AI and the Billable Hour: How AI Is Reshaping Legal Pricing

Richard Susskind has argued for decades that the billable hour is a perverse incentive structure — rewarding inefficiency and penalising the adoption of better methods. For most of that time, the legal profession has been able to absorb his critique without acting on it. AI is now making that position untenable.

The data from 2025–2026 makes this concrete. Client encouragement of AI use by outside counsel rose from 63% to 78% in a single year, according to the Blickstein Group's Law Department Operations survey. That is not a gradual trend — it is a market signal. Clients are not waiting to have this conversation. They are already having it.

The Pressure Is Already Here

The pressure is coming from two directions simultaneously.

From clients: Mark Smolik, CLO at DHL and a Fortune 500 benchmark for in-house legal expectations, puts it directly. Law firms increased innovation headcount by roughly 2,000% between 2014 and 2025 — from 16 innovation-titled roles in ILTA's 2014 roster to nearly 400 by 2025. Yet the Blickstein Group found that almost two-thirds of clients disagree or strongly disagree that their outside firms are innovative. The reason is directional: firms invest in technology that drives internal efficiency, not efficiency that appears on the client's invoice.

From within the economics of practice itself: the Blickstein Group projects average partner billing rates reaching $984/hour by 2030. CFOs and CEOs are already inserting procurement into outside counsel negotiations — treating legal services as a supply chain to be optimised, not a professional relationship to be preserved. GCs who cannot demonstrate cost discipline to their CFO will be forced to bring work in-house or route it to alternative providers.

The third pressure, specific to the Swiss and German market: only 18% of legal professionals currently measure AI ROI, according to the Thomson Reuters 2026 AI in Professional Services Report. About 40% are unsure whether their firm tracks these metrics at all. Firms that measure and can show AI ROI will win the client trust conversation. Firms that cannot will find themselves unable to explain why invoices look the way they do.

There is a fourth pressure, more direct than any of the above: clients are now using AI to audit the bills themselves. A pricing industry veteran with 25+ years of experience reports that law firm clients are already asking a specific, measurable question: "How much did you save working on my matters through generative AI last quarter?" This is not a general conversation about innovation perception — it is a line-item query entering the invoice review process. Firms that cannot answer it will not lose the argument. They will lose the client.

Simultaneously, the volume of client-generated AI drafts is growing. Business owners are using generic AI to draft contracts, employment letters, and compliance checklists — and sending them to outside counsel with "is this okay?" The billing model must account for this reality: the client who has already spent two hours with ChatGPT does not expect to be charged for four hours of drafting.

The Innovation Perception Gap

The gap between how firms see themselves and how clients see them is worth dwelling on. Firms spent heavily on innovation. Clients see no evidence of it. The reason, as Smolik explains, is that the investment went inward.

This is the "innovation perception gap" — a concept that applies with particular force in the SME legal market in Switzerland and Germany. A 15-lawyer Basel firm cannot spend CHF 500,000 on an innovation team. But it can be transparent with clients about what AI does and how it affects their bill. That transparency — showing clients that AI accelerated research, that the invoice reflects that — is visible client value at essentially zero cost.

The client conversation script that works: "We use AI to accelerate research and first-pass drafting. This is why your invoice is 30% lower than it would have been in 2022 for the same work — and the quality is higher because I spent my time on judgment, not lookup."

That sentence is more powerful than any innovation headcount.

Alternative Fee Arrangements in the AI Era

The good news for firms willing to adapt is that alternative fee arrangements (AFAs) become more economically viable, not less, when AI reduces the time cost of execution.

Fixed fees have historically been challenging for law firms because the variance in how long a matter takes makes pricing risk difficult to manage. AI reduces that variance for well-defined task types — contract review, due diligence, regulatory filings, standard pleadings. When you can predict with confidence that an AI-assisted review will take four hours rather than somewhere between eight and twenty, fixed-fee pricing becomes manageable. The AI variance is now your risk to absorb — and increasingly, your ability to do so is a competitive advantage.

Value-based pricing — billing for the outcome delivered rather than the time spent — is the logical endpoint Susskind has long advocated. A well-drafted SPA clause that prevents a €500,000 dispute is not captured by "2 hours at €400." Value pricing requires a different conversation with clients, but clients who understand it often prefer it. The challenge is that it requires firms to understand their own historical data well enough to price outcomes with confidence.

Subscription models are emerging for high-volume, predictable practice areas. A fixed monthly retainer covering all routine employment matters, for example, becomes viable when AI handles the volume work and human lawyers handle the exceptions. Mid-size company GCs particularly value the budget predictability.

A Phased Engagement Model That Works for Smaller Firms

Pure value-based pricing can feel abstract for a 10-lawyer firm. A more accessible starting point is a phased engagement structure that captures AI economics without requiring a complete pricing overhaul:

1. Scoping phase (flat fee, 2–3 hours): assess the matter, identify what AI can accelerate, provide a written estimate. Charge for this. It demonstrates methodology and builds trust before the main engagement.
2. Fixed-fee estimate: based on what you found in scoping, provide a fixed fee for execution. You have now absorbed the variance risk.
3. Execute at fixed fee: AI variance — whether a task takes 3 hours or 6 hours with AI assistance — is absorbed into your margin, not passed to the client as time-uncertainty.

The destination this model points toward is what some practitioners call the "concierge" model — analogous to concierge medicine, where clients pay for direct access, personalised attention, and proactive guidance rather than reactive problem-solving billed by the hour. By automating routine legal work with AI, lawyers reclaim the time previously consumed by production for the activities that actually differentiate them: understanding a client's business deeply, anticipating problems before they arise, providing strategic guidance that is specific to their situation.

A 10-lawyer Basel firm serving Basel's pharma SME corridor cannot out-resource Walder Wyss on a major transaction. But it can provide a quality of client relationship — rapid access, genuine sector knowledge, proactive monitoring — that a large firm structurally cannot offer. AI makes the economics of that positioning viable at smaller scale than was previously possible.

This structure is more accessible than pure value billing and more honest than hourly billing when AI is involved. It positions your firm as the one that prices with transparency rather than the one that bills whatever the hour count turns out to be.

Measuring What Matters

The Thomson Reuters 2026 report found that the 18% of firms that do measure AI ROI focus predominantly on internal metrics — cost savings, usage rates — rather than client satisfaction or revenue impact. That is the wrong focus.

The metric that wins clients is not "we saved 40 hours internally." It is "your invoice reflects 40% less time on research because we used AI, and here is what we did with the remainder: deeper judgment work, a second review pass, proactive flagging of three risks you did not ask about."

Showing clients how AI reduced their bill is not a threat to firm revenue. It is the conversion of efficiency gains into client loyalty — which, as the data shows, is currently the gap between what firms are investing in and what clients can see.

The Billing Leakage You Are Already Losing

Before restructuring your pricing model, there is a simpler intervention available: recover the revenue you are already performing but not capturing.

Research on attorney time-recording practices is consistent: attorneys who record time at the end of the day lose approximately 10% of billable hours. Those who wait until the end of the week lose up to 25%. In a small firm, this "billing leakage" is the difference between a thriving practice and a struggling one — and it has nothing to do with pricing models.

AI-assisted time capture tools — which reconstruct billable activity from document timestamps, email threads, and calendar data — address this directly. The intervention is not about changing how you price work. It is about capturing the work you have already done. For a firm with three to five fee earners, recovering 15% of currently lost billable time typically exceeds the cost of any AI tool in the stack.

This is the "start here" for firms not yet ready to renegotiate pricing with clients: fix the capture problem first, then redesign the model.

Practical Steps for Transitioning

Audit AI-automatable tasks first. Identify which tasks in your highest-volume practice areas can be partially or fully automated. Document review, research, first-draft contracts, regulatory submissions, and compliance checklists are the typical candidates.

Model the economics before piloting. For each task type, model three scenarios: bill traditional hourly, bill AI-assisted hourly (same rate, less time), and bill fixed fee. Understand your margin profile under each model before committing to a client conversation.

Pilot with one trusted client. Choose a relationship with sufficient trust and volume to test a new pricing model. Frame it as a pilot. Measure satisfaction, time, and margin over six months.

Update engagement letter templates. Most standard engagement letters are silent on AI use and AI billing. In Switzerland, the SAV/FSA AI Guidelines (adopted June 2024) require transparency about AI processing with clients. Clarify your firm's position on AI use, disclosure, and pricing before it becomes a dispute.

Train relationship partners. The billing conversation is a client relationship conversation. Partners who can explain the new model clearly — and quantify the value delivered — will retain clients through the transition. Partners who cannot will lose them to firms that can.

Start Here

If you take one action from this article, make it this: identify one practice area where AI has already changed how long a task takes, and have a single honest conversation with one trusted client about it.

Not a policy announcement. Not a fee negotiation. A conversation: "We have been using AI to assist with [task type]. Here is how it has changed the work, and here is what I think it means for how we should price your matters going forward."

That conversation — done once, done well — builds more client trust than a year of internal AI investment that clients cannot see.

The billable hour will not disappear overnight. It will erode — first in the most commoditised task types, then progressively across the stack. The firms that emerge strongest are those designing a coherent pricing model now, piloting it deliberately, and building the client communication skills to explain the value they deliver.

The Information Arbitrage Collapse

The deeper structural shift beneath billing disruption is the collapse of information arbitrage. Professional services firms historically sold access to expert knowledge clients lacked. AI now enables clients to approximate much of this analysis themselves. Professional services face "dual margin compression" — pressure from above (clients demanding outcomes rather than hours) and below (AI automating the work that justified those hours).

The durable moat is not analytical competence — it is trust infrastructure. The client who uploads a contract into Claude and gets 80% of what your associate provides still needs your liability backstop, your professional indemnity coverage, and your judgment on the 20% the AI cannot assess. Build that trust position now, while the transition is still orderly.

Get in touch to discuss how to model the economics of AI-assisted pricing for your practice areas.