Freshfields and other Magic Circle firms are ditching the billable hour. Bloomberg's reporting shows that AI is forcing a reckoning with the time-based fee model that has underwritten UK legal practice for decades. This matters to mid-market firms because it signals something more than efficiency gains—it signals that clients have finally stopped accepting the fiction that a junior associate taking three hours to draft a clause is delivering equivalent value to a senior partner doing it in thirty minutes. AI doesn't fix this moral hazard. It exposes it. And the SRA, already alert to fairness concerns under its 2024 Code provisions, is watching how firms transition away from hourly billing without creating new compliance risks around transparency and fee management.

What we are witnessing is not the first wave of legal AI adoption—it is the second wave, and it is angrier. The first wave gave us document review automation (Luminance, LawGeex) and contract analysis tools that promised to make existing workflows faster. Those tools worked. They saved time. They did nothing to challenge the economic model. Now firms are facing a harder question: if AI can do substantive legal work—research, drafting, precedent spotting, compliance checking—why are we still charging by the hour at all? The pressure is coming from clients (who now understand what GenAI can do), from regulators (who care about value for money), and from competition (firms that move to value-based pricing will win mandates). This is not a technology problem. Technology was never the constraint. The constraint was business model lock-in. AI has broken the lock.

Here is what we see firms getting wrong. They deploy Harvey or Copilot Legal and expect the billing problem to solve itself through marginal productivity gains. It doesn't work that way. A 20% efficiency boost in legal research still leaves you charging by the hour. You still have the same misaligned incentives. You still have the same opacity about what AI actually did versus what a human did. At Trovix, we take a different view: AI integration must sit inside a firm's actual business model, not on top of it. That means using tools like Trovix Aria not to make people faster at billable work, but to enable a transition to outcomes-based or fixed-fee work by giving fee-earners the confidence to commit to realistic timelines. It means using Trovix Sift not to extract data faster, but to eliminate entire categories of manual work so you can price on value delivered, not hours burned. The firms that will win are the ones that use AI to change the question, not to answer the old one more quickly.

What should a mid-market law firm do this week? First: stop asking 'how do we make AI faster' and start asking 'what should we stop charging for at all?' Second: audit your current matter intake and billing practices against the FCA Consumer Duty guidance (PS22/9) and the SRA's fairness expectations. Are you currently transparent about how fees are calculated? If not, adding AI on top of opaque billing will only amplify the problem. Third: identify one practice area—employment law, conveyancing, regulatory advice—where you could move to fixed or outcome-based fees within six months, and pilot an AI-enabled workflow there. Use the regulatory monitoring data you already have (or build it with Trovix Watch) to stay ahead of changing compliance rules as you experiment. The firms that move now will own their pricing models. The firms that wait will have clients and regulators decide for them.

Source: Bloomberg News