Options for AI Charges in Law Firm Fees
- James Markham

- 1 day ago
- 2 min read
The landscape around how AI vendors charge for use is becoming more complex
On the basis that most firms using AI likely find themselves:
(a) in a transitionary period moving from fixed AI subscriptions to consumption based charges,
(b) with a mix of fixed and hourly rate arrangements with their own clients, and are
(c) struggling to move away from cost- or competition- based pricing towards value based pricing
The matrix below shows a simplied view of what you might consider in navigating the next 6-12 months
Because this is a 2x2 matrix, it goes without saying that this uncovers a deep and meaningful foundational truth about the universe 🙂 , but some practical footnotes (if not outright caveats) may help nonetheless...
Firstly, this is what you might do to defend profitability in the light of increasing costs and a changing basis for those costs being charged. You would likely do different things if you were trying capture volume or market share. Intent around outcomes is important
More specifically:
(1) and (2) need be no more complicated than to say we've added x% to the cost base, we need to increase realised rates and fixed fees by y% to square off profitability (easy to say, if not easy to do)
For those now looking at variable charges from AI vendors, I suspect (3) is the common path of least resistance for most firms with a mix of fixed fees and hourly rates
Whilst relatively straightforward, the main disadvantage is you are reinforcing cost to deliver as being the main driver of the fee, as well as opening up line item level challenges from clients on token use
For the more adventurous looking to break that link, (4) is hard and (6) is harder still - you're into the dark arts of activity based costing and would need to model with some confidence typical consumption patterns to then bake that back into the fixed fee or an uplift in rates. I don't think many firms are able to do this outside of highly standardised services, just for the fact it's all still rather new
But the prize here is you can start to divorce price to client from cost to deliver, whilst taking comfort that you're not taking a bath on matter level profitability
In practice, this can all get (even!) more complicated quite quickly with more exotic AFAs, and once you pull on the thread of pricing you will likely need to consider broader questions around clients and markets, service design and delivery models
And, separate to how do we deal with the AI costs, is the more substantive question around how do we share the efficiency gains with clients in an equitable way - but that is a post for another time
As a starter for ten, hopefully helpful!




Comments