I remember doing this interview in Swedish. There's a saying like blood and smoke and you like you taste the like you taste the blood because you you worked so hard. Yeah. She publishes the article in English at Legora. We wake up with a metallic taste of blood in our mouths and people in the company go, "Holy [__] is Max a vampire or does he just floss badly?" Like what's going on?
It's not like a culture that I think would quite work in San Francisco. Like I don't know if that's something that you can do.
Well, when we open our San Francisco office, they're going to taste the blood. And yeah, they're going to taste the blood.
Well, this is going to be a cool new format. I'm here with my new partner, Chathan and Max. And Max, you're the founder CEO of Legora, which is an amazing legal tech company that Chathan sits on the board of. And so, I just feel really lucky to be doing this with both of you. So, you guys, thank you for thank you for making this happen.
Thank you so much, Jack. It's great to be here.
Okay, so I want to start with the topic of competition. And Jason, when you invested in the company, there were already competitors out there. This was I think it was only two. It's crazy because Legora is like 2 years ago.
It's a big company already.
Yeah. Almost 400 people, you know, but this this the seed was 2 years ago. And at the time of the seed, it was an early market, but there were competitors out there. And so I actually want to start with you, Jason, like what was in your head at the moment you invested? Were you thinking about the landscape around like were you just Max is so special that I don't care? Like what was going through your head when you did that?
The first meeting that we had was with Max was with me, Peter and Max actually in the other room. And interestingly, I had invested in two other legal software companies pre-AI.
And so there was a shape of the legal market that I intuitively understood because I participated in the market. And so I sort of understood the different kinds of lawyers. Who buys software? Do in-house lawyers buy software? Do law firms buy software? How they sort of there was an intuitive understanding that I had and there's sort of like two things that happen when you've like sold into an industry before either you end up hating it or you have some strong bias against it. So there was always this idea that there's opportunity for AI in the legal market. And you know there was a player in the market that had already raised at a billion dollar valuation. And when Max came in to chat with me and Peter, the thing that immediately jumped out was the clarity of thought that Max had on why the general foundation models had a lot of room to grow in intelligence and how that was going to be a huge boon for the legal profession over the next couple years. And so he had this like very strong viewpoint that there was something about legal data that the general models were going to serve in a very unique way.
Max, since you're here, can you like explain like what was what what was that?
So I think it's it's worth to go back to 2023 and 2024 when I think part of the paradigm was you should train your own models and like the general models aren't great and fine-tuning is going to be really important for two reasons. We were like [__] that, right? Um one fine-tuning doesn't really seem to work at least on the scale that we were operating right to train the new generational model you had to put billions of dollars into it and secondly there was so much application that you had to build on top of the models to make them useful in your environment and back then I mean just solving like basic data compliance privacy and sort of great file uploads and like great parsing and great chunking and all of these things that was where the value was.
And there was another part of your experience which was that you were actually embedded in a law firm.
And so you were studying the shape of like what data law firms had in a way that was you know Bill talks about this a lot is like does an entrepreneur strike you as a a learn it all? And it was clear that the early Legora team when we invested it was like five people. They were just trying to learn everything they could about how the legal profession worked and they didn't have any bias towards it. And so the other thing Max said which you should you should share with everyone is because they were embedded in a law firm in a windowless conference room in Stockholm.
Sounds great. Uh they had a deeper understanding I think of like how a law firm like the data model of a law firm in ways that most of us didn't. And I mean just to take it back even further like when we started um I offered to buy a lot of lawyers lunch on LinkedIn because I wanted to learn. So like literally cold write them and I would say hey I'd love to meet I'd love to talk about like IP law. I'll offer to pay you your hourly fee and lunch and they were all too nice to like make me pay for lunch. Uh they'd often end up paying for it, right? But that did as as Chathan put it like I think it allowed us to uh work with customers from the very beginning. So the founding team at Legora were all engineers. The first lawyer didn't join until 9 months into the journey.
When you had a lawyer join, had you already sort of like set the plan and the goal for the company? Like was that done without experts? And was that important to do without experts?
So it's actually funny. I mean the the founding and this is a bit of the Legora untold um like the first reveal the company formation was in 2020 and there were four co-founders.
And I was not one of them.
Right they were sort of working on this intersection between AI and law for three years with the early BERT models and even a Swedish trained version called SweBERT but it was impossible to work with uh not only was it not very intelligent. It was also blatantly racist cuz it had been trained on like the Swedish like um like forums. Um some racist data there. When the LLMs like 3.5 came, that was when the moment shifted, right? And so we turned this into a company. Two of the co-founders left. I joined and then we basically said like we're going to work in the intersection between AI and law. We don't know what that product is, but we're going to run like hell in this direction. And funny enough, the first lawyer who joined was a sort of soon-to-be customer of ours. So he was the CIO at one of the big firms in Sweden that we wanted to sell into. And he had built a early version of like a GPT plus the document management system. So basically uh like an LLM that could RAG into the existing precedent and data that the firm was using. And he basically said, "Well, these guys are going to run faster than me, and if you can't beat them, I might as well join them." And that turned out to be a good decision.
Are you surprised by like how strongly the legal market has adopted AI? If I had thought, you know, in 2023, let's say, or 24, like what's going to really adopt quickly? Like, I don't know if personally I would have seen it coming that lawyers would be near the top of the list.
I don't know. I mean, you've invested in this stuff before, too, so I guess this is for both of you. But like, has it been a surprise over the last two years, the rate of adoption?
Yes, it's been like vivid. But, but second, and maybe more importantly, um the law firm market is very interesting because it's it's like this perfect equilibrium with frankly like pretty low innovation. Like if you need to do a a VC deal here in the valley, like you could go to any, you know, the top five firms and you're going to get roughly the same thing. If one of them starts leveraging Legora to offer a better service at a better price point faster, all of them have to adopt it. So the equilibrium like shifts down and then everybody has to move. So what happened in the law firm market was as soon as one big firm in a market uh sort of adopted Legora and went public with it, everybody else had to do the same.
That's not necessarily the same in the in-house legal sector, right? Like if one big bank says it like another big bank doesn't necessarily need it.
But was there something about the process of the way work got done or the structure of it that allowed Legora's product to drive so much value so fast in a way that it did force that sort of prisoner dilemma?
I just think the legal sector was so underserved with great software for such a long time that there was this like a lot of built up problems that we could easily solve with LLMs but they were really hard to solve like pre-LLM. I also think you guys had a great insight early on which was that there was like a a difference in respect to the customers that lawyers are really smart. They're extremely well educated. They're tech savvy. They're not programmers but they're very tech forward. They use like the latest software. They use the latest devices. And so they were all going to be playing with ChatGPT and Claude.
And so if you showed up with a legal AI product, it had to be better than the foundation models. Otherwise, they were just gonna say, "Why are you deserving of my dollars?"
And Microsoft Copilot rolled out very quickly, like every law firm in the world is a Microsoft shop.
Like everybody works with Outlook, uh, Microsoft Word, and you know where they store their documents basically.
What have you found it like to the point of you have to be better than the models if you had to break down like as a vertical AI application what have been the things that have allowed you to just be so much better than the models that it's worth you know the incremental investment?
So I think in the beginning there was a lot of just foundational problems with the models like you had to guard rail them very hard to make them useful you had to build citations you had to build good RAG systems you had to overcome context window problems and there was a lot of rate limits issues. So you had to like juggle different models for different types of tasks. There's just so many incremental basic things to solve. I think as time has progressed um our product has moved further away from what the foundation models are and much more into this enterprise-wide platform where you know we're going to transact billions of dollars of legal work on the platform. And we've moved from building a lot of the agent work ourselves and we sort of let the models rip a little bit more like Open Claw or Claw Bot or whatever it's called these days where like with Opus 4.5 and Opus 4.6 uh there was an extraordinary difference in in level of intelligence and like instruction following capability and so I see our job as let's provide the model the right environment and the right tools and skills to leverage and then let's build a UI and a and an interface to the rest of the business so that they can all leverage it uh comfortably and with a lot of trust. I do think that, you know, the model capabilities improving so quickly makes us run faster because we have to be three standard deviations ahead of any general capability.
And that's like a very good motivator.
As an as somebody that's invested in a lot of software companies, like one of the unique things about an AI software company is that it's tactically built differently than a traditional software company. And I think it's becoming more known now, but when you guys first started and you guys built up this org, the way you designed the org made a lot of sense for the product you were building and what you just described, which is like we need to deeply understand model capabilities and then we need to bring that to our customers in a way that's deeply differentiated, which as you explained to me we need to invest heavily in understanding the models, which then would lead to understanding what to build. But as models got better, your features may not matter in six months.
And so talk about how that led to an organization that was heavily technical, heavily engineering and researcher led and for as a company as big as you you are, you have very few product people. The number of product people you have essentially rounds to zero. You have like leaders, you have a couple of leaders, but that's it.
I mean the founding team were were three engineers and so you know the the most natural hires were let's grab all the smart engineers that we know from college and let's add them into the org. In the beginning you know we had to build our own agent framework because like LangChain and these things that we initially built on like couldn't get customized to the level that we needed back in 2024 as we understood more about the model capabilities but also of the problems we wanted to solve like let's take due diligence as an example um it's really hard to solve of a due diligence task in a chat-based format because you need to review hundreds of documents and hundreds of documents are never going to fit into the context window of a single model call at least not back then and you know probably not now either. So we built this new product that we called tabular review big matrix where you would throw in you know tens of thousands of documents and you'd throw in all the prompts and it started running all of them in parallel and what we basically did was we just said okay three engineers you're now on tabular review this is your own company run over 10% of the EPD or at Legora are XYZ founders so our head of engineering Jake who joined he was a solo founder in YC our VP product Adrian was also a legal tech founder in YC and happened to be both GC and a lawyer. And so as we've progressed, engineering and product has sort of stayed at the core of of who we are and what we do. And I also think that everything else is sort of an expression of that. Like we can only market what we actually build. We can only sell what we actually build. And product lead compounds.
I think as as you put it in the beginning, we did not show up first. Like Legora was not the first product that many legal teams looked at because they were earlier entrants. So we knew that we had to show up and be best and if you want to be best well then you need to invest in product. You need to invest in engineering. And I think you need to build that culture of like reliability first. We actually had a time period in the company for for 6 months where we didn't sell basically because we weren't ready to like hit the gas on onboarding a thousand lawyers a day and knowing that the product was going to keep up with that. So we took the early hits of like investing in that.
Talk more about that period specifically. You know the seed round you did with us was in March of 2024. The product went to GA October 1st, 2024.
And you called me early September 2024 and said, "You need to come to Sweden because all of us need to sit in a room and just talk about where we are and what we need to do to get this thing out in a month." And you know, we came and we sat the whole literally the whole company, which wasn't that big back then. It was only 10 people. Yeah. And so it was like the whole company, the founders, chicken wings and beer.
Yeah. And we and peanuts actually. Those were the three things served.
And there was like a very open dialogue of like how do we get this thing out in 30 days because at that point cuz you were essentially, you know, weren't facing the market test. You were building there were 10,000 things you could build and that the sort of like outcome of that discussion was that we're only going to focus on three use cases.
So talk about like well one you know you calling me to tell me to come to Sweden to have that discussion and you actually showing up.
I did show reflecting on it that was one of the most important things that that you did in the company the founders did in the company was at that moment say we have 30 days to go we're just going to sprint at these three things not the 15 things that we could do.
Yeah. So I think there was this feeling of like you got these LLMs, they're so powerful. We learn about all these use cases in the firms and with the clients that we work with. Let's go solve all of them. Like wrong decision. Like you can't solve 15 things at the same time. And so we had to to kill a few darlings and we had to like really double down on the stuff that we thought was going to work. And we looked at on the market and we basically saw a few things that were really working like as a paradigm for LLMs in legal. One of them was this big sort of tabular extraction. Another one was embedding it deeply into Word and Outlook. So basically having Legora be accessible wherever the lawyer is already working. And we were still called Leia back then like this was very early. We took the entire company. We had like a town hall and I remember showing some numbers where like a particular company that just had one of these features were doing more revenue than us. We were doing like 1.5 million at the time. That felt very painful because we thought that we had a better suite but we didn't have as much revenue because we were based in Sweden and we were sort of mostly selling to still European firms at the time. So we just said let's do these three things. Let's do them better than anyone else and it's going to be worth to buy our suite over anybody else's. And so I wrote this like very short product manifesto, send it out to the entire company and we sort of rallied the troops. I think it was off the back of that that we had our first quarter where we doubled revenue. So we went from like 1.5 to four. We're like, "Oh, this is like ripping and it's flying off the shelves." And then in Q1 we had another quarter where we doubled where we went from like four to eight. Like whoa okay now we're talking and if it became time to launch in the US uh we hired Patrick and Evan who joined from a competitor and we sort of had our first boots on the ground in the US and then we felt like okay what we have is like a winning formula so we just need to crunch it out everywhere and now I think we're at another interesting point in time where we've built all these different tools but the paradigm from now onwards is humans are are probably not going to work with all these tools like basically agents will leverage the tools that we built. So I remember early when MCP came uh our CTO basically went well now Legora has two users it's human users and agent users and every new feature that we build has to be able to uh cater to both and now we're seeing more people like basically use our agent that uses the tabular grid or like our agent who uses our word editing capabilities than humans actually going and using those features at all.
Chathan made a cool point to me recently, which is that um because you have we were talking about how you know companies that are pre-AI and companies that are just fully AI native just have to be built differently in various ways and the fact that you didn't build a pre-AI company I think gives you sort of like you know this unshackled mind to like you're not even trying to think about some past alternative you're just like given what's in front of me what should a company look like and you know you talked about how like having YC founders inside the company has been helpful and I'm sure there's like a lot there but I'm curious about like what are like the main tenants that you've observed you know cuz now you've probably hired a lot of people who did you know work and built companies pre-AI what do you think are like the main tenants ideas cultural concepts that have been important to you just to like make it work in like a fully AI native world.
Yeah so I think this idea that Chathan brought up around um like you have to be willing to like kill the stuff that you've done in the past is very important because I think in in more traditional software you had to build the foundations and then you build the stuff on top of it and you sort of kept building the stack and in that world it was also very good to have like a technical architecture where one feature would rely on the same microservices as like other features but the problem is in AI like maybe that feature now needs to scale really really quickly and the cost of writing software is so low that it's basically better to build your own stack for like each thing and now that we hire you know finance professionals or even like lawyers internally to Legora or we just hired our first like tax person.
I think they come with a set of ideas of like oh this is how I used to do it in my old company and everybody's forced to relearn I think and also question what their value is on top of the general model capabilities in a way which is like very painful.
Totally. Brett Taylor talked about this on this podcast too basically that like people are going to build something and 6 months later we might just kill that thing and everybody needs to be comfortable with that which I think historically would be a lot of painful internal conversations so like do you have to change is that like a different culture for people.
I think it's a different culture completely I mean I think the culture is you don't maximize for your function like you maximize for the company always and I'm very upfront with every exec who joins Legora that in a way like you're joining with an expiration date and you have to continuously prove that you like scale out of that in a way because the company is scaling so exponentially and I don't know if it was Mark Zuckerberg or somebody talked about like let's hire people with high y slopes and not like high y intercepts. I think about that a lot mostly because I've had to do that right like I did not join or start Legora with a lot of experience but I've proven that at every new point in time I've scaled with the business and so other people in the Legora needs to do the same and I think that goes for every function and you know I think an engineering team that's shipping the amount that we do previously had to be like 500 people and now we can get away with being 50 and I think there's even a question of like Do we need to be more than 100 engineers or is the bottleneck here really knowing what to build and building it in the right way and designing an experience that works for you know hundreds of thousands of people that we now have on the platform. I think the paradigm is like shifting all the time. What's nice about our work is that engineering is sort of a a roadmap of what's going to happen in other industries too. Like I think the general models have come the furthest in coding but also those organizations are very quick to adopt and shift and so like engineering orgs are today looking slightly different and I think we can expect the same in you know legal organizations.
Two things you brought up that you should it would be great if you could dive into. One is Legora doesn't really have a long-term roadmap like you guys react and build today and you know when you first got started you had this like nearly weekly cadence where that's how long you would roadmap to and these days it feels like you almost roadmap on a daily cadence and so like things change tomorrow like you wake up and it's like we have to do something different talk about that lack of roadmap and then also the other thing that you've invested heavily in is just understanding model capability and the sort of proprietary eval infrastructure you've built where you've had these conversations with the foundation model companies of how you're able to identify latent model capabilities that they themselves are not aware of.
I mean on roadmap like way back every new model just like unlocked new things right and when we got early access to GPT like 4.5 and you just realize that holy [__] like now it can finally draft like an end to end thing and we don't need like all these harnesses and like things around it that's amazing like let's unleash it in a way that that works.
By the way to do that you need sort of like a low ego organization cuz you build all this IP and all this software and you're like, "Okay, now the model can do it."
We worked really hard for 6 months.
We're deleting everything.
Yeah. It's incredible.
But I think a lot of the things that we have built, we know that we're going to delete someday.
And I guess you kind of need people to opt into that at the front end for that culture to really work.
We've also talked about it as like if we were if we're here today and we start building for the future that's over here, like that's too far out. Like our customers are not going to adopt that. They don't understand it yet. So we need to take them on the journey and we need to take them on the path of being successful which I think every iteration cycle now is shorter like back in 2023 2024 I think it was like slightly longer like you'd have a quarter or two quarters because the models weren't moving that fast every upgrade was pretty incremental but now it like flipped Opus 4.6 flipped uh in capabilities so now we have to revisit a lot of the things that we built.
Do you know what the next flip you're waiting for is? Like is there a thing?
So, actually I don't think so. So So it was funny. I was at the customer advisory board at Anthropic yesterday, which is I'm wearing my my like Dario shirt here.
Thank you. Most of that conversation was about the models are now intelligent enough where they're no longer the bottleneck. The bottleneck is all of the software around putting the models in an environment where they can execute and do work and humans can review that work in like a trustworthy way. They're seeing that across basically every single vertical and every single company. So I don't really think that we're waiting for new model capabilities anymore. There's like nice things to have like it's nice to have better context windows. It allows us to, you know, do less garbage and context.
Management or like when you overflow the context in memory and so on, you have to deal with it to refresh it. So there's nice to haves, but I think we're at a point now where we just have so much building in front of us in terms of bringing the model capabilities into our world that that's where all of our focus is. I think on sort of discovering what the models can do, we thought very early on that evals were going to be important and both building up like an exercise of building new evals but also building out evals for all the use cases that we want to cover because in the beginning it was a lot of like, oh, how good is Gemini, how good is GPT, and so we had to test them on the different evals. And a lot of our customers actually contributed this, so they would give us manual tasks that they used to do and they tell us, here's the evals and we're going to call you when we can get to 100% on these evals. And I actually remember it was a funds related use case, an LPA key term review report that a Danish law firm was spending like three days on. Basically like an associate would spend three days putting together that report. In summer of 2024, we had like 60% accuracy on that task. By the end of that summer, we had 100% accuracy. And once you got to 100% accuracy, I mean that task is done. Like it's over. I've adopted this mentality internally that if AI can do something, it will do it. And so our product, we think a lot about solving legal tasks end to end and once a task is conquered, it's done. Like we just strike it out. And we're on this path of solving more and more complex tasks. Like you start with NDAs, but at some point you get to full-on share purchase agreements which are very complex, but we're going to get there. I think the question for these organizations who are maybe more traditional and trying to keep up with the pace of AI is how do you do that while at the same time do your normal job, right? I think a lot of the organizations that we work with really struggle with keeping up with the technology uplift even like our developments. And so we're struggling by getting all the latest models and then we're turning that into product and then they have to adopt it and then they're customers and it's...
Yeah. This is a question I think for both of you as I'm listening to you talk. I'm sort of like, you know, I can sort of see the hill climb that you're on where you've like tacked one part of it and the next one's coming and the next one's coming. And one of the things I'm thinking about is for, let's say, a new startup in legal, what would the right strategy be for them? Like how do you possibly get into the mix fast enough or all of these things and...
Exit to sell to Legora. That's a good one. How urgent is it to grow really big really fast for Legora given all of the dynamics around Chathan? I'm curious like how do you think about this? Like is it the same urgency as always or do any of these dynamics mean that like getting to real scale is more urgent here than other places?
We can go back to sort of launch day October 2024. So when they launched, you know, roughly the ARR of the business was rounded to a million dollars. If you just go back into that moment, you know, there was this exercise of should we make a budget and what we all decided around the table was there was no reason to make a budget because we don't know anything about the market. We don't know if people even like our product. We had instincts, but we just needed to go literally as fast as we could to get the product in as many hands as we could because ultimately the whole theory of the company didn't work until we got product feedback. And so that was literally the aim. The aim was like get this out as quickly as possible in as many hands as possible. And I think one of the things that Max did, again, this was like it's cliche to say it's first principles thinking, but it is because the team was unbiased by how to build a software company. And so one of the things that you learned in SaaS was the way you do pilots is like you would go in do like a time trial pilot where you would give them access to the application and the minute the trial was done you would turn it off and then they would have to make a purchasing decision.
A big thing that happened with Legora is they would go put Legora into your organization and whatever you put into Legora they would leave behind even if you didn't want it.
And so there was this idea that like, hey, you adopted AI, you did stuff with AI, you built some practices but you're not like whatever skills you built or whatever IP you built, it's kind of yours and we can leave that behind, it's not a big deal, it's like it's your skills, it's your things that you learned. And then Max went around and just gave people like 30-day pilot, 60-day pilot, whatever they wanted, 90-day pilots and they would run these competitive pilots, right? So they would say, okay, there's a couple of companies on the market, we're going to want to AB test all of them because it's really hard to pick just based on the feature set on your website. And in those pilots, I think we did an extraordinarily good job of delivering value. And so when the sort of 30 days were up, like if we shut it down, it would be a riot, right? Like people would roar and they'd be like, 'We've never seen software adoption like this in a legal organization. We need this and we need it now.' And in those pilots, we would demonstrate much better than any other company the value that the product and the service around the product could bring. So we hired all these lawyers who are now called legal engineers. It's a great term I think, forward deployed legal.
I was going to say what about FDLE?
FDLE. That's right. FDLE. And they're amazing. Like they're the most tech-savvy lawyers in different organizations who don't want to make partner because that's one type of life and they want to work in a tech company and now they get to work with their practice that they're amazing at and technology and then they get to work with the best legal organizations in the world and driving that change.
