The impact of AI on billable hours has been the pressing question for law firms over the last couple of years as AI adoption has increased significantly. The structural question underneath it — what happens to the leverage model — is the one that may have an even greater long-term impact on firms. The good news is that leverage isn’t breaking. It’s being rebuilt, and the firms that redesign it deliberately, practice by practice, get to do it on their own terms.
By Dan Safran, Unbiased Consulting, and Marci Taylor, Mantra Legal Consulting
For two years, the AI conversation inside law firms has circled the billable hour. Will it survive? Should we move to fixed or value-based fees? How much efficiency gets passed to clients? These are reasonable, necessary questions. However, the real disruption is the economic engine beneath it: leverage, the ratio of junior producers to equity partners that has powered law firm profitability for a century.
Leverage is a machine for monetizing junior hours at margin. AI compresses precisely those hours— the low value, high-volume work that fills the base of the pyramid.
But “AI breaks the leverage model” is the wrong headline because it is not breaking. It is being rebuilt from an engine that runs on the volume of junior hours into one that runs on judgment and technology. The base compresses. The apex appreciates. A new hybrid middle emerges. And the firms that redesign that machine deliberately, rather than letting the market redesign it for them, are the ones that come out ahead.
The shift is already demonstrated in industry statistics
Start with basic law firm economics, because the change is already visible there. The Am Law 100 posted record profits in FY2025, but the gains were strikingly uneven: profit per equity partner grew 18.5% in the top quartile and just 5.3% in the bottom, with the gap inside the tier now widening as fast as the gap between tiers. Meanwhile the classic pyramid is already mutating. Non-equity partners now outnumber equity partners, and they have grown in number more than three times faster. Leverage was shifting before AI fully entered the picture. Now, that shift is changing and accelerating with the AI adoption.
Adoption at the frontier of the industry is near-total. Daily AI use among lawyers in tech-forward firms runs from the low 80s in percentage terms to nearly universal, with some firms logging hundreds of thousands of queries a month and usage multiplying several times over in half a year. Consider these firms as the leading edge, not the market average. However, the edge is where the model gets rebuilt first. The only consistent non-adopters we see are lawyers approaching retirement.
Then there is the law firm client, and here the Deloitte 2026 survey of legal-department leaders[1] is blunt about what is coming at firms. According to the study, clients expect a 20 to 40% external spend reduction over three years and the share of work billed hourly to fall from roughly 72% to 44% within two to three years. They believe AI could automate about 28% of their own legal work, concentrated in the same high-volume categories that firms staff with leverage.
Interestingly, in the same Deloitte study, 58% say their outside firms rarely or never raise AI benefits with them. Clients are not waiting. Some are already running AI across years of invoice narratives to decide what to insource. Some are testing outside counsel tasks with their own AI tools for efficiency and improved work quality. Silence from outside counsel does not preserve the status quo. It just means the client sets the terms.
Four systems you are actually rebuilding
Rebuilding leverage means reworking four connected systems, and each are critical to the outcome.
The revenue math. Leverage monetizes junior hours; compress the hours and margin does not survive on volume. But efficiency is not a return. Reducing the hours you bill without changing how you price simply lowers revenue. The rebuild is to change what you sell and how you charge: value- and outcome-based pricing, productized services that can be sold repeatedly, and matter economics built around smaller, more capable, tech-enabled teams.
The training math. The document review, diligence, and first-draft work being automated first is precisely the work that taught junior lawyers judgment. Remove the apprenticeship and you do not merely lose cheap hours — you lose your pipeline of future partners. Leaders are rebuilding development on purpose: rotations and continuing education, simulation and scenario training, and competency-based progression in place of years of experience.[2] Firms that wait for a regulator or a law school to solve this will be a cohort behind.
The talent math. The pyramid is becoming a diamond: fewer junior and senior generalist roles, a thicker middle of hybrid legal-technical talent such as legal engineers and embedded technologists, and a smaller, more senior apex. Crucially, that apex appreciates. As routine work commoditizes, deep expertise and client judgment become scarcer and more valuable. AI is raising the price of the best lawyers, not lowering it, and the hybrid talent are a key area of investment that cannot be overlooked.
The comp math. This is the one that most firms are still resistant to address, and it is the one that decides whether the other three hold. Leverage is not only a staffing shape; it is monetized through the compensation system — in profits per partner, origination and hours credit, the equity/income split. Change the shape without changing the incentive system, and partners experience AI as a pay cut and quietly refuse to adopt it.
There is no settled answer yet. Some firms grant billable-hour credit for AI work under innovation budgets, often with undisclosed caps; one prominent firm refuses credit on principle, treating AI fluency as part of the job rather than a bonus; another separates training credit from committee credit. The point is not which model is right. It is that compensation system design is now the problem, and firms that leave existing systems on autopilot will watch any hopes of AI transformation die quietly in the partnership.
Why the rebuild is practice by practice
Other than framing the leverage issue and making it a priority, very little of this happens at the altitude of “the firm.” Leverage is not one number. Corporate, litigation, and regulatory run entirely different pyramids, automate different work, serve different clients, and price on different logic. A single firm-wide roadmap papers over exactly the differences that determine where value sits and how to capture it.
The rebuild is practice by practice. Firms must start where the work is genuinely AI-ready and where the practice can actually convert the gain. Ripe practices areas are those where pricing can flex, comp can adapt, and, perhaps most importantly, a partner will lead it. Prove the full loop in one practice or a few, then expand. The alternative — a firm-wide platform rollout with no practice-level redesign — is the pattern that produces adoption without transformation, which is where most firms are stuck today.
The rebuilt model is better, not just leaner
Framed as loss, this is a grim story. Framed accurately, it is an opportunity. The rebuilt model is not only leaner; it is better. Although 78% of in-house respondents in the Deloitte study cited cost reduction as the most important benefit from their providers’ AI use, nearly 60% also cited improved quality as a benefit.[3] AI helps a team to review every relevant document rather than just one historical sample, deliver more consistent and insightful work, and frees senior lawyers for the judgment clients actually pay a premium for. It may also grow the market rather than shrink it: as the cost of delivery falls, work that was previously economically out of reach for the client becomes worth doing, and the real competition shifts from lawyers-versus-AI to firm-versus-firm for demand that does not yet exist. More than half of firms adopting AI already report revenue increases, not merely cost savings. A pyramid rebuilt as a judgment-and-technology engine can be more profitable and produce better client outcomes at the same time.
Get ahead of it, or explain it later
That leaves timing and posture. Firms that shape the client conversation, bringing clients into how AI is deployed and how benefits are shared, develop AI capability as a collaborative negotiation. Firms that stay silent meet a client who has already decided what the savings should be. The same holds internally: firms that reshape leverage before market pressure forces the issue will manage the transition; those that wait for client attrition to make the decision will be in a much worse position.
The question that decides the decade
The uncomfortable question the profession keeps deferring is whether the structures we inherited are the right ones for what comes next. Firms built on leverage, careers built on the pyramid, comp systems built on time-driven personal productivity have the challenge and the opportunity to design the law firm model(s) for the future. The firms that will define the next decade are not the ones with the largest technology budgets; they are the ones that rebuild their leverage model deliberately, practice by practice, before the market does it for them.
[1] Deloitte Legal, “The AI Imperative: Reshaping of the Legal Industry.” June 2026.
[2] Ibid.
[3] Ibid.