What can the legal industry learn from medicine?
That’s the question I put to Liam Brown, the founder and CEO of Elevate, in the latest episode of Without Limitation. Liam founded the company in 2011 - but if you rewind all the way back, he trained as a doctor in London in the late 1980s.
He left medicine, but it completely changed how he thinks about process - triage; checklists; the difference between brain surgery and routine care. Medicine has worked out, over decades, how to unbundle work and give each part to the right person or system.
Liam has applied what he learned at medical school for more than 35 years. He looks at a process and asks: why does this work have to be done this way, by this person, in this place?
The technology has changed several times in those 35 years, but this question has pretty much stayed the same. And right now, with AI forcing every firm and legal team to rethink how work gets done, I think it is a more important question than ever.
🎧 Listen on Spotify or Apple Podcasts.
Lupl was acquired by Elevate in August 2026, so Liam and I now work together. It was a fun experience interviewing a colleague for the first time.
Two key lessons from medical school
Liam took two lessons from medicine into business.
The first came from the AIDS crisis in the 1980s. Back then, a needlestick injury with infected blood could kill you, so the NHS issued new safety syringes. The catch was that they needed two hands, which meant someone else had to hold the patient. Almost nobody used them. About six months later, a one-handed version arrived, and within two weeks it was everywhere. Liam’s lesson: if you make something easy to use, people adopt it. He saw the same pattern later with M&A lawyers and bankers: give them a better tool and they will change how they work.
The second lesson is about unbundling. Junior doctors at the time worked brutal hours and complained about it constantly. Then the NHS let other clinicians take patient histories and do initial examinations. The doctors, he says, were furious - how could they give good care if they hadn’t done that work themselves?
Liam’s diagnosis is candid:
“What was really going on was that we thought of ourselves as special, and the idea that you could unbundle any element of our work and give it to someone else really rubbed up against our sense of specialness.”
The end result, in Liam’s opinion, was that doctors actually had more time to talk to each patient and understand them as a whole person.
A career challenging the status quo
The unbundling thread runs through pretty much everything Liam has built since.
Typesetting in London. After leaving medicine, Liam ended up running the document production group of a financial and legal publisher: a heavily unionised shop, in a recession. Work on IPO prospectuses was expensive, slow and done in the city, with premiums if it ran into the weekend. His question: can’t we move this work electronically somewhere else? Desktop publishing made the answer yes, and that led to round-the-clock delivery and a very different cost structure.
A data room business. In the mid-1990s, Liam was running business development for a public company and was frustrated by how labour-intensive M&A diligence was (flying to another firm’s office and reading paper documents at 10pm, etc.). This again was one of the processes most ripe for unbundling. So, he started one of the earliest virtual data room companies, backed by Silicon Valley VCs, and sold it just before the dot-com bust (”more by luck than judgment”). That business went on to become Datasite.
Integreon. If work could be digital, maybe knowledge work could be unbundled too. Integreon’s first customer was McKinsey: moving research and presentation work to India so consultants could spend more time with clients. Then investment banks. Then, as the late adopter, law, working with Clifford Chance and Microsoft on what we would now call legal process outsourcing. Integreon grew to a couple of thousand people.
Elevate. After a strategic disagreement with Integreon’s private equity investor (Liam says, cheerfully, that he “actually got fired”), he sold his stake and used the proceeds to fund a new company. Instead of raising venture capital, he asked a group of people he respected, including John Croft, to go without pay for a couple of years in exchange for equity. The elevator pitch was “Accenture for law”: consulting, software, data and global delivery for the run-the-company work that has to be done right.
The new paradigm = “Experts + AI”
I asked Liam what has changed in the 15 years since Elevate launched. An oversimplified summary of the early days might be “move the work somewhere cheaper”. What is it now?
Liam’s view is that in-house teams increasingly want to buy complete solutions, rather than the constituent parts. He walks me through the moving parts of transformation in legal departments:
Legal process outsourcing. Unbundle some work and send it offshore.
Flexible lawyers. Hire experts locally, at lower cost than law firms (the Axiom and Lawyers on Demand model).
Legal tech. An explosion of tools, which someone has to buy, configure and run.
Legal ops. A new function to manage all of this complexity and decide what goes to the law firm, what goes to the ALSP and what stays in-house.
The problem with that sequence is that the client ends up doing the integration. They have to weave the experts, the delivery teams and the software back together themselves.
Liam tells me that Elevate’s response has been to build all the pieces under one roof. Legal professionals in the US, the UK, Australia and elsewhere. Global service delivery in places like Poland, India and the Philippines. Regulated law firms (Elevate acquired Halebury, an ABS in England and Wales, and has its own US law firm and an ABS in Arizona). And, since acquiring LexPredict in 2018, its own AI capabilities. As Liam put it: “We don’t pay Harvey or Legora. We know which machine tools to use for which steps in the workflow.”
In Liam’s view, customers want an outcome, delivered by experts who “look like us, sound like us, and have walked in our shoes”, without having to assemble the supply chain themselves.
I think this is a useful test for the current wave of AI-native law firms, many of which say they sell outcomes rather than hours. Who carries the responsibility? Who chooses and runs the tools? If the answer is the client, the client is still buying parts.
Two steps forward, 1.97 steps back
None of this was easy, and Liam is honest about that. From 2019 to about 2023, getting lawyers, process experts, delivery teams and AI to work together in a multidisciplinary way felt like “two steps forward and 1.97 steps back”.
Then ChatGPT arrived, and the tools got much more powerful almost overnight. But the interesting part of Liam’s account is what made the difference for Elevate: years of change management. The company already had a workforce used to experimenting with process and service design (and redesign). So, the new tools landed on prepared ground.
Which brings me to checklists. Liam and I are both fans of Atul Gawande’s The Checklist Manifesto. Surgeons and pilots use checklists because the work is too complex and too fast to rely on memory. Law adopted that discipline late. What I find interesting is that AI agents need checklists too (I wrote about this in How To Train Your Agent). A firm that has already written down how it does its work has a large head start in teaching an agent to help.
Is the billable hour finished?
In Liam’s view, no. Firms will build a portfolio of models, with hourly work alongside productised and subscription offerings. The reason is buying behaviour. For high-stakes work, he does not believe the unpredictability of the hourly bill causes enough pain to outweigh what clients get from an elite firm: certainty of getting the deal done, the brand, and the protection that comes with it. Clients will keep paying by the hour, he thinks, as long as they see reasonable attempts to be efficient and to use technology.
I broadly agree. Clients often want to pay by the hour for some work, and sometimes it would take longer to scope an alternative fee than to do the work. And, as I wrote in 2,000 Years of Legal Fees, business model change always lags technology change.
The more interesting prediction, I think, is about separation. Law is not one thing. Even inside one firm, there are many practices with different economics, and within them hundreds of tasks, some very exposed to AI and some not at all. Liam expects many mid-market lawyers to find they no longer fit inside firms whose rates and ambitions are moving upmarket. Where do they go? Independent practice, new firms, or platforms like Elevate. That is a significant shift, and I think we will see it play out over the next few years.
“A Precambrian explosion of lawyer entrepreneurs”
Liam has a strong view on the idea that lawyers inherently aren’t innovative. He called it “complete claptrap”. In his experience, law is full of innovative, client-focused people who need a platform to experiment.
He says he has had more calls in the last year than he can count from lawyers at firms like Latham and Kirkland who want to leave and start an AI-native firm to solve one specific problem. His phrase for it: “a Precambrian explosion of opportunity for lawyer innovation, because it’s easier than ever before.”
I see the same thing. My AI Firm Index keeps growing. Earlier this month, the FT reported that senior Freshfields partners had discussed spinning out the firm’s legal tech unit, in emails that ended up being sent to the whole firm. And not a week goes by without a partner asking me whether they should go and build something.
How many will? Fewer than are thinking about it, that’s for sure. Law firm economics are very strong right now (Partners get paid unbelievably well). And lawyers are trained to spot risk, so when a pitch deck lands on their desk, they read it differently from a pure entrepreneur: who owns the IP, what happens to the data? (I have had to rein in that instinct myself. You have to look for the opportunities as well as the problems.)
Part of Liam’s strategy for Elevate is to lower that barrier: to give entrepreneurial lawyers the infrastructure, brand and back office so they can focus on the problem that gets them out of bed, rather than on payroll.
AI resilience, AI opportunity
The last part of our conversation was about the next ten years. Liam frames strategy in two halves:
AI resilience. Do not build a large part of your business on products, services or capabilities that AI is about to eliminate.
AI opportunity. Build offerings that were not possible before AI.
He describes himself as “a species-level AI doomer, but a very, very business-level AI optimist”, especially in law. He is also clear that responsible use matters: AI should not be an excuse to hollow out a workforce. His aim is for Elevate to be “led by lawyer experts who are amplified by AI”.
He is backing that view with his own money. Elevate (according to Liam, around $150 million of revenue and 3,000 people) is in the middle of a management buyback from its investors, because he sees so many signals of latent demand for these new ways of working.
What this means
For law firms: Look hard at your work, task by task. Some of it is brain surgery. A lot of it is routine care that should be done very differently. It makes no sense for the partner in the corner office to do a task someone (or something) else could do better.
For in-house teams: Ask whether you are buying parts or buying the outcome. If your legal ops team spends its time stitching vendors together, that is a cost too.
For AI-native firms and founders: If you sell outcomes, be ready to own the result, including when things go wrong.
For anyone rolling out AI: Adoption follows ease of use. And do the change management early, so the tools land on prepared ground.
Final thoughts
I’ll leave you with Liam’s reading recommendations.
The first is The Mars Project by Wernher von Braun. It is mostly maths, but the introduction is about how you assemble the imagination, and then the social, political, economic and technological will, to do something very ambitious.
The second is Marcus Aurelius’s Meditations, on being clear-eyed about what you can and cannot control. As we agreed, you have to be quite stoic to run a company :)
