Clients Bypassing Law Firms with AI
- James Markham

- Mar 24
- 2 min read
Today, on "this time it might just be different"...
Increasingly it seems to me that there is a particular type of practice that is particularly exposed to AI in the short to medium term
It's not the consumer, or the SME business owner having a go with ChatGPT - I don't think many of those folks were paying clients in the first place (and I suspect they're creating more legal work in the long run!)
Rather it is the larger corporate buyer that has historically bought in moderate volume (in contrast to "high volume", "commoditised" and all the emotional baggage that comes with those terms)
There are a fair few law firms that have courted these businesses in recent years, with the rationale - why act on a corporate M&A/transactional real estate etc deal for one buyer/seller, when a serial acquirer will have multiple transactions we can act on
In the current climate, I think a lot of those serial acquirers are now exploring the use of general purpose AI tools (e.g. MS Copilot), and more specific legal (but still general, not point solutions) tools such as Harvey and Legora
I think there's a short term wave coming on the next panel renewal, or the slowing down of deal flow to a trickle, as corporates effectively start insourcing the work
This is a different dynamic to moving the work to a different law firm, significantly increasing headcount, or pushing the work to an ALSP. Those options were always there and still are
And let's not get carried away, it's likely not all work - there are other reasons beyond access to a bit of tech for external counsel in more significant bet-the-company type scenarios, and it may be disaggregation of parts of the work and a rebalancing of who does what tasks rather than it all goes
But from a law firm partner's perspective, I would be looking at large clients that send repeat instructions with a view to assessing the impact of AI on that work
Not "how can my law firm use AI to make that more efficient", but "where might clients start doing this themselves"
As a starter for ten, I would be looking at banks (some big investments being made in AI tooling there and they'll want a return), higher volume real estate clients (e.g. developers, asset managers) and I can see the somewhat excited rush into Private Equity M&A work of recent years come unravelling with DD on lower risk transactions moving inhouse
This is wholly different to "can you do us an AI efficiency discount"
It's also potentially far more impactful - there are whole practice groups and departments built around this sort of work for these sorts of clients
What are they going to do if (and I increasingly suspect when) that work is no longer being sent their way?




Comments