Why Legora isn't building its own model

0:03Hello and welcome to the Sifted podcast. I'm your host Freya Pratty, Sifted's Associate Editor. We're recording live from Tech Barbecue in Copenhagen today in a pod booth at the side of the event. So if there's any noise around, that's why. And I'm joined today by Jacob Lauritzen, CTO of Swedish legal AI firm Lugora. We're really excited to have you on the pod. Thank you for joining us. Thanks for having me. For anyone who's been living under a rock, give us the Legora pitch in a sentence to start us off. If I'm explaining Legora to my grandma, I'll just say that we are in the business of making legal work much more rewarding and fun and exciting.

0:43Great. I like that. But that's missing out lots. You have raised a lot of money. 550 million dollars in March, I think, led by Excel. You also have General Custer List. Reports of more to come, but I won't try to get you to answer on that. And five acquisitions so far this year, I think. So that pitch is condensing down a lot of achievements. Yes. It's been a lot in the past two years. It's really impressive. I'm interested then in what your, so your CTO, what does your day look like? what does it look like to work at one of these of one of Europe's kind of buzziest companies? Do you have like a particular productivity routine?

1:25What's it like? Take us inside Lagora. That's a good, well, so for me personally, productivity routines, yeah, I mean I have a, obviously I'm super air pill to say everyone at the company. I have a my own, sort of, I call it my EA digital EA, which is like a collection of skills and an agent that I run every morning, which triage just my email and my Slack and what's happening in linear with the different projects, highlight any important architectural decisions that would be made in our pages that have sort of these types of things in Notion. And I run that once a time today, and then I don't really look at my email or any of these things because I just get it flagged and automatically sort of unsubscribes and handles all the stuff that I would do.

2:10I try to have the mindset that I should only do things once ideally and if i if i see myself doing it multiple times i should automate it try to jump on whatever is the most urgent whatever's the most whatever's the biggest bottleneck right now to moving faster and iterating faster is where i've tried to focus my time doing something once and then automating is a really that's a really good rule talk us through how your team is organized i think our readers are really interested in the kind of new uh cohort of ai companies how they're organizing engineering differently perhaps in relation to product like how many people report What do your departments look like?

2:43Talk us through all of that. So there's many different talks right now. You need to have small teams of three people or only one people, or you need to have huge teams of 30 people. It's like, I don't know if we've really conversed on what jobs are, or if it's really that difficult from what it used to be. For us, the key things to organize for, and I don't think this has really changed with AI, is I want very clear accountability and responsibility. so everything there's there's always one person who is accountable and responsible and then i want to give them as much ownership as possible and interfere as little as possible basically giving them a very clear target give them everything to succeed to run at that and then optimize there we used to have a lot of small teams and now we sort of group them into larger groups and because what ended up happening if you have a bunch of really small teams they're optimizing for their small little part of the product um so let's say you had you know i'm making this up let's say you had a team that was focused on um onboarding or you had a team that was focused on um automations or something like this they would really only care about making automations great or making onboarding great but sometimes you have to rethink the entire thing or there's a bigger thing that neither of them see because they're just focused on their own product and so we've organized it into larger groups of you know let's say 30 people are bigger than that and there's a product responsible and an engineering responsible to optimize globally across all that and i think that's extremely good because then there's lots of flexibility within those big groups so like maybe sometimes onboarding doesn't need that much work and then we can focus on all the other stuff instead so i think the flexibility is very key right now and do you see yourself i see lots of talk about yeah as i said product team morphing within engineering i think we're definitely uh allowing pms to ship stuff and to actually make code changes um and that's just I mean, our engineering always focused a lot more on the sort of system, rather than just writing code to build our product.

4:39They're focused on like, how do we create the system around agents that allow them to build code really fast, really good. And if that system works really well, then that also opens up for designers and PMs and theoretically salespeople and whoever else in the org to actually ship code, or at least to initiate it and then get it prioritized. And engineers can say, no, this is a horrible idea. Let's not do it. Or yeah, this probably is good. and then an agent can pick up and do most of it and probably can deploy it and get it out there. So I think the lines are blurring between product and engineering to some extent because PMs can ship stuff.

5:10But I do think that we still need both because engineers are very systems focused and product people need to be really good at going out and talking to customers and sort of doing the discovery of what's the actual problem that you're trying to solve. Not just, we're not just going to build what you're asking us for. We're going to figure out what you're actually trying to achieve and see if there's a better way to achieve it. Yeah, that's interesting. I talked to a lot of VCs who comment on their products, like iteration cycle, going to the customer are worried that that might be lost when some of these teams collapse together.

5:36So you're right, that's a good thing to point out. Let's go through some of your recent announcements then. There was the integration with Google very recently. Explain what that means for Lagora, for you guys. Well, so Google has the ecosystem enterprise. Actually, there's lots of people who use Gemini for their enterprise workloads. And it's just another way to access the depth of the Gora and the legal expertise that you get from the Gora to do that directly from GeminiWareSit anyway. And so for us, it's just another way to basically access all the things that we do that are very specialized and very unique and have a lot of depth to make that easier for our users.

6:20Was there a worry when you made that deal about moving focus away from people that come directly to the Gora or that's not a concern? Not, well, not really. The reason being we're building Legora as a great agent that you can access from a lot of places anyway. So you can run from your mobile, you can email it and it will email you back. You can kick it off from other services. Probably maybe you can talk to it. There's like many ways that you should be able to engage with the agent. It should just be someone who's there with you. And then one of those access points is probably going to be other AI tools.

6:52Okay, that makes sense. And you've made, I think it's five acquisitions in 2021. What does that look like from your perspective? Like how many engineers do you take from those companies? How do you integrate tech? So, I mean, it's not, when you look at the number, it would sound like we have this very aggressive M &A strategy that we're like, we need to buy all these companies. That's not how it works. Generally, it's just we happen to meet great people that we really like and they really like us. And we just think it would be much better if we work together to create something even bigger. And so that's what happens.

7:23Typically for most of these acquisitions, it's been maybe adjacent or maybe even not adjacent, but it's a really good team that we want. And so we just bring them, everyone in, and then we co-create something better than either. What's that like for you? What are your lessons in terms of integrating an outside team within your one? That is a good question. I think we've managed to interview and do our duties very well on most of these companies. And the most important thing is that the culture is the same. I think that's like really difficult if the culture is different then integration is not going to work for obvious reasons um and especially when we're still relatively small team and acquisition actually takes up a lot of space and so they need to be culturally very aligned they need to be super bought into the types of people that we want and that will be successful at the gora they need to be super aligned on that and i think that's something that we've done really well and that has been a key to making this a success because if we're just you know buying random companies because they looked good from the outside, we could have destroyed both companies.

8:25Yeah. How do you actually work out the culture fit? We spend a lot of time with the founders. Hang out with them. Yeah, that's probably the best way. Exactly. And we hang out with the engineers and with the product people. We try to challenge them. And I think very quickly, you can also figure out how they've run their company. If they've run their company in a way that's very similar to how we run Legora, then it's probably also a good fit. So it's just spending time with them, honestly. I think some of this culture stuff is like, you have to have a good culture, gut feeling magnet somehow. Yeah, you can't put a kind of, you can't put data on it or anything like that, right?

8:56Not really. More on the like technical aspect of the M &A then. When you're acquiring a company, what do you think? Oh, we can build this. We should bring in a team that have already done it. How do you make that decision? Have you done that in the previous five and what do you think about going forward? So I think it's a mix of, do we think this team is exceptional? If they're exceptional, then we will always buy it basically if we can. So you're buying for the team, not their tech. Yeah, sometimes it's both and sometimes it's blended. If it's purely tech, then we will think, okay, how long would this take us to do ourselves?

9:27How difficult is it? Try to figure that out, try to talk to them. If it's extremely difficult and this is something that they've built over a very long time, I think CrewRare is one of them. It's a legal research platform, basically, where we're like, okay, this is such a difficult problem. It would take us really long to find out. It's much better to... And they also happen to be an amazing culture of fitness. If it's both, then it's really good. But if it's either, then that's also good enough. Okay, nice. My next question, Harvey, who you are often compared to or put alongside, has launched his own model.

9:59Would you do the same? We're not looking to train our own frontier model. I don't think that really makes sense business-wise. Why is that? Because there's new best frontier models every two weeks. and I think I don't think we can compete with the likes of Anthropic and OpenAI to create level models I just don't I don't think that really makes sense for our clients for us to spend energy on there are certainly specific tasks which makes a lot of sense to post train models and so we're doing that for you know very specific things we can get a much better intelligence to cost ratio than any of the open providers because there's like a very it's very specialized but for a general sort of not very specialized, let's call it knowledge work, AI, agentic AI.

10:49It's a difficult sell for us. Okay. Yeah, I guess hugely, hugely expensive. Who do you look to as competition? I know you're compared to Harvey in terms of the end product, but who do you look to perhaps in terms of technical capabilities or maybe not competition, but who are you kind of watching and thinking they're cool, they're doing it right, let's keep abreast of them. How do you map the ecosystem in that way and who do you compare yourself to? I think there are lots of similarities to coding because it's going through coding maybe six months ahead of legal. So, you know, I like a lot of stuff that Cursor is doing.

11:22I think they've been great. Grogbot is a really cool, I mean, it's basically open call, but it's like the UX is really nice. And so we're definitely looking to all of these that are doing similar things, particularly on how to solve UX, how do we package extremely capable agents that do very complex work in a very simplified, easy to understand way. even customer support tools we're always staying very close to the ecosystem of what's happening because there's lots of parallels particularly on UX that we can learn from each other. Last question on kind of recent announcements and things like that you've announced a shift to consumption based pricing for some tiers I understand.

12:00What are your thoughts on that? Is that a move to fully as in move all of the tiers towards that and what does that feel like for you as the CTO? What does that made for you guys? So yeah, the, the idea is essentially marrying incentives in a way. So the great thing about this is if we marry incentives of our clients and us, we can optimize really well. So, um, let's say we have a new legal research product. This is not truly hypothetical, um, that can do amazing things and it runs for like three hours and it's obviously super expensive to run but it creates an immense amount of value we had actually had this problem there was like a specific thing it cost 200 to run and people loved it and so if we have consumption based pricing that's great because they can use it now we can build really expensive things that are like 10x give 10x value what they cost and we can just add that to the product because everyone just they if they don't feel like paying for that they won't but if you have a team it's very liberating absolutely plus it also means we can make great cost um reductions and it goes straight to the client.

13:04And so it's like if we had a flat fee, then we wouldn't invest into making these really crazy, extremely good intelligent things because they would be too expensive to run. So suddenly we would be in this conflict of interest where we would not launch a great feature because we were worried about our own costs that it wouldn't make sense. And so now we have this thing where our incentives are completely aligned with our clients, which is for us extremely liberating because we do all sorts of cool things and we can also do all sorts of cost optimizations and it's all in service of our customers that's cool i haven't really heard that perspective but across across about it so you think in general ctos this is something that model of pricing would be something that the technical team are going to be really they're super excited about you can do i mean i see a bunch of people doing all sorts of cool stuff now and also more of our engineers are super excited about uh optimizing for cost not for our own reasons but because this is like now a challenge yeah and it's like something that goes directly to our clients it's like part of our North Stars, we will be the most intelligent, cost-efficient place to go.

14:02Yeah, that's really cool. What else should we look out for from Lagora? I know there's lots of rumors about funding, etc., but in terms of technical progress, what should we look out for from you guys, from your department? Yeah, we're working on so many cool things. I mean, yeah, I don't know how much I can say. That's okay. It's a tricksy question. Yeah, I can say that we're looking to do more stuff proactively for lawyers. Okay, nice. And then my last question, what is going on in Stockholm, in your opinion, as in why is it having such a moment? You probably get this question lots, but I still feel like we haven't got to the bottom.

14:36What's in the water in Stockholm? I don't know. There are, yeah, there are many hypotheses. I think the two ones that are maybe, well, I don't know. So there's like multiple waves. The first one that Max likes to point out, and I'm not sure how, you know, how much, excuse me, causation there is here. But in the 90s, Sweden introduced this tax break that basically made ensure that everyone had a PC at home and then in school as well. So I think that probably bred a level of technical knowledge very quickly. And then I think there's been the wave before us of tech companies like the King and Klana and Spotify that also relocated a bunch of people.

15:20so they brought a lot of technical talent and that means there's now really high sort of talent density, technical talent density in Stockholm and now we have an X-Wave like us and Lovable and Tent of Health and other ones and I think that's sort of, now we're starting to have like a self-sustaining more startups starting to come out of even Lovable and even us When you're looking for engineering talent, are you taking people from Stockholm or is that now very competitive? Are you bringing people in? Where do you look for? Both. I think we've spoken to most people that are in Stockholm at this point.

15:54So we're looking to bring people in. I think that's the great thing about the EU. We can actually move people from all across. We can move people from Amsterdam, from the UK. We can move them from Norway a lot. We've had lots of Danes. So I think there's lots of strength outside of Stockholm, just in the Nordics, for sure, but also in the rest of Europe. Yeah, yeah. You're Danish, as I understand. Do you hope that Copenhagen can experience a similar moment Why isn't it? Or perhaps it is. And it's just about... That's a great question. I really hope so. I think there was a moment Copenhagen hasn't managed to really get that ecosystem going.

16:31I think probably the only one I can think of is like Playo did well some years ago and nothing really came out of that too much, too successful. So I think that's a shame. I'm obviously now trying to bring all the best engineers from Copenhagen to Stockholm. So I'm not helping. But I think in general, the Nordics, I think it's great even for Copenhagen that's going to well in Stockholm and likewise for Norway and Finland. So I think there's a revival in the Nordics that I think is great. And there's so much really strong talent here. Yeah, yeah. It's exciting to see all we've got time for. But thank you so much for joining the pod.

17:04And at short notice, because we only asked you yesterday. So it's a quick turnaround and we really appreciate it. Listeners, if you're interested in following Lagora, it's a company that we track step by step. So you should sign up to our daily newsletter. We'll drop a link in the episode description. and you'll find all the Lagora updates there. You can also catch Lagora CEO and co-founder Max at our Sifted Summit in London on September 31st and October 1st. We'll put a link to the tickets in the episode description as well. That's going to be great. So please come along. And as always, please rate, review and share the pods.

17:40This episode was produced by Aditi Meshwa.