Building a law firm from scratch: what five people who actually do this for a living told a packed room at LegalTechTalk 2026
Paul Foster, Shainul Kassam, Clare Murray, Lucy Murphy and Rory O’Keeffe
Today, I moderated a panel at LegalTechTalk2026 grandly titled "What If You Were to Build a Law Firm from the Ground Up?" It promised forty minutes of intellectual sparring, no script that anyone was legally permitted to adhere to, and five people who fundamentally do not agree on how the world works.
Our architectural team consisted of Clare Murray (the boutique employment maestro of CM Murray), Shainul Kassam (who built Fortune Law from a kitchen table), Lucy Murphy (Chief Growth Officer navigating the glorious supertanker that is Linklaters), and Paul Foster (ofsa.global, who watches law firms make expensive tech mistakes across six continents).
I deliberately approached the session like an aggressive Grand Designs project. First, the foundations (structure), then the mortgage (capital), the household rules (governance), the staff (people), the smart-home automation (technology), and finally, the slightly damp truth in the cellar that none of us wanted to admit was the actual answer all along.
20 years ago, if you had handed me a blank sheet of paper and asked me to design the law firm of the future, I would have got it wrong. Confidently, expensively, and spectacularly wrong. I would have built something that looked exactly like every mahogany-heavy firm I had already worked in, but with a slightly flashier logo and perhaps an unnecessarily artsy water fountain or fish tank in reception.
The reason this panel mattered to me is that none of the four people on it would make that mistake. And as is always the case with lawyers, watching where they aggressively disagreed told us infinitely more than watching where they politely nodded.
Structure: Get the Blueprints Right
Clare opened the bidding with "Norm Law", the Blackstone-backed outfit boasting a tiny core of senior lawyers and a small army of legal engineers. They built an AI platform first, and then essentially wrapped a law firm around it like a high-tech blanket.
Now, I pushed her on this. "Norm Law" is currently the single most overused case study in legal tech. It’s the darling of every PowerPoint presentation on earth. I wanted to know if she actually believed the hype or if she just found it aesthetically pleasing.
She landed somewhere beautifully honest: it’s not that every firm should slavishly copy it, but that the sequencing is the real lesson. You build the digital platform first, and let the practice follow. You do not retrofit AI onto a 19th-century partnership model that was never designed to hold the weight of a smartphone, let alone an algorithm.
My own view, which I tossed directly at Paul, because he is the only person on stage who observes this madness across global markets rather than from inside the bubble of a single firm, is that most "AI-first" claims by law firms are pure decoration. They are the legal equivalent of putting a spoiler on a Nissan Micra.
Paul’s answer confirmed this without him having to alienate his entire client base. He politely called it "noise," noted that every single firm claims to be entirely unique and special (spoiler: they aren’t), and observed that the attitude has only genuinely shifted from marketing fluff to reality in the last six months.
The Takeaway: If you are building a firm today from scratch, ignore what other firms say they are doing on LinkedIn. Build your own platform around your own data from day one. That is the part you cannot retrofit later, and it is also the part nobody puts in their glossy pitch decks.
Capital: The Question I Think Most Founders Ask Backwards
Shainul built Fortune Law from a literal kitchen table with zero outside money and an enviable amount of grit. Clare’s world, by contrast, now dances with private equity, litigation funders, and Alternative Business Structures (ABS) that simply didn’t exist when either of them took their first steps.
I framed this for the room as a classic, irreconcilable tension: Private Equity wants a highly profitable exit in three to five years; a great law firm requires decades of slow, boring trust capital with its clients.
Nobody on the panel actually resolved that tension, mostly because it cannot be resolved in the abstract. However, what both Shainul and Clare agreed on, from diametrically opposed starting points, is that the real question founders ask is completely backwards.
They ask, "Can I get the money?"
What they should be asking is: "What am I actually trying to build, and exactly how much of my autonomy am I willing to hand over to a guy named Rupert in a gilet to get there faster?"
Asking for the cash before answering that second question is the single most expensive mistake I see in the modern market.
Governance: The Sharpest Answer of the Day
I asked Lucy directly whether Linklaters would still be a traditional partnership if she were building it from scratch today. I asked this fully aware that there was a very senior, very real Linklaters partner sitting in the front row, staring at her. It was a beautifully uncomfortable moment.
Her answer was the absolute highlight of the session. Yes, she would still build a partnership—but she would build it like a three-headed beast.
|
BUSINESS ONE
High-End Advisory & Complex Disputes
Pure Judgment
|
BUSINESS TWO
Productised Legal, Fixed-Cost Tech
Law-as-a-Product
|
BUSINESS THREE
Ongoing Subscription-Style Services
Always-On Coverage
|
Three separate business units. Each with its own profit and loss sheet. Different recruitment, entirely different pricing, and critically, different routes into partnership depending on whether your core strength is brilliant legal judgment, sophisticated technology, or data architecture.
Then she dropped a statistic that should cause anyone currently planning a traditional legal career to spill their coffee: across the largest US firms, the share of partners holding full equity plummeted from 72% in 2010 to just 43% in 2024.
Fewer than half of the people with the word "Partner" on their email signature actually hold the ownership stake that the title used to guarantee.
I don't think that’s necessarily the apocalypse it sounds like. A narrower, nimbler equity tier that can actually make a executive decision without forty people needing to sign off on it in triplicate might actually be a functional business. But it is a massive structural shift, and the vast majority of lawyers under thirty have absolutely no idea it is happening to the ladder they are currently killing themselves to climb.
People: I Do Not Believe the Disappearing Junior Lawyer Story, and Neither Does Shainul
There is a terribly lazy, dystopian argument floating around conference circuits right now that AI will entirely eliminate the need for junior lawyers.
Shainul doesn't buy it. In fact, she brought a live exhibit to prove his point: Tycho Orton, a legal engineer at ClearyGottlieb, who happened to be sitting in our audience. Tycho carved out an entirely new, indispensable role for himself by simply raising his hand for every single unglamorous AI project that the senior partners were too terrified or too bored to look at.
It’s a genuinely inspiring story. It is also, if we are being completely honest, a terrible talent strategy for a firm. You cannot hire for serendipity. You cannot build a business model on the hope that your graduates will randomly happen to be self-starting tech geniuses.
What junior lawyers need right now is not less training; they need a faster route to judgment. The repetitive, mind-numbing groundwork, the 18-hour days spent manually checking contract cross-references, which used to accidentally teach that judgment over five years, is vanishing beneath them.
Clare’s worry sat one level higher. The explosive growth of consultancy-style practices (firms consisting entirely of senior hotshots with zero junior intake and no structured knowledge transfer) solves an immediate corporate cost problem. But it creates a terrifying ten-year supply chain crisis.
My honest reaction on stage was one of mild existential dread: nobody in that room had a good answer for who exactly becomes the senior specialist in 2036 if every firm quietly stops training the 24-year-olds today. I left it as an open question because I certainly haven’t solved it either.
Technology and the Liability Problem Nobody Priced In
Clare described her own experience testing the current crop of legal AI tools as "strategicwhiplash." I told her on stage that it was the best phrase I’d heard all year, and I stand by it. It perfectly captures that dizzying feeling of alternating between "Wow, this is magic" and "Oh dear god, it’s hallucinating fake case law again."
Her key finding was the massive gulf between tools built on technically verified, closed legal databases and those built on general, open-world models. And then there is the lock-in risk: if you build your entire workflow inside someone else's proprietary tech platform, you will quickly discover the exit costs more than the original contract ever disclosed.
But the part of the tech conversation I bit into the hardest was client liability. It is the most dangerously underpriced risk in the profession.
Clients are now regularly turning up to meetings with AI-generated legal analysis, often drafted by a sleep-deprived founder at four in the morning, asking their lawyer to quickly double-check it.
Shainul delivered the sharpest line of the day here:
"Clients are no longer paying us for information. They are paying us for conviction."
I agree entirely, and I pointed out to the panel that this quietly and radically inverts the entire commercial relationship. The client genuinely believes they have done the heavy lifting themselves and now just wants a cheap rubber stamp.
The lawyer, of course, cannot give that stamp. The output isn’t verified, the prompts that produced it are totally invisible, and suddenly, the partner isn’t just supervising junior associates, they are supervising the client’s unvouched homework. And they are doing it for far less goodwill and a lot more client resentment.
Data: Where Everyone, Including Me, Ended Up Agreeing
Paul closed the technology segment with the neatest framing of the entire forty minutes. If you strip away the sci-fi noise, every single law firm on the planet sits on the exact same four pools of data:
People
Clients
Matters
Finance
He calls this "domain intelligence." This data already tells a firm precisely what it is uniquely brilliant at and precisely where it is bleeding money—and yet, almost nobody looks at it.
His comparison between the BlackBerry and the iPhone landed beautifully in the room. The firms that move now get that crucial, year-long head start that separated Apple from Research In Motion permanently. The firms that wait get to watch history happen to someone else from the sidelines.
This was the exact point in the session where I stopped playing the neutral moderator and made my own case. The panel had spent thirty minutes talking about structure, capital, governance, and people, and every single thread led back to the exact same room.
Not one of those four things is your actual differentiator. Data is.
The firm that understands its own data, and possesses the raw courage to act on what that data says (even when it violently contradicts the partners' favorite, deeply cherished assumptions about themselves) is the firm that wins the next decade. The other three models are just infrastructure for getting to the scrapheap.
The Close
As we entered the final stretch, I banned all corporate fluff and demanded a single, definitive closing thought from each panelist. No caveats. No legal fine print.
Thanks to the tape, we have their exact testimony:
Shainul went straight for the entrepreneurial juggernaut: > "Starting a law firm can be quite easy, but sustaining a law firm is a completely different matter."
Lucy struck fear into the hearts of traditionalists everywhere by predicting the inevitable collapse of the old-school pyramid: > "You can imagine over the next few years that the number of equity partners within law firms is going to reduce, but I think that's good, because you'll see your decision making and governance will [otherwise] be an inhibitor to grow."
Clare threw her weight behind the fluid, evolutionary power of a modern firm: > "The legal sector is great because it has so many different models... but there's multiple forms to do things in law, with strength in being able to train and mentor and career along."
Paul brought it all home with a stark, digital evolutionary warning: > "What's really important is what the data is telling you, and how you're going to get back them that allows you to use that data before... within a year, 18 months or worse, you're going to be left stranded."
Then, it was my turn.
I looked out at the room, at rows of faces filled with the exact same nervous, electrified energy I had when I sat where they were sitting twenty years ago. I didn't want to leave them with a warning; I wanted to leave them with a blueprint for victory.
I told them that the traditional law firm model is a structure built for a past that no longer exists. But that isn't a crisis; it is the single greatest opportunity our profession has seen in a century. If you are starting a practice tomorrow, you do not carry the dead weight of legacy tech, bloated partnerships, or decades of bad data hygiene. You are light, you are nimble, and you get to build the foundations correctly on day one.
The law firm of 2036 will not be defined by the size of its mahogany desks, the complexity of its billing hours, or how many millions it throws at flashy AI tools. It will be defined by something far simpler: Courage.
It will belong to the founders who are brave enough to look at what their data is actually telling them, intelligent enough to act on it, and human enough to mentor the next generation of talent to do the same. The future isn't something happening to us; it’s something we get to build live, question by question. Go build it.
Rory O'Keeffe moderated this panel at LegalTechTalk 2026. He is a solicitor and the founder of RMOK Legal, fractional general counsel, a commercial law practice specialising in technology, AI governance, and digital regulation. He is the author of AI Advantage (2025) and a formally accredited Leading IT Lawyer with the Society for Computers and Law.
The Beyond The Fine Print podcast covers technology law for in-house teams and founders. Available on Spotify, Apple Podcasts, and YouTube.