The legal industry is obsessed with what AI will do to law firms. The more important question is what AI will allow their clients to stop buying.
For decades, firms sold three things in one package: judgment, labour and capacity. You hired the partner because she had seen the problem 50 times before, but also because your legal team didn’t have another 50 hours to give.
AI is pulling that package apart. It makes legal production cheaper and lawyers more productive. It does not make judgment, ownership or operating experience abundant.
The incentives are different
When AI saves an in-house lawyer an hour, the company wins. That lawyer can handle another contract, clear part of the backlog or avoid sending work outside.
When AI saves a law firm an hour, the economics are very different. Utilization, leverage, staffing and pricing enter the conversation. One side is rewarded for eliminating work. The other has spent decades monetizing how much work gets done. In 2025 alone, law firm rates increased 7.4%, compared with 2.8% inflation (Thomson Reuters Institute, Law Firm Rates Report 2026).
That makes the rush by firms to build AI and Knowledge Management teams fascinating. They’re helping clients automate work, build capability in-house and, theoretically, need fewer outside hours.
But it raises an awkward question: if a partner helps a client cut next year’s legal bill by 20%, who gets rewarded for that?
Big Law may be dipping its toes into McKinsey work while still operating on law-firm economics.
More work is going to move inside
At 鶹Ƶ, we’re seeing the shape of that shift.
We’re working with one of Canada’s largest funds to embed an AI-specialist lawyer part-time alongside its legal team. His role is focused on using AI to make the legal team more efficient.
But what makes it work is that he understands how in-house legal teams actually operate. He’s been one of them.
Legal departments are full of work that should be happening but isn’t. Contracts pile up. Sales waits. Product needs support. AI governance lands on a docket that was already full.
Plenty goes outside not because it requires exceptional judgment, but because the internal team has run out of room.
If five internal lawyers suddenly have the productive output of seven, the question isn’t whether you still need seven lawyers. The better question is: What can we stop sending to external counsel?
This is how the law firm gets unbundled
Historically, companies bought scarce judgment bundled with labour. You wanted the partner who had handled 30 similar acquisitions, but wrapped around that partner were associates, research, diligence and enough horsepower to hit the deadline.
AI gives legal departments more of that horsepower internally. Experienced embedded lawyers give them something equally important: ownership.
AI cannot own the commercial docket, build trust with Product or understand which risks matter to the CEO this quarter. An experienced operator can.
So the question moves from “Who can do all of this for us?” to “Which parts actually require external judgment?”
The future legal department will be layered: a core team that knows the business; AI handling more production, research and triage; flexible embedded lawyers bringing operating experience and ownership; and elite outside counsel reserved for exceptional judgment and bet-the-company matters.
That is a very different model from the binary choice most GCs inherited: hire someone or call a firm.
More work moves inside. Scarce judgment stays premium.
That isn’t the death of the law firm. It’s the unbundling of it.
For most of the last century, the legal industry was organized around the law firm. AI may finally reorganize it around the legal department.
.png?upscale=true&width=1120&upscale=true&name=Mona%20Newsletter%20(2).png)