What actually happened
The news is short and the signal is large. According to Artificial Lawyer, Sunny Mann, the Global Chair of Baker McKenzie, told the Wall Street Journal that the billable hour will no longer be the default method for billing clients. That is the entire announcement as reported: a change of default, not a formal abolition of time-based fees.
Artificial Lawyer frames it plainly as “a vital step in the right direction” that “opens the door for large commercial law firms to really use AI at scale and on valuable work, rather than just using it where tasks are unbillable, or of low value.” The same piece notes it “may not be a total move off the hour,” but argues that when a firm this large and established openly welcomes the shift, it matters.
Two framing details are worth holding onto. First, this is a statement of intent from a firm leader, not a published fee schedule. Second, no figures are attached: no target percentage of revenue under alternative arrangements, no timetable, no named AI platform. Anyone presenting this as a finished pricing revolution is ahead of the facts.
For context on how AI is already being deployed across legal teams, see our overview of AI for lawyers.
Why the default matters more than the rate
The interesting word in the announcement is “default.” Law firm economics rest on time as the unit of value: the pyramid model, in which human legal labourers count up their time inputs to generate income. Artificial Lawyer puts the tension bluntly — automation “is the antithesis of the time-based pyramid model” because, as the piece repeats, “AI makes you go fast.”
When speed destroys the billing metric, AI gets pushed to the margins. Firms adopt it where work is unbillable or low value — knowledge management, internal drafting, research triage — and hesitate where it would compress hours on a profitable matter. Changing the default reverses that incentive without requiring anyone to ban the hour outright. Fixed fees, matter-based pricing and outcome-linked arrangements can coexist with hourly billing; the difference is which one a partner reaches for first when scoping a new matter.
Artificial Lawyer also notes a nuance worth keeping: there can be exceptions, for example where AI surfaces new facts that feed into a time-based system. The claim is not that AI and the hour are mutually exclusive, but that time-based value is broadly the wrong anchor for AI-assisted work.
How it differs from what came before
The source gives us no competitor pricing data, so an honest comparison is directional rather than numerical. What we can say, based on the reporting, is this:
| Dimension | Established default (pre-announcement) | Post-announcement position |
|---|---|---|
| Baseline billing method | Billable hour as the default | Hour no longer the default, per Sunny Mann |
| Where AI is economic | Unbillable or low-value tasks | Valuable work at scale becomes viable |
| Value metric | Time inputs | Not specified in the source; a shift away from time as the anchor |
| Scope of change | — | Explicitly “may not be a total move off the hour” |
The distinction from the decade of commentary Artificial Lawyer describes is authorship. For ten years, the argument for moving off the hour came largely from observers, consultants and clients. Here it comes from the global chair of one of the largest firms. That is a different kind of data point: not a prediction, but a stated internal position.
Note what is absent. There is no claim that Baker McKenzie has abandoned hourly billing, no published alternative fee menu, and no comparison to peer firms. Those gaps are where scepticism belongs.
Three concrete uses
1. Fixed-fee diligence and review for a firm. Take a document-heavy matter — a disclosed bundle, a set of contracts, a regulatory filing pack — and scope it as a fixed fee rather than an estimate of hours. The firm prices the output, absorbs the efficiency gain from AI-assisted review, and keeps the margin. Artificial Lawyer describes a fixed fee using AI as “a win win for firm and client,” which is the commercial logic in one line. The practical prerequisite is knowing your own cost base well enough to price without hours as the proxy.
2. Portfolio pricing for an in-house team. A legal department can ask for matter-based or subscription pricing across a defined stream of work — routine commercial contracts, say — with AI handling first-pass drafting and comparison. The shift in default gives procurement a legitimate reference point in negotiation: a firm leadership has publicly said the hour is not the starting assumption. That is useful leverage, though it remains a reference point rather than a published rate card.
3. Scoping discipline for founders and small firms. If you sell legal or legal-adjacent services, this is a prompt to define deliverables before estimating effort. Our guide to turning an ambiguous client brief into executable specifications is the operational version of the same idea: agree what is being bought, then price it.
One caution across all three: faster work is not automatically better work. Verification still sits with a human, as we set out in AI hallucinated citations: verify legal research before court.
Who and what this affects
Large commercial firms are the obvious audience. If the default changes at a firm of Baker McKenzie’s scale, the internal conversation about utilisation targets, leverage ratios and associate career paths changes with it. Those metrics were built on the hour; the source does not say how they will be reworked, and that silence is the hard part.
Clients gain a stronger negotiating position, particularly sophisticated buyers who already push for alternatives. Artificial Lawyer’s decade-long question — why clients had not forced the change — becomes less pointed if firms move first.
Legal AI vendors get a clearer commercial case. Selling efficiency into a time-based model means arguing for revenue cannibalisation; selling into fixed-fee matters means arguing for margin. The second conversation is easier.
In-house teams should expect pricing conversations to become about scope and value rather than hours. That rewards teams who can specify work precisely and audit output quality. The client relationship dimension of this is covered in our piece on the legal client journey.
Junior lawyers and trainees are affected structurally. If work compresses and value is no longer counted in inputs, training models that relied on volume of hours need redesign. That is analysis, not a claim from the source — but it is the logical consequence of the stated shift.
What to watch
Implementation, not intent. Watch for published fee structures, partner compensation changes and utilisation targets. A default only changes behaviour if the metrics behind it change too. The source names no timetable, so treat any announced dates from elsewhere with caution.
Quality controls. If AI is doing valuable work rather than marginal work, review standards must rise accordingly. Pricing innovation without verification discipline is a risk transfer to the client.
Confidentiality and data handling. Any AI-assisted fixed-fee model depends on knowing where client data travels and who can access it. This is a procurement and information-security question before it is a pricing one.
Intellectual property in outputs. When firms sell deliverables rather than time, ownership and licensing of AI-assisted work product becomes a contract term worth negotiating explicitly rather than assuming. Our note on copyright and ownership of content covers the general principles.
The limits of professional use. No pricing model changes the professional obligations that attach to advice. Faster and cheaper must not become synonyms for unverified.
How and when to try it
The source gives no product to trial, no pilot programme and no date. What it offers is a strategic prompt, and the useful response is internal. If you run a firm or a legal team, the practical step this month is to pick one recurring, document-heavy matter type and price it both ways: on an hours estimate and as a fixed fee informed by your actual AI-assisted throughput. The gap between the two numbers is your negotiation room.
If you are waiting for published terms from Baker McKenzie before acting, the source does not suggest they exist yet. Treat the announcement as a signal about direction and a useful reference in client conversations — not as a template to copy.
For background on where agentic tooling fits into this picture, see AI agents for lawyers.
Sources
- Artificial Lawyer: Baker McKenzie Shows the Way with Billable Hour Shift
Frequently asked questions
Did Baker McKenzie abolish the billable hour?
No. According to Artificial Lawyer, Global Chair Sunny Mann told the Wall Street Journal that the billable hour will no longer be the default method for billing clients. The source explicitly notes it may not be a total move off the hour. No published fee structures, targets or timelines were reported.
Why does a billing default matter for legal AI adoption?
Because AI makes legal work faster, and speed undermines a revenue model built on counting time inputs. When the hour is the default, firms are incentivised to deploy AI only on unbillable or low-value tasks. Changing the default removes that disincentive and makes AI-assisted work on valuable matters commercially coherent.
Should other firms copy the move immediately?
The source reports a leadership statement, not a published model, so there is nothing concrete to copy yet. A realistic first step is to price one recurring matter type both on hours and as a fixed fee informed by measured AI-assisted throughput, then examine the gap. Watch for named fee structures and revised internal metrics before restructuring.
Want to apply this to your own work or your team? Tell me where you stand, and I will reply personally.
Get in touch
