In-house teams will force the change law firms won't

Most of the conversation about AI and the legal profession assumes the change will come from law firms. Firms will adopt the tools, firms will redesign their workflows, firms will invent new pricing. I’ve spent enough time selling into firms to doubt it. Not because the people in them aren’t capable - many are excellent - but because the incentives point the wrong way. The change is far more likely to come from the other side of the table: from in-house legal teams, who have every reason to want it and, increasingly, the means to insist on it.

The incentives sit with the client

A law firm that makes a task take a third of the time has, under hourly billing, made two thirds of its revenue on that task disappear. Firms know this. They talk about efficiency in pitches and then, quite rationally, move slowly in practice. There are ways around it - fixed fees, higher-value work, different leverage models - but none of them are comfortable, and comfort is a powerful force in a partnership.

An in-house team has no such problem. It’s a cost centre. Its general counsel is measured on outcomes, risk and budget, and sits under steady pressure from a finance director who would like legal to cost less next year than it did this year. Every hour of external work that becomes unnecessary is a clean win for them. When the economic upside of a technology lands almost entirely on one side of a relationship, that side is going to push.

In-house teams are building the capability themselves

The more important shift is that in-house teams are no longer waiting for firms to bring them technology. Many now have legal operations functions, their own tooling budgets and access to the same AI models as everyone else. A small in-house team with a well-configured tool can handle work that would previously have gone out to a firm without a second thought: first-pass contract review, NDAs, routine lease renewals, standard supplier terms.

This matters because it resets the baseline. Once a general counsel has watched their own team turn around a routine contract in an afternoon, they will never again happily pay for a week of a firm’s time on the same thing. The question stops being “is AI any good?” and becomes “why am I paying you to do this the old way?”

The questions are already changing

You can hear the shift in what clients ask. Panel reviews and pitch processes increasingly include some version of: how do you use AI on our matters, and how is that reflected in what you charge us? A firm that answers with a vague reference to its innovation team and an unchanged rate card will find that answer landing less well every year.

A general counsel I spoke to put it simply. She didn’t much mind which tools her firms used. She minded paying for time that no longer needed to be spent. That isn’t a technology demand. It’s a commercial one, and commercial demands from clients are the only kind law firms have ever reliably responded to.

What gets pulled in-house first

The obvious prediction is that routine, high-volume work moves in-house first, and complex, high-stakes work stays with firms. That sounds reassuring for firms until you think about what the routine work was doing for them. It was profitable, because it was leveraged through junior lawyers. And it was where those junior lawyers learned their trade. Losing it doesn’t just reduce revenue - it hollows out the training model and the margin structure at the same time.

Firms that assume they’ll keep the interesting work and let the rest drift away are betting that the interesting work alone can sustain the same pyramid. I’m not convinced it can.

There’s a subtler effect, too. An in-house team that has automated its routine work starts to understand that work in a way it never did when it simply sent it out. It knows what a sensible turnaround looks like, what the real effort is, and which parts genuinely need a lawyer. That knowledge doesn’t stay confined to the routine matters. It travels into every fee negotiation on the complex ones, and the information gap that firms have quietly relied on for decades starts to close.

What firms should do about it

The firms that come out of this well will be the ones that stop treating in-house teams as buyers to be managed and start treating them as design partners. That means a few practical things. Price on outputs where you can, so efficiency stops being something you hide from the client. Work inside the client’s systems rather than insisting they come into yours. Share the tooling and playbooks for routine work, and make your money on the judgement that sits above it. Ask your best clients what they’d take in-house tomorrow if they could - and offer to help them do it.

That last one sounds like commercial self-harm. It’s actually the only defensible position. The work is going to move either way. The question is whether the firm is part of the change or the thing the change happens to.

Professional services rarely reform themselves from the inside. The structures are too comfortable and the incentives too well entrenched. Change arrives from whoever pays the bill, once paying the bill starts to feel optional. In-house teams are reaching that point now, and the firms paying close attention to what their clients are building - not just what they’re buying - will be the ones still in the room when the new model settles.