When the work takes a fifth as long, hourly billing turns doing it well into earning less. That is a structural conflict, not a transitional one.
Start a conversation with the AI Adoption Concierge, already scoped to the business model. Pick a starting point, or describe your situation directly.
Hourly billing still accounts for the large majority of law firm revenue — commonly estimated at around 80% of fee arrangements — and it has one property that AI attacks directly: it ties the fee to time rather than to the result. A firm that adopts a tool cutting research or first-draft time substantially, and continues to bill by the hour, has invested in reducing its own revenue on that matter. Every partner understands this instantly, and it explains a great deal of quiet resistance that gets misdiagnosed as scepticism about the technology. The pressure is uneven: it lands hardest on high-volume, repeatable, document-heavy work and barely reaches genuinely bespoke matters. Sorting the firm's book on that axis is the useful exercise.
Ranked from most exposed to least. The top of this list is where clients ask questions first.
Large compression, easily measured, and clients already know it. Most exposed.
Routine agreements and filings. First-draft time falls sharply.
Orientation is much faster; verification is not. The saving is real and smaller than claimed.
Large volume, defined questions. Compresses well and is priced competitively already.
Judgement-heavy and relationship-bound. Barely touched.
Time is not the value and clients do not want it optimised. Least exposed.
What firms are actually trying.
Whether efficiency shows up as margin, or as a smaller invoice.
A partner whose compensation runs on hours is being asked to bill fewer of them. No amount of training addresses that. Firms that fix the incentive see adoption move; firms that run more sessions do not.
No, and almost nobody is. The realistic move is a mixed model: keep hourly where time genuinely tracks value — complex disputes, novel advisory, anything unpredictable — and price the compressible categories differently. That is not a philosophical shift, it is a segmentation exercise, and most firms already do a version of it. What does not work is keeping pure hourly across the book while adopting tools aggressively, because that combination guarantees the firm captures none of the upside.
Answer it directly, because the client has almost certainly already formed a view and an evasive answer is worse than an unwelcome one. Sophisticated clients — particularly corporate legal departments, which are adopting faster than firms are — are asking this now, and they are comparing answers across their panel. A firm that can describe where AI is used, what it changed, and how that is reflected in the fee is in a considerably stronger position than one that treats the question as an attack. Some firms have found the conversation strengthens the relationship, because it is evidence of a firm paying attention.
Change what you measure, because exhortation does not survive contact with a compensation formula. If partner compensation is driven by hours billed, adopting a tool that reduces hours is a personal financial loss, and everyone can do that arithmetic. Firms that have made progress typically adjust the metrics — crediting matter profitability or realisation rather than raw hours — or start with fixed-fee work where the incentive already points the right way. Both are harder than running a training session and both actually work.
Measure now, reprice when you have your own numbers. The common error is repricing off vendor productivity claims and discovering the compression on your matters is materially smaller — usually because verification, client communication and judgement did not compress at all. Run a set of matters, record the actual hours against comparable prior work, and price from that. It is slower and it produces a number you can defend to a client and to your own partners, which is the number you need.
Describe your practice mix and how it is priced. The Institute will help you sort the book.