In episode 50, I interviewed Harvey founder Winston Weinberg, but I specifically waited until episode 52 to interview Legora CEO Max Junestrand. Hey there, I'm Zach Abramowitz and I am Legally Disrupted. This is episode 52 of Zach Abramowitz's Legally Disrupted. But first, here's my hot, disruptive, legal take.
Now, in 2023, shortly after the launch of ChatGPT, all anyone wanted to talk about wasn't Harvey or Legora, it was actually Counsel. The 10-year-old YC startup originally called CaseText that had gotten its hands on LLMs before just about anybody else and they had figured out how to wire these models into legal workflows. It's hard to believe it now, but the conversation around Thomson Reuters' acquisition was, "How could they possibly justify paying $650 million for this company?" Their institutional investors were asking that question the very next day after they announced the deal. But by the end of 2023, that price didn't sound nearly as crazy anymore because Harvey had reached a valuation of more than 700 million by its Series B offering led by Kleiner Perkins and Elad Gil. And for most of 2024, the question I began to get was, "Harvey or Counsel?"
Meanwhile, in early 2024, Y Combinator funded another legal AI startup called Leia, which would eventually rebrand as Legora and position itself as Harvey's primary competitor. That positioning has paid off. Counsel began to fade from the center of the conversation and for many firms whose attorneys were asking, "Hey, why don't we have Harvey?" But the firm maybe had a bad experience with Harvey's sales team or they had felt left out of the somewhat stealthy go-to-market, Legora now emerged as a compelling alternative. While Legora has been playing from behind, its catch-up game has been remarkable. By the beginning of 2026, the legal AI ecosystem was no longer all about Harvey. The conversation had definitely shifted to Harvey versus Legora.
So, why did I then specifically separate these two interviews? And in fact, in episode 51, we had Will Chen from my co-host who has been one of the most vocal outspoken critics of Harvey and Legora. Well, I think a bit too much has been made of Harvey versus Legora. Earlier this year as an example, Harry Stebbings featured Winston and Max in back-to-back episodes of 20 VC, really amplifying the Harvey v. Legora Coca-Cola versus Pepsi narrative. And at a high level, I get it, but Legally Disrupted isn't only about covering the high level. Harry is covering Silicon Valley and startups in general. I'm covering the impact of AI on the legal ecosystem. And from my perspective, Harvey and Legora are two different companies led by founders with different visions. And while many people tend to see them as essentially the same company, the same product, I think these differences are going to become more apparent over time.
In both episode 50 and 52, I went deep with each founder to give you the viewer a better understanding of these companies on their own terms, not just in comparison to one another. Because ultimately, if you as a lawyer or a law firm decision-maker are choosing one of these companies as your transformation partner, you should be diligencing the founder's vision and their ability to execute just as much as the product itself. So, without further ado, let's go deep with Max. Let's learn more about his vision for Legora and the legal profession. Let's get disrupted.
So, Max, I feel like part of Legora's magic has been your ability to meet users and lawyers where they are. So, we are recording this on May the 29th, and it seems like where the entire legal industry is today is obsessing about this press release that came out yesterday from Kirkland and Ellis talking about how they're building their own new solution. And certainly all the lawyers on X and LinkedIn and social are having a whole time with this. I know it's like fresh. I also know that you're pretty connected over even at Kirkland. So, despite the fact that you might not have been mentioned in this press release, what's your reaction to the news? And how do you think what's the important context?
I think that the really large clients and enterprises are starting, you know, we're a couple of years into the AI journey now, and they're starting to ask quite fundamental questions around how and where are these firms really applying and not just augmenting the way that they used to do things, but actually reinventing parts of their services. And different firms are on different trajectories of that journey, right? Like you have the sort of the bell curve, and I think if you're a Kirkland and you're number one by a lot of metrics, you really need to show a lot of strength and a lot of power to those clients to say we're taking this seriously. Now, I don't think the entire equation is, you know, spending hundreds of millions of dollars on building your own thing to show that you're serious, but I think it's a good part of the equation. And they're using a muscle that not every firm has, right? And so, while a million is a rounding error for Kirkland, right? I mean, this could be a total screw up and it wouldn't impact them at all. And so they can go out and market an orchard while everybody else is marketing apples. And that's very appealing. And so I think they're doing a really smart move of leveraging part of their weight in a way that nobody else can. And we'll see what comes out of it. It's certainly a good headline. But at this point a couple of years into the legal AI sphere that we're in, I think press releases are one thing, results are another, and I'm very excited to see the results.
One of the things that's interesting about your company and your product. I mean forget the crazy growth which we'll get to is you were initially designed for law firms. I know you're also selling into corporate legal and that's certainly like an area of emphasis right now and you guys are showing up at all the major legal conferences for in-house. Legal and corporate legal and in-house legal teams are so drastically different in many ways than law firms. And very often what I'll hear from people who are in-house is when I came here from the law firm this is so different and in many ways I wasn't necessarily prepared. How do you design a product that has to be number one for law firms and number one for legal departments?
Yeah. That's a great question. And I think zooming out a little bit the litigation department of a really large corporate is actually more similar to the litigation department or practice in a law firm than it is between the M&A practice at a firm and the litigation practice at a firm. So already from the beginning because we went quite horizontal and wide within our initial target group, we had to serve M&A, IP, bank and finance, litigation, and in-house use cases. And you know, I think what ends up happening is every time the models improve, they improve at almost every single task. And so for the first couple of years now, we've been in the business of building the tools, the scaffolding, and the context layers required to put these models to work. And then they sort of a little bit out of the box can handle the differences between the practice areas. And I think the main challenge that sort of separates the two is most law firms don't really work with volume the same way an in-house team does, right? If you're a very large in-house team in let's say insurance, and you handle a lot of claims, like that is structurally a different business than Kirkland's M&A business, right?
Right, it's very different. And what we've had to do is we've had to do like the difference between shooting free throws. So, you could if the only skill is, "Hey, we just need to shoot free throws." So, as many reps as you can get, it's the same thing kind of over and over again, as opposed to like, "Hey, we now have to strategize for an 82-game NBA season." Right?
Exactly. And that's exactly right.
And I do think like the complexity of the work is simpler in handling mass claims versus, you know, the never-seen-before transaction or, you know, take a company public or bet-the-company litigation, right? And so, I think naturally there's been the most applicability and sort of immediate return on investment in the lower parts of the value chain. But it's increasing, right? And everybody can see the writing on the wall that the complexity of the tasks that AI can handle is becoming higher and higher every single day. And so, if you extrapolate that, then you understand that even if you're a firm like Kirkland, you know, you have to start your AI journey 3 years ago, and you have to build up to that world. I think what we're juggling is the different incentives, right? So, the incentives in an in-house team is, you know, let's cut down costs, let's cut down time, let's bring as much work as we can in-house, and let's be as effective as possible. Now, many of the firms have realized that this is happening, right? And they are actively transforming their service model and their delivery model to meet this new world where, frankly, agents can perform tasks. And we're a little bit of the opinion that if an agent can perform a task, it will perform that task. The only question is, where does it get transacted, or where does that work get carried out? And to Legora, it doesn't really matter. It can happen in the firm, it can happen in the in-house team, it can happen on the software. The important thing is that it gets carried out, and that software is used to do it. And so, I think we'll continue to see the world twirling going forward, but it really comes down to the incentives that the in-house teams set out for their firms.
So, there's so much to unpack here, but let's start with the point about your belief that the task needs to get done, it's not specifically important who the party is. One of the things that when I've spoken with clients, we show them sort of different levels of transformative impact. And one of the ways that we measured that, one of the questions that we would ask and sort of score this is does it change who owns the process? Right? Meaning if this was done before by a law firm and now it's done by a law firm but done maybe more efficiently, yeah, that's impactful but it's much more impactful if okay, this used to be something that was done by the law firm but now because AI or because agents, what have you, it's now done by an in-house legal team or it's done by a provider in Manila or whoever it is, that's already much more impactful. Are you seeing that already?
Yes, and but I think a third option or maybe a fourth option there is, you know, does the law firm enable that new service in a very different, AI-forward way where actually the human involvement is very low, right? Because there is an option where you say, okay, a law firm whose service, you know, 15 large banks as clients, they have economies of scale in solving, let's say, different compliance use cases at scale because they can bring that use case to 15 clients. Whereas innovating within the bank, you know, they're just solving it for themselves. And so I actually think that many of the law firms have an opportunity now to scale or productize, really, their offerings and then take them to the market at scale. And the real question there becomes kind of a market share game, right? How quickly could you as a service provider take something to market, get scale on it, get many clients on it, and then start to iterate? Because when you productize things, you need to have a quick iteration cycle or you're iterating with the clients. I think that's been one of the sources of strength for Legora. But that will then allow them to build a bigger and bigger moat around that service.
One of the other things that you said before was talking about the journey and how law firms and really anyone for that matter start in one place, but you're somewhere else 3 years later. So I've discussed this many times and said, "Listen, if it's just about turning on a switch, then you could theoretically be late to the game." But the compounding impact of early adoption is the AI brain is beginning to sort of instill this AI mindset such that you know better where and how to use the products and you're also able to sort of now forecast where things are going and you're able to start in some ways changing or maturing how you work with these tools. So my question for you is this obviously presents a big challenge for you because on the one hand you as a company are farther ahead both because you started earlier than most law firms, you know, thinking about AI seriously, but also because I imagine that you've got deep ties with all the AI labs. Your team you've got you're working with top engineers. So you guys are like, you know, probably 6, 12 or maybe even 18 months ahead in terms of what you know, but many of your lawyers your customers are in let's say maybe where you were in late 2024 in terms of how they're using the tools. How do you do that? How do you build for the customer where they are today, but also knowing where things are going from here?
Well, I think one thing that I've been positively surprised by is that I actually think many of our clients are much further ahead than what you may give them credit for.
I actually think so. And I think it's because the light bulb moment has happened for enough influential companies and partners frankly, where it's like, "Wow, okay. It's here. It's transformational. We need to get going." And part of our job is to help them get as far ahead as possible, right? I like to say that every firm that's on Legora, every company that's on Legora should be in the top quartile of AI users, adopters, and sort of appliers in the way that they change their processes. And what we try to do is we of course talk to all the labs. We drill that down into what will this mean for Legora? And what will this mean for the legal industry? And then we package that and we give that to our firms and to our companies. So, you know, I run a lot of road map sessions where truthfully, I have no idea what the world looks like 18 months from now. But thank you for giving me credit for maybe thinking that far, but I think we have a very good idea sort of within 6 to 12. And frankly, it's also been very compressed because of what happened over Christmas with Opus and Opus 4.5, 4.6, and now GPT 5.5. And the increase in capability of those models was so large, it was almost like moving from a pre-ChatGPT to a post-ChatGPT world.
Where just so many of the fundamental problems around tool calling, sequencing, long horizon tasks, were solved with these newer generation of models that it accelerated what was already moving fast. And I think the really difficult part as anyone in a company, you know, responsible for change management or responsible for innovating in the legal team is it's happening a hundred times faster than any previous technology wave. Right. And nobody is prepared. Everyone is unprepared. And you know, you wake up every week and you're like, "Oh my god, like there's just a flurry of new things and I haven't even gotten started on the thing that happened three months ago." It's just, you know, I think people thought we were going to work less because of AI. I'm certainly not feeling that. I don't think anyone of our clients is feeling that. We're all working n times harder, and you also sometimes feel like you're building sandcastles because you just know that something new is going to come three months from now and it's going to crash over your sandcastle and I have to start over.
I say this all the time to lawyers. I'm like, "Listen, I've got good news and bad news for you. Good news is AI is not going to take your job, but the bad news is AI is not going to take your job." Right? You might be busier. I asked a room of lawyers this a couple of weeks ago at a partner retreat. I said, "Raise your hand if you feel less busy since AI emerged." No hands, right? This is it. It's very hard to predict. By the way, I like to flex on certain predictions that I've made and been right about, but I'm also like, you know, I was deep into ChatGPT, so deep in that at some point I decided late last year, you know what? I don't really think I need my Anthropic account. I'm just not using Claude at all, right? So like this and I feel like I've been directionally accurate on a lot of things when it comes to AI and large language models and even I, you know, had to go through that period of feeling deeply behind and then kind of eventually caving, going back, you know, with my tail between my legs to Claude and then being like, "Oh my gosh, this is amazing." And now, you know, shortly after ChatGPT is releasing many of these features and moving in that direction as well. Since we're talking about the frontier models, I have to ask. I'm sure like you're getting these questions all the time. Are law firms, legal teams, are they asking you, "Hey, why shouldn't we just be building all of this in Claude or maybe the question will turn into Grok at some point." How are you helping them like think through that and justify working with Legora as opposed to working, you know, directly with one of the models?
Yeah. Well, so we certainly got the question, right? And I think that's part of the credit of the very noisy marketing machine that Anthropic is. And whenever they do a press release, you know, the world goes completely crazy and sells off 20% of the public legal tech stock. And it was actually funny because like a bit of inside baseball, but we were raising our series D as that announcement came out. And then every investor that we were talking with was like, "Oh my god, like tell us about Claude for legal." And we just turned around the computer and we showed them the three markdown files that could I think it was like triage an NDA, like, you know, generate a playbook and it would like give out this...
An NDA, really? Like, wow.
Wow. And we were like, you know, like it gives us this table with like a green, amber, red type of thing. And then you like, you know, copy-paste that over to Word and then you're like, okay, please implement this, blah blah blah. So, so here's what I think of it. I think it's very good that the world is like waking up to the fact that these foundation models are very applicable in law. You know, even if we're out there like screaming it from the rooftops, it's very helpful when a big microphone like Anthropic comes and tells the world about this. I think engineering has had its moment. Now legal is having its moment, right? The foundational models themselves and the harness of, let's say, Claude CoWork, however, is quite thin and quite shallow. And for any serious worker, if you try to put in, you know, an entire data room into Claude CoWork and you go, okay, read through every document here. Here's the guidelines. Here's the report I want you to fill in. It's going to fail. Now, why does it fail? It fails because the harness, which is effectively the environment that you put the model in and the tools and the way that you allow it to reason, is built for everyone. So, at the end of the day, it's like built for general office work and built for no one. And we have the luxury of obsessing over building one type of harness, our type of tools, our type of integrations, our type of governance and security checks so that no documents get shared where they shouldn't so that you have an audit trail of everything. And we get to obsess about that for one type of user. And that allows us to go much, much, much, much deeper than what the foundational models can because they have to play for every type of use case productively. And so, I would really encourage everyone to try Claude and try GPT and try Gemini and because the world is moving so fast, you need to try all of these technologies every quarter or every other month.
I'll tell you an anecdote from one of your competitors from wordsmith.ai and they're really focused, you know, on in-house legal, but their CEO told me the story that he says when companies will come to him and ask about using his product, he'll say, "Well, do you have any internal chatbots?" And if their answer is no, he says, "Go get one of those first because if you don't, you're not going to wrap your head really around AI to begin with. And our product's going to be too sophisticated for you. And on the other hand, you're constantly going to wonder should we just have gone with one of the base products?" He said, "I want you to use it, wrap your head around AI, and also figure out where this product comes up short, and then come back to me and we'll talk."
Yeah, and I think that's exactly the right way to think about it. And as you start to really push at the frontier like we opened the conversation about Kirkland. If you're a company that's going to push the frontier of what's possible, you know, that's what we want to bring. And today if flavor of the month is Claude, maybe flavor of the month is OpenAI next month. And so, it's important for us as well to have the relationship with all the foundation models and the labs to make sure that we can serve the latest and greatest through the product. And ultimately, I don't think that we are in the business of just selling a tool and saying, "Good luck, have fun." We are in the business of making every Legal AI client as successful as possible. Now, that means that software is a little bit like a garden and every time the ground shakes we have to replant the trees and we have to go cut the grass and we have to reimagine what our software will be. I think Zach, two years from now you will have sub 10% of our existing product still alive. I think 90% of everything we built so far I don't think will exist in Legora two years from now.
So this is really interesting I don't know if you've ever seen there's a fantastic South Park episode where the kids come up with this idea this is back in like the day of all the crowd sourcing the crowdfunding companies and they said they were trying to build a crowdfunding company but the business model was we're going to sit on our ass do nothing and make a lot of money and this is the entire time they're talking about this is like their business model and I think there is sometimes this impression with tech companies that oh you just build it and then everyone uses it and the company sits around does nothing and makes money this could not be farther from the truth in an age of AI in particular where it seems like one of the things that customers pay for meaning if I'm a law firm and I'm going with Legora the reason I'm going to do it is I'm going to say listen it's great that we have this product right now but there's going to be a massive change and I would rather outsource to a team of top engineers who have ties with the labs that to me is going to be the argument they're going to because they're going to do that work every single time there's a massive change they're going to come in. Does that feel and I got you I'm sure you're talking with other entrepreneurs, other VCs. Does that feel like that's different in an age of AI much more so than it was like if you were building Salesforce?
Much, much. It's very different. And I think it's different because of a few things. One is the underlying rails upon which we are building are constantly being reshaped. And we're still learning so much about applying the technology in practice, right? And when you really see, you know, a private equity firm leaning in on what AI can mean for their companies and for their business, this is all new territory. Like nobody has really done that. And then you learn things in the field and you go, "Wow, okay, this thing doesn't really work." So let's take that back in a really quick feedback cycle, build it, and get it out again. And I actually think a company's worth in today's age is more tied to the velocity of their iteration cycle than it has ever been in the past because ultimately, as you say, you're looking at a snapshot of today and the technology and the companies today. But you should probably pick a company and a partner to work with that has the fastest slope on their curve going forward. And pace of innovation. Pace of innovation, right? And not to toot our own horn, but I think that was...
Like one of the big reasons why Legora came up from a, you know, very small legal market initially. Like we started in Sweden. Sweden is a smaller legal market than Kirkland.
And now we scale like enormously and yeah, I'm spending a lot of my time building the organization now and thinking about how do we build a company that's going to be able to at scale serve our customers but maintain the velocity and the culture of enterprise innovation. I think that is...
How has your day-to-day changed since you started the company? Like what would you be doing on a typical Friday in 2024 or 2025 versus what you're doing today?
Well, so towards the end of, well, summer of '23, I was still coding. So, there's a few commits from me still in the repo. Towards the end of '23, I was doing 15-minute demos and I was like back-to-back to back from 9:00 a.m. to like 9:00 p.m. every single day with demos because over the first summer after ChatGPT, I think a lot of people played with it and could see like, oh, it would help me write a speech. It would give me a really good recipe. And people got to see it during their vacations and then they came back in like September or like, okay, we're going to see about how we can apply this technology in our company. And then, you know, they would go on the ChatGPT website. They would see that there's no European processing. None of the data is confidential. Like, okay, nobody can apply this. And then, we were there to pick up a lot of that work. So, I've moved from doing a majority of I'd say customer work to maybe, you know, splitting my time between recruiting and like org building, product, and working with our top customers, on, you know, what's next, effectively.
And I think the strength of a company like us now is that I'm privileged to recruit and get leaders who have seen scale before, but are really hungry for the pace that we have. So, David, our CFO, who just joined from Vanta, is a great example. He built that company from 200 people to 1,300 people, 300 million in ARR, and now he's coming here. He's like, 'Wow, like what we're doing in a quarter, you know, you maybe you just take a year in the old SaaS world.' And Sinead is a great example. She just joined us from Atlassian, where she was the CMO.
And you know, she used to have a 500-person team at Atlassian. This is like, 'Honestly, the entire Legora org.' And now the marketing team at Legora is 25 people.
So, our mutual friend, our mutual friend John Levy, from Y Combinator, who's one of my, he and his wife are two of my favorite people in our space, and they're so humble, and they've picked hand-selected some of the top legal AI companies, you know, in history at this point, meaning over the last, you know, really 20 years at this point, and Legora included. He has a line that I just keep coming back to over and over again. As I asked him on stage, actually, this time last year, I said, 'You know, I see that you're investing even more in legal. Is that because you care about the vertical? Is this because you see LLMs as a great fit?' And he said, 'No.' He said, 'We don't care about verticals. The only thing we care about is talent. We are like, look at, we look at this as like the NBA draft, and we want to get the absolute top talented entrepreneurs. It just happens that the most talented people right now want to build companies in the legal space and I think you can see that not just with you. I think I saw Scott Stevenson brought someone to Spellbook, very impressive tech exec in the Spellbook right now and you know, you mentioned before that you know, coding AI coding had its moment and now AI for legal is really having its moment. It seems like people want to get into this space from, you know, who've had really illustrious careers because they see the opportunity. So, let me ask you a question. Do you think that Legora is a trillion-dollar opportunity?'
Yes. Absolutely. I mean, we wake up, you know, it's fun cuz you're so in the arena and every day you wake up and you think about, you know, what can I do today to move forward on our mission? And at the same time, you need to zoom out and go, 'Okay, are we on the right directional path to that opportunity?'
I feel like not enough people appreciate that you think this and my sense is that I think Harvey and the folks over there probably think the same thing and I think that there's this impression maybe that like you guys are trying to just like, you know, smoke and mirrors it until someone buys you for, you know, some amount of billions of dollars and I've been saying recently I'm like, 'No, I think that they believe it's a trillion-dollar opportunity.' So, it's interesting to hear you say that.
It's also, you know, it's like I've set up the company in a way so that like this is my life's work and I decided that after we were invited to the, also inside baseball, we got invited to the most prestigious alumni event of YC 4 months after doing YC, which is very strange. We were the only company from our batch to get invited. And Brian Chesky from Airbnb was the main keynote and he went up and he was going to talk about the founding story of Airbnb, but then he went like, 'You know what? Like I have said this story a hundred thousand times. I'm just going to like rant about what I've done in the last like 4 years following COVID.' And he described, you know, the feeling at Airbnb when they went from, you know, top of the world to COVID comes, they lose 80% of the market cap. And he had, he was working with his exec team, but the exec team didn't really care about the company, they cared more about protecting their own reputation. And he just, you know, described this thing of, 'All right, like I have to decide now. Is this my life's work? Am I going to like fight so hard one more time to like get this company out of the hole that it's in?' And he, you know, described what Paul Graham would later write an essay on, which is like the founder mode thing. And I think like there's good parts and there's bad parts of that. I think if you have a really well-functioning exec team, maybe you don't have to go founder mode on everything and you have other people who can go founder mode on stuff on your behalf. But I came back from that alumni event going, 'Huh. If he decided Airbnb is his life's work, I'm going to decide like what is my life's work.' And that makes you think very differently. It makes you, you know, every fundraise we've done I have fought really hard to maintain board control, which we still have in the founding team. And even though we've raised close to a billion dollars at this point, and it's like I'm not going to let, you know, VC or private equity incentives guide this company because I think I have a much better view of how it gets to a trillion than anyone else. But I think that's also just a number. Like I care about the impact that we have in the world, and then you can get a valuation as a function of that.
Right. No, this is like Warren Buffett has the same advice when you pick stocks, think about buying the stock and asking yourself, will this be the last stock that I ever purchase? And one that I'm going to hold really for the rest of my life. And it does, it just, it completely changes your approach to that.
Right now, you mentioned private equity before. I was joking on X the other day. I said, AI for M&A due diligence. Boring AI use case or the most boring AI use case? Now I know that this is one of the main use cases that everyone's focused on. But I look around and like, you know, M&A deals are still pretty pricey and you still hire lawyers for them. And in many cases because of reps and warranties and insurance, the diligence doesn't have the kind of importance that it might have had even, you know, 10, 15 years ago when I was practicing.
It will save you a lot of way.
Yeah. So, but it feels like we kind of got like very focused on here. And by the way, I remember one of your team showing me a demo of the AI of the M&A due diligence product and I was like, wow, this is absolutely amazing. But is that like what is the impact right now and from your perspective of AI tools on M&A deals? And how are you seeing the best firms adapt? Is there a repricing mechanism? Is there a rethinking of the work? How are people beginning to grapple with that?
