So, Max, when we worked together at Norrsken, there was one thing that we talked a lot about, which was speed of execution.
And here we are two years later. You've built a unicorn from being an intern in just two years. What took you so long?
I mean, I learned everything I know from the Norrsken Accelerator, of course. But look, I think when the process of working with 10, 20 different companies, as we did in a sort of very pressurized two-month session, basically taught me — or I think you saw — how quick you can grow in two months, right? And then if you propagate that over 24 months and you keep the same pace of change, then I think you've got something real there. In Y Combinator, there's this sort of post-batch depression after you get the pressure from working with all the other companies in the batch, and then you need to do it on your own. I think the moment we started Legora, I actually sat down with my two other co-founders and I told them, 'You've never worked as hard as you're going to have to work working with me.' And I think that set the pace for the first summer, and then just like...
So I think we're going to use this session to just creatively and concretely jump into the different parts of building a unicorn in such a short period of time. But before we do that, maybe you could just give a 30-second pitch to what Legora is for the ones who don't know.
So Legora is an AI-powered workspace for lawyers. We basically help them review, draft, and research at speed using generative AI. We started the company a little over two years ago in Stockholm, and since then we've grown from three technical founders, three engineers, to a team almost close to 200 people spread over Stockholm, London, New York, and we're soon to open our office in Australia. And it's been a real journey working from sort of mid-sized Swedish law firms to now working with the largest law firms in the globe, like Linklaters. And the change management and the work and process re-engineering that AI brings is crazy. And I think the legal space is one that hasn't seen a lot of previous efficiency gains, digitalization waves. And so when AI came, it sort of hit a bit from the left flank.
And law firms and in-house legal teams have sort of stressed about how can we incorporate this, and we've been their partner on that.
And you knew nothing about law.
Absolutely nothing. In our first Y Combinator interview, one of the partners, Tom Blomfield, who was the CEO and founder of Monzo Bank, asked us which type of lawyer are you focusing on, and our answer back was, 'Oh, are there different types of lawyers?' That was not a great response to that question, but I was very conscious in the beginning to both learn myself and to also be associated, from a brand and company perspective, as if we were working with the best lawyers. And it was actually through an introduction by you that we got introduced to Mannheimer Swartling, which was the biggest law firm in the Nordics. And then we started working with them, we got a bit of brand recognition, and then things sort of took off.
And just to set the stage a little bit, you're in a head-to-head competition on life or death with a company called Harvey in the US. It's like Pepsi and Coke kind of, and it seems like they were born out of a legal context a little bit more than you were. Why do you think that you're going to win against them in the long run?
So maybe I'll just explain why. When we started, we had a $50,000 angel check, and that same month Harvey came out of stealth with $20 million from Sequoia and OpenAI. And then one week later, another big company in the space called Casetext got sold for $650 million to the biggest incumbent. So we were sort of the very, very, very small fish in the little fishbowl of the Nordics versus these two behemoths of American companies. And I think, frankly, we were quite lucky that they were so unfocused. Harvey was started almost from this research angle that training your own LLMs and training a legal-specific LLM was going to be important, which we didn't think at all. So they basically spent their $20 million on fine-tuning models, and that was their moat. But it turns out LLMs are no moat at all, which we thought already back then.
So from the beginning, and also given that we weren't lawyers, we were sort of notoriously product and engineering focused. So it was like product, product, product comes first. And we were quite lucky that they were more sales-driven and spent a lot of time on research. So when we think ahead and why we're winning a clear majority of the head-to-head and A/B tests we go into, I think again it's product focus. If you put the different systems shoulder to shoulder, there's very clear differences, and product moat actually compounds. So if your product is already better, then it gets better over time, right? And also, if you're moving 100 miles an hour and both companies are doing this, you can't afford any U-turns. So I think that has been very important from a product perspective, that every bet we make has to be correct.
Yeah. And how do you... that little detail, obviously. Yeah, that little detail. So, I mean, I've been fortunate enough to come to your bi-weekly product update meeting, and the cadence of delivery is just insane. But how do you manage to balance that, like short-term, let's get the product now into a shape where we can sell next week, versus like the two-year strategy? And do you have time to focus on that?
So I say that I think there is no two-year strategy.
And one of our other investors is the VP of Product at OpenAI, and they have within their team like a two-month product roadmap. That's it, because the underlying models were changing so rapidly that it was very hard to be very confident in your direction. But I think we had a very healthy push and pull between product and sales. So in the beginning of the company, I took on the more characteristic CEO sales role, and my two co-founders were engineers. So I was out there and I was selling stuff that we absolutely didn't have, running back to the office and saying, 'Hey guys, we've got to build this, and we need it in two weeks because then I'm onboarding this client.' And I think we bred that into the culture very, very early on. So even now we do company-wide — and we're almost 200 people company-wide — demos every other week where everybody in engineering needs to demo what they have built. And if you haven't built something, demos are very, you know, it's like the shame walk from Game of Thrones.
Okay. But we don't undress.
Yeah. You don't have to ring the bell.
But so before we get into like sales, and it sounds like a pretty harsh culture, like how do people have fun? Is it hard to maintain people?
I think people self-select for this, and we're quite open about it, and we're quite upfront with it. And you know, sometimes if I'm interviewing a consultant from McKinsey, I usually draw this curve. It's like, okay, you have tech highs where you have the dot-com and the mobiles, great times to build companies, and then you have the valleys where it's probably better to work in consulting. And now we're definitely in this sort of up wave. It's like, dude, it just makes sense, come and work really hard. Like the work you're putting in now, you'll have exponential impact and leverage on. And for the engineers, it's the same thing. Like you want a nine-to-five, go to Klarna or go somewhere else where you know you can build on a small feature and you can get, you know, like do that, that's fine. But if you want to work really hard and you want to build really impactful things and you also want to be part of a journey that probably doesn't happen that often, right? Like you had Spotify, you had amazing companies being built. Now there's a new set of companies on top of LLMs. We're definitely one of them. And so you sort of self-select for people who want that.
How do you incentivize them? Like is it just fixed pay or everyone has...
No, everybody has stock options in the company. And if there's anybody in government, like please fix Swedish stock options because they are awful to work with.
Now you haven't seen Spanish stock options.
So let's get into sales, because I think this is one of the things that, I mean, you mentioned that Harvey focused on sales quite a lot, but then they didn't have a product, and you didn't have a product either, but you built it. But curious to hear how you, as a founder, are involved in sales. Like very concretely, do you take meetings every day or like any time of the night? Like how does it work? How do you think about this?
So I've done everything from, you know, onboarding an Australian firm in the middle of the night to, okay, we need to get on a flight to Mumbai to meet with the managing partner of that firm to close the deal. And now my rule is I'll work on the top 10 deals every quarter proactively, and then for every deal over $500,000 I can be involved one time, and then the account executive can choose when I should be involved. But in the beginning, almost the first $2 million of ARR I did myself solo, and I did all the onboardings. And I think this was quite useful for me because I got to know our customer, and this was really important because that then allowed me to scale the team. So our first set of sales reps actually all came from McKinsey or a BCG background, and I think that was important because our customers weren't just looking to be sold to. They looked for somebody who was going to take them by the hand and say, 'Hey, look, AI is really hard, but together we're going to figure this out, and we're going to make sure that you win, and if you win, we win.' So you actually needed a bit of this sort of partnership, change management, strategic chat with the partners to get them involved.
And then, of course, you know, then the product becomes clearer, and it's easier to sell on the ROI and, oh, here's all the tasks that you can solve. But in the beginning, it was much more on the relationship.
And super concretely. So you're selling across the world. I mean, you can basically just translate the model to anywhere. How do you select customers? Which CRM do you work with?
