Thank you all so much for attending this talk. As you heard from the intro, my name is Chathan. I'm one of the general partners at Benchmark and honored to be on the board of Legora, where we led the seed round in March of 2024. Max is a remarkable CEO and is building a remarkable company. So, we're going to spend the next 25 minutes talking about that. Max, why don't we start with telling the audience what Legora does and we have questions for you.
Hi everyone. It's great to be back at Slush. We were here last stage on the small screen, but now we're on the big stage. At Legora, we're supercharging lawyers with AI. We launched in October 2024 and since then we've scaled the company from 20 people to over 200. I think some of the numbers were outdated. Last time I was on the small stage I announced a new deal and I'm also happy to announce that we moved from 400 clients to 550 clients in the last month. So we've been quite busy.
Amazing. Amazing. So start from the very beginning. What was the inception idea for Legora? How did you guys come up with the idea? And tell the story of watching your friends that were lawyers, you know, interns that were doing really terrible work and you guys just wanted to go hang out and they weren't using AI. So start with that story. I think that's an awesome story.
Right. So the seed that eventually became Legora was founded back in 2020, but that was with a very early generation of the BERT models, the first GPT versions. And it was effectively impossible to solve any real and meaningful tasks with those NLP models. And when ChatGPT was released and GPT 3.5 became available on the API, we said, "Hey, let's go retry." And so one of our good friends, he had spent the summer in a courthouse basically being tasked with summarizing court cases and finding similar court cases. And vector search had just become a thing. So we said, well, why don't we try and find similar ones? And the first task that we set out to solve was understanding our own stock option agreements at different startups. And the idea was really quite simple: that's combined legal work with AI and a sort of rush in this general direction. And then just one month later, we ended up signing with the biggest law firm in the Nordics, Mannheimer Swartling.
So the thing that stuck out to me when we first met in March of 2024 was how you guys literally camped out in a windowless conference room. At this amazing law firm in Stockholm. They have beautiful offices and you literally picked perhaps the worst conference room I had ever seen.
It's the one where they didn't take the clients.
Right? So, I think this was a remarkable story of you find this inspiration because you're watching your friends that are lawyers not use AI when they clearly should be using AI. And instead of just sort of building a product in a vacuum, you decide to go to the biggest law firm you can find in Sweden. And then go hang out in their basement. So, how did you get that first meeting with Mannheimer? And what did you say to the managing partner to let you come into the basement?
So, this is a funny story. At first I couldn't get a meeting with him. So, one of our first angel investors said, "It's harder to break into Fort Knox than to break into Jan's calendar." But what I ended up doing was I texted him and said, "Hey, I'm about to meet with your second and third biggest competitor unless you have time." And suddenly he had a slot.
So we go there and I'm sort of a little bit nervous. And we talk a big game and he had said, I think a couple of years earlier in the news that he thought AI was more artificial than intelligent. So I was like, "Oh, this is going to be a tough crowd." But I actually mentioned that and I said, "Jan, no, it's this time it's different." And I don't know anything about legal work and frankly you don't know anything about AI so we probably need each other. And then he said, "Max, I'm a guy who likes to take decisions on the spot." So we shook hands and then we ended up moving in and I basically became the guy who would bring my laptop to the lunch canteen, sit next to somebody new and go try it. I want to see how you use the software, right? And I think that gave us the sort of initial six, nine months that we needed to incubate and build up the company.
I think what's underappreciated about Legora broadly, you know, today it's one of the fastest growing AI software companies in the world but I think what's underappreciated is how long you invested in the product. So you were playing around with this before a company was formed with prototypes with the early AI models you were trying this out with your friends then you get elevation that you know we should actually try to create a company around this you create the company and for 6 months you're building literally with lawyers inside of Mannheimer. And then we meet of course once you're in Y Combinator March of 2024 and what you show me is a prototype, powerful prototype, looks great, but it's a prototype. You then take the decision to continue building from March of 2024 to October of 2024.
And so like the product doesn't launch. So it's a long time of just iterating on the product without releasing it to the world.
So what were you waiting for? What was the constraint that you realized you needed to see before you felt like the product was ready for the world?
Yeah. So, I'm not sure if I've said this publicly, but in our first board meeting, which was in May 2024, and Redpoint had just invested $25 million after meeting me for, I think, an hour in this 25-year-old kid. And I sit them down and I basically say, "Hey guys, we're not going to sell for the next 6 months." And I can sort of see them starting to sweat a little bit like, "What's going on here?" And I think we love Redpoint. They're amazing. They're amazing. But like the truth is you get one shot when you show up and when you want to implement your product and if it's not great and if a partner in a law firm has a bad experience they're not going to try again. So we basically used the fact that Europe goes on vacation between June and September and we said oh we're sort of at capacity we can't onboard more clients let's take you after summer right and we wanted to be in a state where both the product would scale but also we sort of felt more comfortable in the design so that we could onboard like a thousand lawyers every single day and that's what we started doing. I mean since then we've more than doubled every quarter and it's been wild.
So, were there technical decisions that you made early on that you look back on and look, legal AI as an industry, you know, if you reflect back on 2022 and 2023, many companies had the idea that they wanted to build AI for lawyers. Many companies raised a lot of money. Released a lot of products that didn't work. How did you know when your product worked? Was there an aha moment for you? Was there something that a lawyer had to say to you to say like okay the product is now ready? And then you know we weren't obviously paying attention to the competitors at all cuz you were just focused on your own product. Just walk me through and walk us through like that moment when you were like okay now it's finally ready.
Yeah. So I actually I feel like I remember a client meeting in Copenhagen and we were pitching a litigation partner and I showed him the product and he basically turns to me and goes like, "Dude, we're unless we buy this right now." And I go, you know, you said it like, you know, at that point they're like, oh, like they see the value like it's so easy to see. And we also had the privilege of hiring a lot of great lawyers into the team so when they started using it every day and like, okay, this is really good then of course it's easier to go out. And I think you're right we were sort of building in isolation and some of the early decisions we made were let's not fine-tune models let's run at building the best application has to be extremely user friendly in the very early days. And when we started pitching to the US people would go, "Oh, this feels like Scandinavian design" and we would go, "Oh, like yeah it is," you know. And I think when you sort of put it in front of a practitioner and they can pick it up and start to roll with it then you, "Oh, okay. This is going to spread like wildfire." And now I actually met somebody here who said that their dad is a lawyer and he wouldn't go to work at a firm like if they didn't have Legora.
And here in Finland, which I thought was really fun.
That's really cool. There were two things that you said in 2024 that were counterintuitive at the time. Number one, you were deeply against fine-tuning models.
And number two, you explicitly said we need to be better than general purpose models and general purpose chat interfaces for lawyers. Like that's the main competitor. It's not another legal AI company. It's sort of the generalized horizontal tool that's the competitor. And unless you prove yourself to be better at the specific task than the generalized tool, like why should people buy you? Yeah, those were two interesting things that are now consensus in 2025, but in late 2023, early 2024, you were one of the few people saying that. Where did that come from? Was that because you were so familiar with AI? Was it because you were messing around with this stuff for so long prior to starting the company?
I mean, I think it was over I think from a timing perspective, we sort of started building before ChatGPT had an enterprise license. We started building before Microsoft Copilot had launched which meant that when those things came it was really apparent that like okay of course this is going to get widespread after they solve some of the initial challenges around European data hosting processing confidentiality and these things. So we knew we had to be better from the start. And I also think that those tools have a disadvantage because they need to solve for so many use cases. I mean the main use case for ChatGPT is homework. And that's the biggest use case. And we're only focused on one type of user which I think gives us a lot of focus.
And okay. So now we're at the point of launch. It's October 1st and you now have reported to everybody like, "Okay, the product is ready. We have this incredible meeting in Copenhagen. The Copenhagen lawyers are on board." You launch. Walk the audience through the first 30 days of demand because I don't know that I've ever seen something like that before in a software tool selling to legal. So you're selling to lawyers in law firms and lawyers in enterprises. So you know these audiences aren't used to buying software. You're not going and replacing an existing piece of software. This is like they're it's net new. There's buying processes. There may be procurement all this kind of stuff. But you launch, you tell the world you're GA and the first 30 days is unlike anything I've ever seen. So just talk about what happened.
I mean, look, we're an enterprise company. We're a very sort of B2B enterprise-heavy motion. And so, I think what we had done very successfully was sort of build up this, you know, we built up to this release and we had built up this pressure and then we said, "Okay, we're ready. So, let's start onboarding people." And at that point, we had built a quite large team of call them legal engineers, but it's ex-practitioners in our teams who were ready to onboard everybody. So, we showed up and we didn't show up with a product and say sort of here, good luck. We showed up and said, "Oh, you're an M&A lawyer. You want to solve this stuff. Here's the prepackaged prompts and workflows so that we know that you're going to get to success." And after that, we sort of built out every practice area. And we just said like open the gates. And since then, it became pretty clear that okay, to meet this demand, the company is also going to have to scale. So at that point we only had the office in Sweden and a couple of representatives here in Helsinki. And very quickly after we started getting a lot of inbound from the US and at that point we sort of had to level up from how do you serve a big European law firm to how do you serve sort of an AM law 200, 3000 practitioner law firm. It's a completely different motion. And when we had nailed that and I think the great thing about working with lawyers is most of them know each other right and so if they have a really great product experience they're going to tell their friends over dinner and that's what started to happening like sure this is an enterprise motion but it's very much oh you have that for your team and your clients know about it I also want that thing and then you know you tell the innovation team procurement goes really fast because the partner says that they want it and then things just spiral right so the first quarter we knock it out of the park and it just basically goes wild.
And then, you know, we're sitting in the board meeting in December 2024. Um, it may not be a board meeting. I think I flew to Stockholm and we're sitting around talking about how to do the 2025 plan. And at that moment in time, you decide to do something that I think is really interesting, which is not to make a 2025 plan.
Because we don't know what 2025.
We did not have a budget for 2025. That felt like waste of time.
That's right. So then it just became we're just going to attack quarter by quarter.
Month by month, week by week.
And then the first quarter in market. So went well. Now it's and we did the rebrand. So that was a big thing for us, right? Like we used to be called Leia. Which I thought was a great name. Loyal to the law in Latin, but we changed to Legora in Q1.
And then Scandinavian design.
Exactly. Yeah. Sounds enterprise, right?
