity is way higher than companies, you know, 100 times our size.

我认为就我们过去一年如何构建公司而言，这本身就很有趣。因为我们从YC毕业时大约有10人，现在发展到100人。这意味着我们平均每周要入职大约两个人。而正确招聘真的很难，这是一项需要学习的技能。

I think that's interesting in and of itself in terms of how we built the company over the last year. Because when we came out of YC, we were roughly 10 people, and now we're a 100. And that means we've onboarded on average, like, two people a week. And hiring correctly, it's really hard. Like, it's a skill you need to learn.

既要为发展速度招聘，为创业精神和不同产品的所有权招聘，同时也要考虑规模，因为公司正在呈指数级增长。是的，所以你需要你的团队成员也能指数级成长。如果人员只是线性增长，到某个时点就会出现巨大差距。

And hiring for velocity, hiring for, you know, entrepreneurship and ownership of different products and things, but also scale because the company is growing exponentially. Yeah. So you need your your teammates to scale exponentially as well. If people scale linearly, at some point, it's a really large delta. Yeah.

然后，事情就开始行不通了。

And then, you know, things aren't working out anymore.

这些大型企业是否存在锁定效应？比如大型法律科技公司？

Do these big companies have lock in, like the big legal tech companies?

这些大公司确实有几个优势，但我认为劣势几乎是优势的十倍。它们有巨大的数据优势，作为现有供应商可以锁定大额合同。

So these big companies have a couple of advantages, but I think the the disadvantages outweigh the advantages almost 10 to one. There were very large data advantages and, you know, being like an incumbent where you lock in a large contract.

确实。

Yeah.

但我觉得买方态度也发生了变化。现在没人愿意签订五年长约，毕竟世界变化太快。所以我们看到他们更倾向签订一年期合同。

But I think the buyers have also changed attitude here. So we're not seeing anybody want to lock in a five year contract. Right. Because the world is moving so Of course. So we instead see them, you know, doing one year contracts.

听起来是促使公司加速发展的好动力。

Sounds like a good motivation for companies moving faster.

确实如此。就连律所也一样，他们也不希望被供应商锁定。

It is. Yes. And and but even law firms. Right? I mean, they don't want to be locked in with a vendor.

所以他们签订一两年期的合同。现在看到很多机构都在重新评估，可能在2023或2024年实验阶段做过选择，但现在要考虑长期部署什么解决方案了。

So they're doing one or two year contracts. And, you know, as we see them now coming up in a lot of places, they're also looking outside of their existing alternatives. So you might have made a bet back in 2023 or 2024 when it was experimentation days. But now you're looking at what are we gonna deploy, you know, more long term. Yeah.

在那里，我看到的是，人们确实关注技术，但更关键的是他们在放大视角，审视你的变革速度。他们希望与能带领他们从起点A到达终点B的合作伙伴共事。目标可以各不相同——可能是希望优先发展AI以推动营收增长，也可能是追求提升盈利能力和精简运营。动机可以截然不同。

And there, what I'm seeing is, yes, people look at the technology, but even more so they're they're zooming out and they're looking at your rate of change. They want to work with the partner that's going to get them from point a to point b. And they can be different things. It might be we wanna be AI first and drive our top line, or we wanna drive profitability and, you know, streamline our operations. It can be, you know, very different motivations.

没错。

Right.

你们的技术栈是怎样的？底层架构是

How does your tech stack look like? What's under

内部架构吗？对。是这样，在构建基础设施之初，我们就明确要基于Azure平台，因为我们的客户都在这个生态里。早期阶段，OpenAI的GPT模型基本是唯一能通过Azure部署的模型。

the Internally? Yeah. Yeah. So building our infrastructure, I think from the beginning, it was pretty clear that we wanted to be on Azure just because it was the same that our customers were on. And in the beginning, I think OpenAI and GPT was really the only model that you could serve via Azure.

现在我们的选择丰富多了——可以灵活切换使用AWS、Claude、Gemini、GPT和Mistral等平台。核心在于构建能随时热切换模型的系统架构，同时确保模型能持续优化。我们还开发了分类模型机制：简单查询匹配轻量级模型，

Right. Now we have much more, you know, options available to us. So we use AWS and Claude and Gemini and GPT and Mistral kind of interchangeably. The biggest thing there has been how do we build everything in such a way where we can hot swap the models whenever we want, and also build it in such a way that the models become better, everything improves. And now we've also looked into classification models where, you know, if you do a simple query, we'll serve you a simple model.

复杂查询则启用高级模型。这样既能控制成本，也符合实际需求——毕竟不是所有场景都需要

If you do a complex query, we'll serve you a complex model. And that's just because that's just, you know, to keep the margins down, but also, you know, sometimes you don't need a bazooka when you just need a, you know, water gun.

那么买家是谁？据我了解律所可能...能否具体说明下？比如合伙人们和其他相关人员...

So who is the buyer? My understanding is that law firms have maybe you could explain to me. Like, I Yeah. There's a bunch of partners, there's other people there too.

对。

Right.

律所或企业法务团队通常如何架构？有哪些角色？采购决策者和实际使用者分别是？

How is the law firm or a legal team at a company Yeah. Generally constructed, and who are there, and who buys it, and who uses offer?

这取决于规模。大型律所主要由合伙人管理层主导，但通常设有创新部门——其影响力因所而异。强势的创新部门拥有自主决策权，负责软件采购和整体创新规划。说实话，我最享受与这些创新专家合作，他们真正懂行。有些人只是应付式地勾选AI选项，而他们才是真正推动变革的人。

It changes a bit depending on size. So if you start with the with the biggest firms, of course, you have the partner group that kind of runs things, but you very often have an innovation department, which sometimes have more or less influence. If it's a very strong innovation department, they make their own choices. They procure software and they're responsible for the entire innovation agenda. I've frankly got the most energy out of working with the innovation folks who are really smart about these things because there's a lot of people that just want to kind of check the AI box and then others who really want to push things forward.

这里有个有趣的困境，他们本质上是在推动整个公司层面的效率提升。但他们自己并非最终用户。不过，你可能经常会看到并购部门、争议解决部门或仲裁部门的创新实践者。

And the interesting dilemma there is, you know, they're basically driving efficiency across the across the stack or across the firm. Yeah. But they're not the users themselves. Right. However, you might often have innovation practitioners that work in the m and a group or the disputes group or arbitration.

然后他们会与这些团队合作推动升级。他们的工作方式会非常流程化。接着可能会用Legora为终端用户构建用例。因为在大型律所工作时，你必须完成计费指标。

And then they will work with those teams to drive an upscale. So they will have a very, like, process minded way of working. And then they might use Legora to build use cases for the end users. Because when you work in a big law firm, you need to hit your billing targets.

是啊。

Yeah.

他们工作强度很大。虽然我们创业者也拼命，但——

They work a lot. Like, we grind as startup folks, but

律师们。

Lawyers.

律师同样拼命。如果你知道有件事需要六小时完成，嗯。而且你清楚具体操作方式，你可能就不会冒险尝试寻找更高效或更优质的解决方案。你只会遵循惯常的工作方式。确实。

Lawyers grind as well. And if you know that there's a way to solve something, and it's gonna take six hours for you to do that Mhmm. And you know a way how to do it in six hours, you might not take the chance of exploring a way how you could potentially solve it quicker or, you know, with a higher quality. You'll just conform to the way you're used to working. Yeah.

所以创新团队其实肩负着重大机遇，坦白说也是推动律所变革的使命。再看中型律所，通常没有专门的创新部门。这时就是合伙人们在推动或决策。但要让全体合伙人都认同很难。

So innovation teams have a huge opportunity and frankly, you know, mission to drive that across the firm. And if you go down a bit, so you have sort of mid sized firms, more often than not, you might not have an innovation department. And so it's the partners who are making the move or the decision. And what I found is it's hard to get the entire partnership to buy in.

这点请详细说说。因为很多创始人问我，如何向金融公司或律所这类机构销售？似乎最难的就是必须说服所有人。

Go deeper on this point. Yeah. Because I know a lot of founders is asking me, how do I sell to, like, a financial firm or a law firm or something like that? And it seems like this is the the tricky part. It's like you have to convince everybody.

要么说服所有人，要么从小范围着手。比如先和某个合伙人及其团队合作，把他们打造成标杆。嗯。然后其他人就会好奇：那家伙在搞什么？

You have to convince everybody or you start smaller. K. You say, let's work with this partner and their team and make them rock stars. Mhmm. And then everybody else looks at them saying, what's that guy doing?

没错。看起来太棒了。我们也要加入。这样就能逐步扩展。

Right. That looks awesome. We also want in. And then you expand.

但关键在于要采取一种销售策略，

But the the key here is to sell sell sort of,

不是自上而下，而是先向高层人员推销。没错。在我们这个行业，自下而上的推广是不可能的，因为软件不是个人采购的。你需要通过采购部门和IT部门来获取。是的。

not not top down, but sell to the senior people first. Right. So there's it's impossible to do a bottom up motion in our industry because you don't procure software individually. You take it through procurement, and you take it through IT. Yeah.

而且还需要通过大量的安全检查。为了在你的系统中处理客户数据，你需要通过许多数据隐私审查。

And there's a lot of security checks. There's a lot of data privacy checks that you need to go through in order to actually, you know, serve client data in your systems.

你23岁时联合创立了Legora。是的。那时你已经做了很多工作。你还在其他一些知名公司有过多次不同的工作经历。是的。

You were 23 when you cofounded Legora. Yeah. By then, you've already worked done a lot. You had some stints at other wise companies, like multiple different ones. Yeah.

在创办这家公司之前，你的背景是什么？

What was your background before you started this company?

当我18岁申请大学时，实际上有两个选择。我要么选择成为职业DOTA、DOTA2选手的道路，要么去上大学。

When I was 18 and it was time to apply to college, I actually had two options. I was either gonna go down the route of becoming a professional DOTA, DOTA two player, or go to college.

我知道这个。

I knew this.

我当时想的是，好吧。每种结果的最佳情况是什么？Dota的最佳情况是赢得国际邀请赛，这是世界上最大的比赛。你可以赚1000万美元。那会很棒。

And my thinking at the time was, okay. What's the best case scenario in each of the outcomes? So best case scenario in Dota would be to win the international, the biggest tournament in the world. You make $10,000,000. That would be amazing.

但后来我想，那之后呢？你知道，感觉就像生活就此停滞了。是的。而上大学的最佳情况基本上就是我现在在做的事情。所以我决定去上大学。

But then I was thinking, what happens then? You know, it kind of feels like then life stops. Yeah. And the best case scenario with going to college was basically this, what I'm doing now. So I decided to go to college.

在瑞典申请大学时，你去一所学校读一个专业。所以工程学院和商学院是完全分开的，我觉得这很奇怪。我们完全不混在一起，这很不好。但有一个小技巧，你可以先申请一所学校，然后撤回申请去申请另一所，嗯。或者撤回申请去申请另一所，然后打电话告诉他们你搞错了，想重新申请。

And when you apply to college in Sweden, you go to one school to do one program. So the engineering university is completely separate from the business university, which I think is really weird. Like we don't mix at all, which is bad. But there was a hack so that you could, you can make an admission to one of the schools and then kind of pull the admission to make another one Mhmm. Or pull your application to make another one, and then call them and say that you messed it up and you wanted to get it, you know, reapplied.

所以我最终安排得可以同时上两所大学。疫情期间这样做时机非常好，因为当两堂课时间冲突时，你只需要摆两台笔记本电脑就行。

So I I ended up making it so that I could go to both universities in parallel. It was a really good timing during COVID to do that because that means when you have two lectures at the same time, you can just have two laptops at

一台一台。

one one.

对，对。有好几次我遇到两所大学考试时间撞车，你就得这边架一台摄像头，那边架一台，假装只参加其中一场考试。这样坚持了一两年后，我开始做程序员工作。

Yeah. Yeah. And there were multiple times where I had, like, exams at the same time with both universities. And you would you kinda sit with one camera over here and one camera over here pretending that you were just doing one of the exams. And so like one or two years into it, I was working as a programmer.

当时我在为电竞博彩建立统计模型，这很有趣，但我也想了解商业方面的情况。后来有幸加入一家叫Norgen的公司，类似YZ但专注社会影响力。

I was building statistical models for esports betting. And that was really fun, but I think I I also wanted to kinda see what the business side looked like. So I had a the the privilege of working at a company called Norgen. It's like YZ, but for impact. Yeah.

公司总部在斯德哥尔摩。在那里我接触到很多创业者，当时让我震惊的是，其中一些人并不满足于现状，而是制定了五年计划要征服北欧市场。这让我立刻有了不同视角。

And it's based in Stockholm. And I think I got a lot of exposure to other entrepreneurs. And what struck me then was, one, a few of them were not super ambitious to build companies that we're doing now, but they sort of had this, like, five year plan to conquer Nordics. Yeah. So I think immediately, like, I had a different take on it.

后来在麦肯锡短暂工作过，也去过Bamlow，还在Depict待了一周。

And then they just short stint at McKinsey and worked at Bamlow and just one week at Depict.

Depict是那种极具人才吸引力的公司，真的。从那里走出了很多厉害人物，比如Lovable的Anto就是创始人之一。

Depict was one of those companies is is one of those companies that was an incredible talent magnet. Yeah. Like, some incredible people have come out of Depict. Like Yeah. Anto from Lovable was one of the founders.

是啊，还有不少其他人。虽然你只待了一周，但已经开始接触到了。

Yeah. But there's a bunch of others. You're starting to go out even though you spent a week there.

确实。

But Yeah.

这种人才磁场催生出这么多酷公司的现象，还挺有意思的。

Is it it's, like, kinda cool how you have these magnets that that that spun off too much about the cool companies. Companies.

不，他们非常出色。而且你知道，我们在斯德哥尔摩都是好朋友。这是个很小的生态系统。

No. They're amazing. And, you know, we're all good friends in Stockholm. It's a small ecosystem.

是啊。

Yeah.

而且互相加油打气也真的很有趣。

And it's really fun to kind of cheer on each other as well.

YC在四月份结束了。你能带我们回顾一下这段时间公司的成长和你个人的发展吗？比如，你们从10人发展到100人，发生了什么？

And YC ended in April. Can you walk us through sort of like the company growth and your personal development in this time? Like, you were 10, now you're 100. What happened?

我们增长得非常快，同时也感受到了阻力。是的，你知道吗？我们把产品推向市场后，通过一次演示就能卖出去。当律所开始在一次演示后就购买产品时...

We grew really fast, and we were also feeling the the drag. Yeah. You know? Like, we we took the product to market, and, you know, we would sell it in a demo. And when law firms start to buy things after one demo

是啊。

Yeah.

说明你做对了什么。所以我们的逻辑是，应该多做这类事情，而且希望同时在全球范围推广。这个领域很明显，法律与大语言模型...

You're doing something right. And so the rationale was like, we shouldn't be doing more of this, and we wanna do it everywhere all at once. And this is also a space where it's kind of obvious that legal and LLM

没错。

Yes.

是绝配。是的。当时行业里还有很多其他公司。我喜欢说，那时候有太多法律AI助手了，现在感觉很多都逐渐消失，只剩下几个赢家脱颖而出。基于这个逻辑，我们也希望引入美国资本，因为当时机成熟时，我们想从斯德哥尔摩转移到美国。

Is a good fit. Yes. And so there were a lot of other companies in the industry. I like to say, like, there there were so many legal AI assistants, and now it just feels like many of them have kind of fallen and they're emerging a couple of winners. With that rationale, we wanted also to get American capital on the in the company because we wanted to be able to make the move from Stockholm to The US when the time was right.

融资后的第一次董事会上，我记得当我说接下来四五个月都不准备销售时，某些董事会成员脸上的表情。原因是每当我们有机会接入一个客户时，都需要大量工作。要让他们充分理解在这个平台上能实现什么，需要付出很多努力。而且，法律专业人士首次登录的体验只有一次机会——如果搞砸了，他们就不会再回来了。

After we raised the money during our first board meeting, we sat down and I remembered the look on on some of our board members' faces when I basically said, we're not gonna sell for the next four to five months. And the reason for that was when we got the chance to onboard a client, it took a lot of work. Took a lot of work to get them to a level of understanding of what they could accomplish in the platform. And also, the first experience of a of a legal professional logging in is the one chance you have. If you mess that up, they're not coming back.

我们曾遇到过几次大规模人员入职的情况，期间确实犯过一些错误。我们不想重蹈覆辙，所以全力提升了系统的可靠性和扩展性，最终实现每天能稳定接纳一千名律师注册。达到这个目标后，我们才真正放开手脚扩张，那也是我们开始大规模招聘的时候。

And we had a couple of of of situations where we'd onboard a lot of people, and we had done, you know, some misses. And we didn't want to ruin that. So we worked really hard on reliability, scalability, got the system to a place where we could comfortably onboard a thousand lawyers a day. And once we had that, we kinda let it rip. And that's also when we really started to hire.

十月份时团队大约25人，仅仅六个月后就发展到100人。当时我们决定要横扫欧洲所有市场，并开始向美国进军。由于我们是瑞典的小型初创公司，初期在美国的洽谈颇费周折——我本人就曾多次往返纽约进行商务洽谈。

So we're maybe 25 in the October. And just six months later, we're now a 100. So what we did was we said, okay, we're now gonna scale across every market in Europe, and we're gonna start scaling towards The US. And our initial conversations in The US, some time because we were a small Swedish startup. So I I made, you know, multiple trips back and forth to New York.

嗯。

Yeah.

现在我们已在纽约、伦敦、斯德哥尔摩设立枢纽中心，同时在西班牙、法国和德国也部署了本地团队。可以说我们是全面出击，抱着'万事俱备只争朝夕'的劲头在推进。

And now we open up hubs both in New York, London, Stockholm, and also people locally in Spain, France, and Germany. So we've really gone at it and just said, hey, we wanna do everything everywhere all at once, and let's do it now.

对你个人和Sigir来说，这段经历带给你们什么收获？

And for you personally, and Sigir, like, what was your experience in that?

我认为最大的成长在于完成了从执行者到管理者的转变。

I think the biggest takeaway in learning is going from being an IC into delegating.

确实。

Yeah.

这个转变的关键在于——你虽然擅长某项工作，但个人能力终有上限。所以必须教会别人，还要招募在各个领域都比你更优秀的人才。我们早期就聘用过另一位YC创始人，最后招了Jake？

And that move you know, you know how to do something, but you that's not gonna scale. So you need to teach somebody else to do it. And you need to hire people who are way better than you on a lot of different topics. So one of the early sort of hires that we made were actually another YC founder. We've ended Jake?

对。实际上我们团队扩张时吸收了很多创业者。这不仅是因为我们需要这些技能，更是因为我们的组织架构——本质上我们是在公司内部并行运作着多个创业项目。

Up Yeah. And we actually, we've scaled the team with a lot of entrepreneurs. And that's not only, like, the skills we're looking for, but it's also, like, the way that we built the company because we're effectively running multiple companies within the company.

这就像许多顶尖公司都在遵循的隐秘法则：首批招聘对象最好是连续创业者。这个建议其实来自早期的Paul Graham——有人担心创业失败会影响职场竞争力，但在初创生态里，这种经历反而会让你更抢手。确实如此。

It's sort of like a secret playbook that a lot of ways to companies, some of the best ones are all following is that the first people you wanna hire are all former founders. Yeah. And it's kind of actually an advice that I got from Paul Graham back in the days is that sometimes you think when a founder that I worked in this company for three years, didn't go well, Am I less attractive in the job market? Like, if you're here or if you're in a start up center, you're actually more attractive in the job market. Yeah.

因为人们实际上想和你这样的人共事。

Because people actually wanna work with people like you.

没错。我们也想雇佣他们。所以是的，这一切都很棒。而且，还有机构以及解决问题的态度，这些正是你所寻找的。

Yeah. And we want to hire them. So Yeah. It's it's been amazing. And, also, the agency and the the attitude to problem solving, that's kinda what you're looking for.

是的。有时候你需要为了规模而招聘。对吧？比如，现在我们有一个庞大的销售团队，你需要一个经历过从1000万到5亿规模的人，因为这就是我们正在经历的旅程。

Yeah. And then sometimes you need to hire for scale. Right? Like, now we have a significant sales team, and you need somebody who's seen the 100 you know, 10,000,000 to 500,000,000 because that's the journey that we're on.

我在Airbnb学到的经验，我相信也适用于你们，最初的文化就是你雇佣的人。当然。是的。

And my learning from Airbnb, which probably I'm sure applies to you, the culture in the beginning is the people that you hire. Of course. Yeah.

当我们现在扩展中心时，我们总是从斯德哥尔摩派一个人和他们一起。这是斯德哥尔摩办公室最优秀的人，然后去旅行并建立新的中心。

And when we've now scaled the hubs, we always send a person from Stockholm with them. It's the best people from the Stockholm office that then travels and setups the new hubs.

你看起来像是那种体现了这些特质的人。你可以直接做事。那么你能告诉我这在你公司是如何体现的吗？

You seem like the kind of person who embodied the attributes. You can just do things. So can you tell me how that is reflected in your company?

你可以直接做事。当我们开始建立这家公司时，我们对法律一无所知。对吧。我想这在我们的第一次面试中表现得相当明显。然后我们，你知道，从那时起做出了正确的行动，到第二次面试时我们证明了我们可以做到。

You can just do things. And when we started building this company, we didn't know anything about law. Right. I think that was pretty apparent in our first interview. And we, you know, made the right moves from them from them to the second one where we showed that we could do it.

你们申请了两个不同的批次。

You applied for two different batches.

是的。第一次进行得不太顺利。关于这个特质，这也是我在别人身上寻找的。在我做的很多面试中，我经常会问一个问题，你知道，你在公司角色之外还做了什么？

Yeah. The first one didn't go as well. And so about this attribute, it's something I look for in others as well. During a lot of the interviews I do, I ask I often ask the question, you know, what have you done outside of your role for the company?

是的。

Yeah.

在这里，我寻求的是创造力、发现并解决问题的能力。是的。还要承担起比手头工作更多的责任。对吧？

And here, I'm looking for creativity, ability to spot problems and solve them. Yeah. And to take responsibility for more things than just the stuff that you're doing. Right. Right?

关于创业和构建未来，坦白说我们需要重新构想许多正在做的事情。我们不想要那些被上司指令束缚的人。我们的组织结构非常扁平化，比如营销团队，我们需要通才利用AI完成过去十倍的工作量。过去需要30人的营销团队，现在可能只需要5人。

And I think in terms of starting companies and, you know, building the future, because frankly, we need to reimagine a lot of like the stuff that we're doing. We don't want people who are bogged down by your boss telling you to do something. Right? We have a very sort of flat organization where, let's say, our marketing team, we want generalists who are using AI to do 10x more work than they could have done in the past. And where you might have needed a 30 person marketing team, you now need five.

而这5个人必须全能，愿意主动突破界限。在这个工具能带来巨大杠杆效应的时代，这种特质愈发重要。完全同意。

And you want those five people then to be complete, you know, yes sayers and to go out, you know, above and beyond. And that that characteristic, I think, is increasingly important as well in an in an age where if you're really ambitious, you can get a lot of leverage out of tools. Absolutely.

那么展望未来五到十年，律师的日常工作会变成什么样？

So if we fast forward, like, five or ten years, how does the day to day job of a lawyer look like?

这很有趣。我们经常思考这个问题。我认为未来更多是进入一个审核工作的空间，而非亲自执行工作。嗯。

That's interesting. We think about that a lot. Right? I I'm kind of viewing it as you're more and more entering a workspace of reviewing work than actually doing it. Mhmm.

你需要管理客户的期望，以及AI代理的工作产出。嗯。实际上你是在指导它们，监督它们执行任务，确保不仅结果正确且符合标准，还要管理成果交付给客户的方式。

And you are managing the expectations from your clients and the expectations and the work from your AI agents. Mhmm. Right? You're you're effectively instructing them. You're watching them go out and do work, and you're making sure that everything they're doing is not only correct and sort of at your standard, but you're also managing how that work gets delivered to the client.

嗯。

Mhmm.

因为最终总需要懂行的专业人士把关。这正是我们与律师而非法律服务使用者合作的重要原因——律师对交付最终产品不可或缺。但预测五到十年后的发展确实很有挑战性。

Because I think, you know, you you will always want somebody who knows their stuff Yes. On this. And there's a big reason for why we're working with lawyers and not with the people who might, you know, use the legal services. Because the lawyers needed and and necessary to deliver the end product. But looking five, ten years ahead in in these days is also It's hard.

确实很难。如果我能预知AMOLs十年后的样子...是啊。

It's hard. Right? If I knew where the AMOLs would be ten years from now Yeah.

我们现在连几周后的情况都难以预料。确实。

We're looking weeks ahead now. Yeah.

这很有趣，尤其是在我们的产品路线图方面。嗯。我试图提前多个季度规划它们。

It's and and that's funny just with our product road map. Mhmm. I tried to do them kinda like many quarters ahead.

是啊。

Yeah.

是的，这很难。真的很难。

Yeah. I It's hard. It's really hard.

没错。你认为大型AI实验室会尝试涉足法律领域吗？

Yeah. Do you think that the large AI labs are going to try to attempt at doing law?

可能不会专门针对法律，但我确实感觉它们越来越像平台公司而非模型供应商。比如谷歌正用Gemini构建Google Workspace，Anthropic则在大力推进MCP理念，打造一个通往众多应用的通用入口。嗯。对我们这类公司的期望其实相当明确。

Maybe not law specifically, but I I do feel like they're more and more becoming platform companies rather than model providers. I mean, Google is building Google Workspace with Gemini. Anthropic is running very hard on the MCP idea of building kind of a universal entry point into a lot of applications. Mhmm. I think the the expectations on companies like us are pretty clear.

你知道，模型实验室产出的东西某种程度上是意料之中的。嗯。而我们在此基础上添加的一切更像是锦上添花。

You know, whatever comes out of a model lab is kind of expected. Mhmm. And then everything else we're adding on top is kind of like icing on the cake.

这种产品市场契合感如何？

How does it feel to that product market fit?

我觉得这种感觉最好形容为一种拖拽感，或者说无限——你正被

I think the feeling is best summarized by almost like this drag feeling or kind of infinite You're being

拉入市场。

pulled into the market.

对。就像字面意思，感觉我们有无限的需求。我认为这是因为产品确实有效，它已经从实验性AI的范畴中脱颖而出。

Right. It like, it literally feels like we have infinite demand. And I think that's it's coming from a point of the product is working, and it's moved from being in this experimental AI bucket

是的。

Yeah.

我们核心工作的交付依赖于这个系统。现在如果出了问题，你知道，马上就会接到电话说，嘿，我们没法继续了。怎么回事？对吧？

Into we are reliant on this for core work that we are delivering. Right now, if something breaks, you know, immediately, we get a phone call saying, hey. We can't do this. Like, what's going on? Right?

然后我们修复它。基本上就是这样，你开始时希望自己在做正确的事，试图让早期合作伙伴对你的工作感到兴奋。坦白说，一开始很多人加入我们是因为他们想参与这段旅程，他们下了赌注。是的。我非常感激并高兴他们这么做了，因为现在我们已经把他们从A点带到了B点，并且正在持续扩展。

And we fix it. It's basically been this point of you start out, you hope that what you're doing is the right thing, and you try to get early partners excited about what you're doing. And in the beginning, to be, you know, really frank, a lot of people got on with us because they wanted to be on the journey, and they took a bet. Yeah. And I am so thankful and happy that they did that because now we've taken them from from point a to point b, and we're continuously scaling from here.

所以我们告诉Weiss公司搬到旧金山

So we tell Weiss companies to move to San Francisco

是的。

Yeah.

一般来说。你决定不采纳这个建议。能跟我们说说这背后的考虑吗？也许还有一些关于不在这里的利弊。

Generally. You decide to not take that advice. Can you just, like, tell us a bit the thinking here? And and maybe, like, if you have some pros and cons Yeah. About not being here.

是的。我们留在斯德哥尔摩的原因是我们需要一个市场来发展。如果你去美国，不仅竞争更激烈，而且我认为这会迫使你成为一个更狭窄的公司。你开始横向发展，然后意识到，等一下，我们在这方面真的很擅长。

Yeah. The reason why we stayed in Stockholm was we needed a market to grow in. And if you go to The US, it's not only more competitive, but I think it kinda pushes you into becoming a more narrow company. You start building really horizontal, and then you realize, wait a minute. We're really good at this.

于是你开始在其他市场扩展，很快发现，啊，我们在芬兰也是最好的。嗯。我们在丹麦最好，在挪威也是最好的。然后扩展到西班牙、法国、德国、伦敦，再到美国。是的。

So you start to scale it in other markets, and you quickly notice, ah, we're the best in Finland too. Mhmm. We're the best in Denmark, and we're the best in Norway. And then you scale to Spain, France, and Germany, London, and then The States. Yeah.

那时，我们已经完成了15个新市场的进入。算法或方法已经基本确立。当然，美国是一个更大的挑战，但我们也已经从一个小池塘里的小鱼成长为现在大池塘里的鳄鱼或鲨鱼。

And at that point, we had always you know, we had already done 15 new market entries. The algorithm or the the method was already kind of established. Of course, The US is a bigger undertaking, but we had also then grown from this small fish in a small pond to crocodile or a shark in in the bigger pond now.

所以你在五月中旬筹集了8000万美元。是的。你在纽约开了办公室。你和美国这里最著名的律师事务所之一一起推出了产品。是的。

So you you've raised $80,000,000, like, in mid May. Yeah. You opened an office in New York. You launched with one of the most famous law firms here in The US. Yeah.

看来你正试图将自己定位为全球AI法律领域的领导者。

It seems like you're trying to position yourself as the category leader of AI law in the world.

没错，百分之百是这样。我认为在很多方面，我们已经做到了。对我来说，这更多是关于雄心和下一步的问题。很容易说，嘿。

Yeah. I mean, a 100%. And I think in many aspects, we're already there. It's for me more of a question around ambition and what's next. It's very easy to say, hey.

我们看到这个问题，让我们去解决它。然后你就满足了。

We see this problem. Let's go solve it. And then you get satisfied.

嗯。

Mhmm.

但我觉得每次我们解决一个问题，就会有新的问题出现。随着我们在整个法律软件栈中越钻越深，我们也发现软件与服务之间的界限正在模糊。AI正在以极快的速度持续发展，这意味着我们也需要这样做。因此在我看来，这个领域的领导者不仅构建软件，他们还作为这些大型律所的战略合作伙伴，帮助他们在这一转型中获胜，因为这是一次非常巨大的转型。

But it feels to me like every time we solve a problem, a new one emerges. Right. And we're finding that as we go deeper and deeper and deeper in the entire legal software stack, we're also seeing that the line between software and service is blurring. AI is continuously developing super, super quickly, and that means we need to do the same. And so in my mind, the category leader in the space does not only build software, they serve as the strategic partner to these large firms and they make them win in this transition because it's a very large transition.

这也是为什么我们在保持文化、紧迫感和速度的同时，尽可能快地扩大了人员规模。

And that's also why we've basically scaled how to head count as as quickly as we've could whilst maintaining kind of culture, urgency, and velocity.

我遇到的很多创始人都在问我如何打造一家垂直AI公司的问题。这似乎是现在人们正在创建的公司类型。对于刚刚起步的创始人，你有什么通用建议吗？

So a lot of founders that I I meet are asking me questions about how you build a vertical AI company. That seems like the kind of companies people are building now. Do have any general advice you wanna give to those founders who are just starting

第一个显而易见的建议是不要被供应商锁定，也不要与AI实验室竞争。嗯。AI实验室在推进，对吧？像Perplexity这样的公司也是。

The first kind of obvious tip is don't get locked in with a provider and don't compete with the AI labs. Mhmm. The AI labs ship. Right? And so does companies like Perplexity and and others.

所以我认为你需要非常清楚和诚实地认识到你在哪里增加价值，在哪里建立长期的护城河。这是我们在Laguerre思考很多的问题。比如，我们如何构建像船一样的东西，这样当潮水上涨时，一切都会变得更好？是的。如果你刚刚起步，你必须意识到你没有能力在任何一个方面超越那些公司。

And so I think you wanna be really clear and honest to yourself where you're adding value and where you're adding long term moat. And this is something that we've thought a lot about at Laguerre. Like, how do we build things as boats so that when the tide rises, just everything gets better? Yep. If you're just starting out, you gotta realize that you do not have the capacity to outperform any of those companies.

你必须找到一个狭窄的、你知道模型无法触及的领域来做这件事？还是

You gotta have to find a narrow a narrow category to do it where you know the models won't get to? Or

要么就是找到一种非常有创意地利用模型的方法，我是说，以一种别人没做过的方式。比如AI记录，这是个好例子。典型的AI记录很难做。嗯。

Either that or finding out a way to leverage the models very creatively, I mean, in in a way that others haven't done it, I think take AI scribing. Yeah. It's a good one. Like, typical AI scribing is hard to do. Mhmm.

你需要嵌入大量定制提示和正确方法，让它使用正确的医学术语，这和法律很相似。就像你需要用律师起草条款的方式来撰写条款。

And you need to embed a lot of, like, custom prompts and ways to get it right so that it uses the right medical language, is very similar to law. Like, you needed to write clauses in a way that a lawyer would write a clause.

对。

Right.

而不仅仅是模型输出的最可能的答案。

Not just what the model spits out as the most probable answer.

如果我在看这个视频时想申请Legora的工作，嗯。告诉我应该对申请流程或工作环境有什么期待。

If I'm watching this video and I'm like, I'm thinking about applying for a job at Legora Yeah. Tell me about what I should expect either from the application process or from working there.

我们看重的是雄心壮志和愿意说‘我们面临这个大问题，这座大山，我们怎么攀登？’的态度。我们也非常直白地告诉候选人这不是朝九晚五的工作，我们不是传统的瑞典工作环境。我们有好的方面。

The things that we look for are ambition and the willingness to say, we got this huge problem. There's this huge mountain. How do we climb it? And we're also very upfront with candidates that this is not a nine to five, and we're not the traditional Swedish working environment. We have the good stuff.

我们有咖啡时间。但我们还有很多

We have the fika. But we we have a lot

更多

more

渴望，坦白说还有更高的期望。我们不仅对自己有这样的要求，对彼此也是，因为我们想作为个人成长，作为企业家、公司和领导者成长。说到申请流程，我们最主要的是做很多案例分析。

hunger and, you know, frankly, a lot higher expectations. And we we want that not only for ourselves, but for each other because we want to grow as people. And we want to grow as entrepreneurs and as a company, as leaders. And I think they're just looking at like our our application process. The biggest thing we do is a lot of cases.

对吧？如果你想加入我们的市场团队，你需要来向我们推销我们的产品。是的，而且你要做一个非常出色的推销。如果是工程团队，我们基本上会要求你搭建一个Legora的概念验证。

Right? If you want to come in and work in our go to market team, you need to come and pitch us our product. Yeah. And you need to do a really strong pitch. And, you know, if you take the engineering team, we basically ask you to build a POC of Legora.

而且，你知道，我们希望你们能使用AI生成的代码工作，但也需要你们能够解释它。是的，对吧？还要设计可扩展的系统。我认为斯德哥尔摩是个小型生态系统。

And, you know, we want you to work with AI generated code, but we also want you to be able to explain it. Yeah. Right? And to design systems that scale. And I think Stockholm is a small ecosystem.

因此，在这里很容易做背景调查，看出谁真正优秀，谁曾在某家公司并助其取得成功。你知道，不仅仅是——

And so it's also quite easy to make references and see who's actually good and who's who who who's been in a company and made them a success. You know, not only

当时确实如此。恰到好处

Was there Exactly. At the right

另一个重要部分是我们在全欧洲招聘。所以我们有员工从马德里、阿姆斯特丹、德国、巴黎搬到斯德哥尔摩。我们倾向于不在天气恶劣的十一月安排入职，但我觉得我们已开始与许多其他公司共同打造这个AI中心，不仅充满乐趣，还能孕育出优秀的公司。

And another really big piece is we're hiring all over Europe. So we've had people move from Madrid, from Amsterdam, from Germany, from Paris, all the way to Stockholm. We tend to not onboard them in November when it gets nasty, but I feel like we've started to build this sort of AI hub together with many other companies that is not only, like, super fun, but also, you know, great companies come out of it.

非常感谢你回到IC。

Thank you so much coming back to IC.

谢谢，古斯塔夫。

Thanks, Gustav.