Max Junestrand, co-founder and chief executive of Legora, tells a Y Combinator Startup School audience how three Swedish engineers with no lawyer among them built a legal AI company that Y Combinator turned down the first time it applied. Gustaf Alströmer runs the Q&A afterward, pushing him on whether domain expertise still matters, which models Legora runs and why, how a founder learns to hire an executive team, and how the company landed Jude Law.
Guest: Max Junestrand, co-founder and chief executive of Legora, the Stockholm legal AI company he says over 3% of the world's lawyers now use
Host: Gustaf Alströmer, a Y Combinator partner, who ran the Q&A after the talk
Published: 29 August 2026 on the Y Combinator feed
Watch on YouTube | Apple Podcasts | 1 hr
✅ Time saved: 31 min
Key Takeaways
Domain expertise is not the moat; a willingness to learn the market is
A Swedish VC passed on the pre-seed because the team had no lawyers, and Mr. Junestrand says the partner he ran into at an airport was "very, very bitter"
He would still hire domain expertise for quantum computing or fusion
Legora bet that the models would improve rather than fine-tuning one of its own
"The models will keep improving."
The first pitch was that it was ChatGPT, compliant in Europe
The first YC interview collapsed on one question about what kind of lawyers they served
Two months later they were accepted under a new name and a different platform
They froze the sales motion for six months, with money in the bank and nothing to sell
"When you work with lawyers, you only really get one chance to get it right."
Hire for the slope of the skill curve, not the y-intercept
"Our top seller today is 23 years old and he has sold more than $10 million worth of Legora."
Evaluation, not the model, is the muscle worth building
Legora hired lawyers and made half their job building the use cases the evals run on
Legora's own benchmark went public last week, and xAI's Grok was the surprise
The models are not on its data processing agreement yet, so customers cannot use them
Banks and big law firms have one hard model preference: "They don't want Chinese models."
The Swedish law of Jante is the thing a Swedish founder has to beat
The values are lean in, fight for excellence, grow together, and the acronym is deliberate
Agents are moving from doing what they are told to starting work on a trigger
"how does one lawyer produce the outcomes of a team of 10 lawyers with the help of Legora?"
The CEO job resets every quarter, and the SaaS growth playbook does not transfer
"If you grow 40% year-on-year, you are doing fantastic." Mr. Alströmer: "That's not enough anymore."
Jude Law said yes because Legora would not stop asking
"I think one underappreciated form of negotiation is nagging."
What Legora Sells, and the Slack Channel Where the Customers Say So
Mr. Junestrand opened by noting that the reception was warmer than the one Legora got the first time it came to Y Combinator, and that the company's YC story started with a rejection. He then gave the audience the two-sentence description YC teaches founders to write.
"Legora is the agentic operating system for lawyers, and it handles complex legal work from start to finish so lawyers can achieve more than ever before."
"Over 3% of all the world's lawyers are now active users of Legora."
The opening was the state of the software, not the state of the models
"The software that we were up against looked like it had been built in the '90s."
Software engineers have been spoiled by good tools, he said, because they build for each other and for themselves. Lawyers had nobody doing that for them
His favorite internal Slack channel is called customer love, where customers post things they feel they could not have done without the system, and the company mines its sales calls for the same kind of line
"Never has the path between an idea and execution been so short."
"If Legora went away tomorrow, I'd just go back to coaching high school basketball."
One customer told him at a golf tournament that the product was the one-word answer to how he juggles work and travel
The revenue curve runs from general availability to nine figures in under two years
"from the time that we went into GA in October 2024 until the last end of quarter, we've grown from one to a hundred million in ARR"
"Starting from just three engineers in Sweden to a company of over 750 people all over the world"
He said doing it from Europe adds another feather in the hat
When Legora arrived at YC it was questioned about selling into law at all, a conservative business that had been slow to take up software of any kind
The Company Started Without Him, Before GPT, With Google's BERT
Legora did not begin with Mr. Junestrand, and it did not begin with him in the room. In 2020, before GPT and before AI was fashionable, a lawyer, a physicist, an engineer and a psychologist started a company on one observation: law students went into their internships mainly to summarize court cases. They used BERT, an early Google model, to test what could be built.
Mr. Junestrand met his two co-founders, August and Sigge, at a volleyball game in the Swedish archipelago, which he conceded was not a high-odds place for a startup story to start
They showed him a demo built on GPT-3.5 that could explain what a stock option agreement really meant
He told them it was a cool project. One of them replied that he was glad Mr. Junestrand wanted to help out, and he ended up helping considerably more than that
He dropped out of college and never finished his master's thesis because "the opportunity cost of not building had become too large"
"When GPT-3.5 came, that was the internet moment of our generation."
They Offered to Pay Lawyers Their Hourly Rate for Lunch
To find out how lawyers actually worked, the founders cold-emailed addresses off law firm websites and messaged lawyers on LinkedIn, asking for a lunch about their practice areas and offering to pay the hourly fee for the time. Many took the offer. Most did not make them pay, and many paid for the lunch.
Once they had their feet under themselves they approached Mannheimer Swartling, one of the largest law firms in the Nordics
Two years earlier its managing partner had gone on television and said AI was "really more artificial than intelligence"
Changing that perception, Mr. Junestrand said, took real work
The economics of legal work explain the caution
"In law, you are not paid when things go right. You are punished when things go wrong."
Earlier AI tools in the space needed an immense amount of training, with a pile of use cases and examples supplied for very little result. GPT-3.5 changed that story
Legora ended up moving into Mannheimer Swartling's offices and working closely with the firm
The First YC Interview Ended When a Partner Started Laughing
They applied in May 2023 with a promise to let users query all legal documents using large language models
The first question in the interview was what type of lawyers they were serving
Their answer, he said, was "What do you mean? Are there different types of lawyers?"
"Tom Blomfield started laughing in the interview when we gave that response, and we knew that we were cooked"
The rejection is the part he tells founders about. They did not work the day after, went back to building the day after that, and were accepted two months later under a new name and with a different platform
From Sweden, he said, the acceptance felt like "we were going to Mount Olympus. We were going to dine with the gods"
Once inside, they realized they knew a lot more than they had thought
His favorite photograph from the whole story is the screenshot he took after the video call following the second interview
A Windowless Conference Room and an Engineer Fanning the Door
After the acceptance, the team moved into Mannheimer Swartling, the biggest law firm in the Nordics and, in his telling, one with no real idea how to work with software.
They were given a conference room with no real windows where the air conditioning shut off at about 5:00 PM
It was close to the Coke fridge, which he counted as the compensation
At about 6:00 PM one of the engineers would get up, grab the door and wave it back and forth to push more oxygen into the room
The chairs, he said, still left an imprint on their backs three years later
A photograph of him in a car with two extra screens marks the move out of the law firm and into Legora's first real office
The lesson he drew was about asking, not building
They learned quickly not to take no for an answer
"it is very powerful to be a person and to be a company that others want to see succeed and that others want to help"
Sales Calls From 1 A.M. to 10 A.M. Under an Influencer Ring Light
The team came to San Francisco, took a room in an Airbnb up in Bernal Heights and built
"During YC, we grew from zero to a million dollars in ARR."
Working across San Francisco and Europe took a heavy toll on the team. He bought an influencer ring light and put it on his laptop
"I was doing sales calls between 1:00 AM and 10:00 AM every single day in San Francisco to Europe because that was where our initial customer base was"
He called the batch a rocket trajectory at the time, and noted that two years on it looks slow next to what some companies are now doing
Eighty Investor Meetings, and a Term Sheet Negotiated on Decimals
It was his first fundraise and he had no idea what to expect: "we had lined up 80 investor meetings in a week and a half", European and American
The takeaway was that fundraising is a storytelling exercise
"you need to become a really good storyteller in order to drive interest and to get people on your team"
He found the two investor pools different in temperament: "in the US, we think more about what can go right rather than what could possibly go wrong"
Legora took money from Benchmark, and he negotiated the term sheet on the partner Chetan's desk with a computer open, going on decimals in Excel to settle the exact number of shares
"They ended up investing $9.51 million."
Redpoint pre-led the Series A three weeks later
The Month Interest on the Bank Balance Beat the Revenue
"There was actually a month following that when we made more money from interest rates on those $35 million than we did from customers."
The company was 10 people with about $35 million in the bank
"That is not a very good sign when you have become a bank."
Going into the first board meeting, with Benchmark and Redpoint in the room alongside the three founders, they made what he called a very hard decision and froze the sales motion
"When you work with lawyers, you only really get one chance to get it right."
A product that does not work, lag that is too high, or a system that goes down under too much traffic on the Azure instance, and "you are toast"
"After that six-month sales freeze was when we really started ripping."
That, he said, is where the curve from one million to a hundred million begins, and where the company stopped looking for product-market fit and decided what its product was going to be
They Used to Pick the Roadmap by Voting
In the earliest days, before the rename, the team decided what to build by taking democratic votes across the whole company.
"if you have too many chefs in the kitchen, that typically does not make a very good dish"
The result was too many features at once, and everything had to be rebuilt for the launch that followed the sales freeze
The rebuild targeted a platform that could absorb constant change underneath it, both in the underlying models and in the frameworks used to build agentic workflows, LangChain among them
In October 2024 the team wrote a product manifesto, gathering everything it had learned into one simple document shared with the 25-person company to rally around
Legora was doing about $1.3 million in annual recurring revenue at the time
"Many of our competitors were doing 10 times that with products that only had one of these features."
Starting in Sweden set the ceiling that forced the move. He said the market is not big enough to house a real unicorn or a decacorn, and one of the earliest questions YC asked was when Legora would move to the United States
The product refocus is what made the jump possible, letting Legora compete on product and then compete for the biggest logos in its markets
Hiring for the Slope of the Curve, Not the Logo on the Resume
"As engineers and as tech people, we often overemphasize the reliance on product and on the tech." Building a product and building a company, he said, are very different jobs, and Legora made plenty of early mistakes on the second one.
The hiring pattern they had to unlearn was the fancy logo on a resume, which the company calls y-intercept
A skill curve can start very high, but without a good trajectory upward the person struggles in a company scaling exponentially
They looked instead for people like themselves who wanted to work insanely hard and had very high growth potential, and doubled down on that talent repeatedly
"Our top seller today is 23 years old and he has sold more than $10 million worth of Legora."
He had no sales background and jumped out of university
Up to about 500 people, Mr. Junestrand interviewed every candidate outside engineering; he now interviews directors and above
The cultural foundation, he said, now self-selects
LFG Is a Company Value, and Jantelagen Is the Thing to Overcome
Legora has three values: "lean in, fight for excellence, and grow together"
"Together they make LFG—let's fucking go."
He said the profanity does its own work, telling leaders arriving from more established backgrounds what kind of company they are joining
Starting a company in Scandinavia meant working against Jantelagen, the law of Jante
"you should not think that you are somebody or that you are more important than anybody else"
Part of that humility is good, he said: the best ideas win and people early in their careers feel free to speak up
But it is hard to build one of the fastest-growing enterprise companies in the world while believing you are nobody in particular
The fix was "a nice cocktail of US, European, and Asian culture all in one"
"For Us, There Is Only Winning"
Mr. Junestrand said a strong culture is the best guarantee of attracting and keeping the best people, and pointed to a piece about the company in Sifted as an example of how that culture reads from outside.
"For us, there is only winning. Everything else is losing in a winner-takes-all market or winner-takes-most market."
"The company has to run like there's no place for number two."
He said Legora had no reason to exist when it was founded, being one of many legal AI companies, and that never feeling safe is what has kept it moving despite extraordinary numbers
"we continue acting like professional swimmers who are looking down in our lane while the competition is looking sideways at what we are doing"
That intensity had to be built in a country where most people want to go for fika at 3:00 PM, meaning a cinnamon bun and a coffee, or go home at five
A Swedish saying became an internal word by accident
"you can wake up with a taste of blood because you're so excited about something"
He used the phrase in a Swedish interview without knowing the article would run in English. The replies asked whether Legora was made up of vampires and suggested the staff floss better
Every employee anywhere in the world does onboarding in Stockholm, which he said is nicer in summer than winter and gives every Legora office the same feel
A large American law firm working with a competitor could not get documents drafted from incoming term sheets, a problem the competitor had promised to solve and had not
One of Legora's US-based legal engineers learned about it, flew to Stockholm, spent a week with the engineers, solved it and delivered it back to the customer
"I think that Swedish base has become one of Legora's real superpowers."
What Mattered Happened in a Windowless Room at 2 A.M.
Closing the talk, Mr. Junestrand said he was sharing the stage that weekend with people who had built at a scale he has not reached, and would leave the big lessons to them.
Three years in, the things that mattered were not the milestones
Not signing a big client, not closing a round of financing, not opening a big office in Manhattan
It happened in a windowless room without much oxygen, on a call on the way to the airport, and in a Slack channel at 2:00 AM fixing a bug before a Monday presentation
"Nobody puts that in a pitch deck, but that is what makes the company."
"You have to love the hustle."
He was born in the Swedish archipelago in a class of 10 other children and had a plan to go to McKinsey after school
"But when GPT-3.5 dropped in our laps, we truly made that into our thing and we captured what I believe to be our generation's internet moment."
"So you can just do things."
The VC Who Passed Because the Team Had No Lawyers
Opening the Q&A, Mr. Alströmer put the question he says founders ask him about Legora: three founders, none of them lawyers, built what he called the best legal AI company in the world, so does domain expertise not matter that much anymore?
A Swedish VC was very bullish on doing the pre-seed and passed because there were no lawyers on the team
Mr. Junestrand met one of that firm's partners at an airport recently and said "they were very, very bitter"
The firm now treats Legora as its lesson that domain expertise is not what you need to succeed
"I think that you need a willingness to learn about the market"
They bought that willingness early by exposing themselves to lawyers and spending as much time with them as they could
He qualified the answer by market. Quantum computing or fusion would be a different matter. "But legal was easy enough for us to pick up as we went along."
Betting the Models Would Improve Instead of Fine-Tuning One
Mr. Alströmer said he had spoken with founders about the next categories where the models are not yet good enough but will be, and asked what gave Legora the confidence to build on a model that was good but not good enough — the same question facing founders in mechanical design or anywhere else the models fall short today.
The tribal knowledge in 2022 and 2023 was that you had to fine-tune a model
He recalled Bloomberg spending "millions and millions and millions of dollars building a Bloomberg law model"
Legora went the other way partly from conviction and partly because it had no money: "The models will keep improving."
He added, on stage at YC, "I'm sure Sam is going to come up on stage and say that later today."
The work is delivering the value the models generate to your market
The first generation of ChatGPT was unusable in law, both because the model was not good enough and because it did not support private conversations or European data hosting
His first sales pitch was "We are like ChatGPT, but we are compliant in Europe." He called it not a very strong pitch, and one ChatGPT would obviously close later, but good enough at the time
"You need to build for the world today and maybe one step ahead."
Maximize the opportunity now and build customer relationships good enough to have the next conversation as the models improve
"I think that we have built 1% of all the software that we will build over the company's history."
Two Camps of Customers: Maximum Intelligence, or Cheaper Tokens
Mr. Alströmer noted the volume of model news over the previous two weeks, said he hoped somebody was making a documentary of the moment, and asked whether the open-weight releases had changed Mr. Junestrand's view that Legora is not attached to any one model.
Different use cases will pull different models, he said
Customer support optimizes for turnaround speed, latency and the cost of solving a ticket; he said Fin used to charge $1 per solved customer ticket
Law runs the other way, because the token and software spend is tiny next to the human expertise applied to the problem. On complex litigation "you're going to want to throw the most brain that you can on that problem"
The user base splits into two camps
One group wants to loop really complex models over problems for long stretches, running up large model costs
The other wants open source because it is cheaper on consumption
He said the answer is somewhere in between
The muscle he told the audience to build is evaluation
"the ability to eval new models and to eval new use cases because that is the superpower that then allows you to route things effectively"
Legora hired a lot of lawyers and made part of their job customer-facing and part of it building the use cases the evals run against
Locking in contracts with the labs and compute providers is its own problem at their scale: "We cannot have our system go down when a percentage of all the lawyers in the world are working with it and their clients rely on it."
Legora Released Its Benchmark on Friday, and Grok Was the Surprise
Legora published the Legora bench the previous Friday, after three years of work kept internal
It stayed internal because the practical choice was an OpenAI model or an Anthropic model, which made publishing benchmarks uninteresting. The system offered those two and let the user pick
The set of options is much larger now
"we found to our surprise that xAI and Grok was one of the best performing models, especially given the costs on our benchmarks"
The models are not on Legora's data processing agreement, so it cannot offer them to customers yet, though it will start to
Customers do hold model opinions, particularly banks and big law firms: "They don't want Chinese models."
His Father Lost Two Dot-Com Companies and Still Does the Odd PowerPoint
Mr. Alströmer said he had had two dinners with Startup School attendees arguing about when to start a company — people working at DeepMind, people one or two years into a job — and asked what triggered Mr. Junestrand to skip McKinsey.
"my dad had two companies during the dot-com boom, both of which went to zero after raising too much venture capital"
His father later ran a boat business in the archipelago where Mr. Junestrand grew up
They speak every day or every other day, and the family sometimes calls him a senior advisor
When the team was stretched for resources on one of its biggest customer pitches, his father came in and built the PowerPoint
He described himself as very competitive: "I used to play Dota 2 more than I went to school."
"It's the biggest arena that you can enter." You leave victorious, defeated, or okay
He still thinks McKinsey is a great place to start and that it teaches work ethic, and he did an internship there during his studies
Before GPT, Alströmer Says the Startup Ideas Were Boring
Mr. Alströmer said the last year or two before COVID produced few new ideas, and that when he sat down pre-GPT and listed every startup idea he could think of, most were pretty boring and most would not have worked
He said small companies then got a really big advantage over large ones, pointing to how long Microsoft's Copilots took to work
Mr. Junestrand remembered the fear on the other side of that: "Oh my God, we're so screwed."
Every lawyer already worked in Word and Outlook, and Copilot looked like the end of the opening. It did not work, and Legora kept building value in the space
"the amount of things you learn is a function of the amount of discomfort that you are willing to endure"
He was super introverted when Legora started, did not feel comfortable talking to customers and wanted to code as much as possible
Parting ways with an employee for the first time was one of the hard ones
He had worked at YC startups before founding his own, and told the audience to think deeply about who they start with and about the problem space
"I think we picked a space more than we picked a problem" — that legal and AI would be a thing was completely obvious, how it would be a thing was not, so they marched in the general direction and figured it out
The alternative, he said, is to find a specific problem, solve it clearly, generate ROI and expand from there
Everybody Knows the Monthly Goal
Asked how a Legora employee knows the chief executive is competitive, Mr. Junestrand said everybody there is like that, and that it self-selects.
"Everybody knows every monthly goal." He said quarterly goals are a little bit tricky
At the beginning there were no goals at all, just maximizing each day's outcome. At 750 people that stops working
"communication is the thing that really breaks when you scale"
The competitiveness shows up in engineering, which stack-ranks Legora's tabular review against the alternatives in the market
Most of the engineers on vacation in Stockholm in July are more productive than they have ever been, he said, because they are taking time off and coding on problems they want to win
Wins are celebrated, losses are mourned, and then immediately replanned
"I feel like people are in solution mode rather than blame mode."
Legora spent its early life chasing legal AI companies that were bigger than it, and is now much bigger than they are
Asked what happens to his motivation when Legora is the biggest, he said the company would either pick another enemy or find a way of being better than itself yesterday
Competitors are a carrot and an easy thing to rally around, provided nobody is overly focused on them
Ambition Is a Peer Group, Not a Personality
An audience question asked how a founder becomes more ambitious. Mr. Alströmer framed it with a company he works with that plans one European country, then another, then maybe the US next year; Legora entered many European countries during the batch alone.
"And now we're in 50 countries." He put that down to ambition rather than the product simply working
"I actually think the best way to become more ambitious is to have a peer group of other very ambitious people."
He was lazy in his final years of high school and his first two years of college, played a lot of video games, had an easy time at school and did not know how big the world outside his sphere was
A new, ambitious friend group made it click: "I went from doing zero internships to doing eight jobs in one year."
He said that is one of the best parts of YC, and cited a Paul Graham essay making the same point about the effect of ambitious peers
"I don't feel like I'm the boss. I feel like I'm a peer in the team that is running the company."
Legora's finance chief sometimes challenges the marketing team to be more ambitious, which he said a CFO typically would not care about
The idols were Spotify and Klarna, the biggest technology stories to come out of Stockholm
"Then we realized, wow, we have a bigger market cap than Klarna."
He called it a weird moment, because "we feel like we have done zero to one. And one to 10 and then 10 to 100 still remains."
Storytelling First, Then Being a Teammate, Then Shipping Speed
Asked which skills are more or less valuable for someone in school or leaving it, Mr. Junestrand named three, in order.
Storytelling is the one he had completely underappreciated
You have to sell the company to yourself first, because working this hard without believing your own story is very hard
"We need to sell the story of how working at Legora is going to be better than working at Anthropic." The company is competing with the labs for talent
Then investors, then customers. He said he does not know how to teach it, or how he learned it beyond a lot of exposure and a liking for stories
The second is collaboration, which he had to learn the hard way
Playing volleyball growing up, he would yell at a teammate who missed a serve, and told the audience what he yelled
"That doesn't get you invited to play the next game."
"It doesn't invite a followership and it doesn't invite a good vibe." He said he understood that in high school and had to pivot
On technical skills, he said velocity and the ability to iterate on customer feedback is probably still the most important thing to get started
He still drops a customer problem heard on a call into the product channel and asks how fast it can be turned around and the customer delighted
"That's a bit strange to do at a thousand person company scale, but still how we do it."
The counterweight is not to zig-zag away from the vision; speed against an incumbent is the superpower and should be leaned into
"Just because you can produce a lot of code with AI, it's not an excuse for not being a very good software engineer."
"We are hiring a hundred of them at Legora."
From Reactive Agents to Agents That Start Work on a Trigger
Asked what the most interesting problems are for someone joining, he named the shift from reactive agents to proactive ones.
For three years the pattern was that a user gave Legora a prompt or a set of instructions and it went off and did that thing
Now the product is being wired to different pieces of context so that a trigger starts the work
A contract arriving with the sales team gets routed to a Legora agent that starts on it in parallel, escalating to a lawyer if it needs to and otherwise executing the contract
Connect an entire data room and the agent organizes it and produces the due diligence report
"It's having the agent do things without you telling it what it should do."
The unlock he is aiming at is leverage per lawyer
"how does one lawyer produce the outcomes of a team of 10 lawyers with the help of Legora?"
There are engineering problems of pure scale alongside it: "We're spending millions and millions and millions of dollars on OCR and document parsing."
"I Need to Re-Qualify for the Job as CEO of Legora Every Quarter"
Mr. Alströmer noted that Legora had gone from zero to 100 million ARR in 18 months as of April and that the job had changed as fast as the company, invoking the Dropbox founder's talk at YC about the sequence of mini games a founder plays. The current mini game is hiring the people who will run the company, and Mr. Junestrand agreed it is not something he has done before.
He said most of the work up to any given point is what qualifies you for the next stage, provided you actually learn during each one
"I think you really need to put the company first, way ahead of your own ego in a way."
"I need to re-qualify for the job as CEO of Legora every quarter." He tells the executive team the same applies to them
"Running a sales team when you are one million ARR is very different from 150 million ARR."
The standard advice — build a network of people who have done it before — runs into how few people have done it this fast
Incoming executives come out of SaaS and its triple, triple, double, double growth model. "If you grow 40% year-on-year, you are doing fantastic." Mr. Alströmer: "That's not enough anymore."
David, Legora's CFO, came from Vanta, another YC company, which he said grew "from 30 million to 300 million in something like four years" and "from 200 people to 1300 people over four years". Legora is attempting both in one
The consequence is compression: say no to things brutally, and understand where your own competitive edge actually is
How Legora Got Jude Law: They Nagged
Mr. Alströmer said that when he landed in Stockholm at the end of April, the ads were all Legora ads with Jude Law, and told the audience to go watch them on YouTube.
The idea came out of a meeting with one of Legora's marketing agencies. The pitch was to redo famous legal movie scenes with Legora in them, on the joke that with Legora the movie would be over in a minute
Somebody then said Legora is AI-powered law, so what if they got Jude Law to be AI powered
"He thought we were completely unserious when we reached out the first time."
Hollywood is currently hostile to AI in scripts and films, and the approach came from a Swedish AI company asking to make billboards
"I think one underappreciated form of negotiation is nagging." He said his father taught him that, and that they nagged and nagged and nagged until the answer was yes
The condition was that Jude Law picked his own scriptwriter and cinematographer
He brought an SNL script director and the cinematographer from Oppenheimer to light the scene
Nobody from Legora was allowed on the set, and the finished films came out a few hours later
The campaign moved the brand outside the profession. He said people who have nothing to do with law now know what Legora is
Mr. Junestrand's bottom line is that Legora's advantage was never legal expertise or a model of its own: it was betting that the frontier models would keep improving, freezing sales for six months until the product could survive contact with lawyers, and hiring for how fast people climb rather than where they have already been.
Products, Companies & Tools Mentioned
Legora (The guest's company, which he calls the agentic operating system for lawyers; general availability October 2024, over 750 employees, and by his account over 3% of the world's lawyers as active users)
Mannheimer Swartling (One of the largest law firms in the Nordics, whose managing partner had called AI more artificial than intelligence two years before Legora moved into its windowless conference room)
Y Combinator (Rejected the team in May 2023 over the question of which lawyers they served, then accepted them two months later under a new name)
Benchmark and Redpoint (Benchmark invested $9.51 million, negotiated on decimals in Excel; Redpoint pre-led the Series A three weeks later, leaving a 10-person company with about $35 million)
OpenAI, ChatGPT and GPT-3.5 (The demo that got him in, the moment he dropped out of college, and the first pitch — ChatGPT, but compliant in Europe)
Anthropic (One of the two model providers Legora offered before this month, and now a competitor for engineering talent)
xAI and Grok (The surprise of Legora's own benchmark on cost-adjusted performance, and not yet available to customers because the models are not on the data processing agreement)
Google BERT and Bard (The pre-GPT models the founding team used to test whether legal work could be automated at all)
Bloomberg (His example of the fine-tuning orthodoxy Legora rejected: millions spent building a Bloomberg law model)
Microsoft Copilot, Word and Outlook (The existential scare — every lawyer already worked in those tools — that he says did not work for a long time and left the opening)
LangChain (Named among the agentic frameworks that kept changing under the product, which is why the rebuild targeted adaptability)
Azure (Where the traffic sits; a system going down under load is the failure that ends a law firm relationship)
Fin (His cost comparison from customer support, which he says used to charge $1 per solved ticket, against legal work where intelligence beats price)
Vanta (Where Legora's CFO David came from; grew from 30 million to 300 million and 200 to 1,300 people over four years, which Legora is compressing into one)
Spotify and Klarna (The Stockholm success stories the founders looked up to, until Legora's market capitalization passed Klarna's)
Sifted (Where an article about Legora's winner-takes-all culture ran)
McKinsey (The career he had planned and abandoned; he still says it teaches work ethic and did an internship there)
Dropbox (Its founder's YC talk on the sequence of mini games a founder plays, which framed the host's question about hiring executives)
Jude Law, SNL and Oppenheimer (The campaign that put Legora on Stockholm's billboards, made on the condition that Jude Law chose his own script director and cinematographer)
Dota 2 (Played more than he went to school, and his own evidence for how competitive he is)
Books & Resources Mentioned
Legora bench (Legora's own legal evaluation set, kept internal for three years and released the Friday before this talk; the source of the finding that Grok was the best performer for its cost)
Paul Graham's essays (Cited on ambition as a property of a peer group rather than a person, which he says is the best part of going through YC)
The Legora advertisements with Jude Law (The host recommended the audience watch them on YouTube)
The Dropbox founder's talk at Y Combinator (The mini games framing the host used for how a founder's job changes stage by stage)
Apple Podcasts (The episode on Apple)
Episode page (The show's own page for this episode)
Get the latest market chatter as it happens: