There’s a delicious little problem stalking the carpeted corridors of law firms. Legal-tech vendors arrive bearing dashboards and the evangelical promise that their software will save 80% of everyone’s time. Partners smile, nod and imagine efficiency. Then somebody in finance remembers the firm sells time. The room cools.
If a machine can produce a respectable first draft of a standard contract in seconds, the old arithmetic becomes awkward. Ten hours compressed into one isn’t merely a productivity gain – under hourly billing – it’s a nine-hour revenue leak wearing a futuristic tie. The technology hasn’t betrayed the firm. It’s exposed what the model was built to reward – elapsed time, leveraged junior labour and the solemn performance of busyness. An assembly line, in other words – only with better tailoring and less visible machinery.
AI and the Billable Hour: When Efficiency Eats Revenue
The contradiction is now measurable. Thomson Reuters’ 2026 AI in Professional Services Report surveyed more than 1,500 professionals in over 24 countries. It found that 40% of respondents now use generative AI across their organisations, up from 22% in 2025. More than 80% of current users engage with it weekly. Yet only 18% say their organisations track return on investment. In other words, the profession has bought the engine, started driving and decided the fuel gauge feels needlessly judgmental.
Thomson Reuters’ 2025 analysis of law-firm economics reports that 80% of law-firm respondents expect AI to fundamentally alter how firms conduct business, particularly how they price, staff and deliver legal work. The same analysis says each lawyer expects to save about 190 workhours a year. Those saved hours are economically useful only if the firm can sell the result rather than mourn the timesheet.
Clients have noticed. In Thomson Reuters’ Tech, AI and the Law 2024 survey analysis, 52% of surveyed in-house legal professionals said traditional billing hampers AI innovation and adoption, while 58% said generative AI should be factored into law-firm pricing. They’re not asking to subsidise a firm’s nostalgia for six-minute units. They’re asking why speed, which was sold as the benefit, is being disguised as an invoiceable tragedy.
Legal AI Fees: You Can’t Bill for Hours Nobody Worked (Duh)
The ethical position is becoming harder to finesse. The American Bar Association’s Formal Opinion 512 says lawyers billing hourly must charge for time actually spent, even where generative AI makes the work faster. It also connects AI use to competence, confidentiality, communication, supervision, candour and reasonable fees. It isn’t binding in South Africa, but the underlying logic travels well – a client shouldn’t be billed for fictional effort merely because the fiction appears in 0.1-hour increments.
South African firms must, of course, apply their own professional rules, mandates and fee agreements. But no amount of local exceptionalism changes the commercial fact – opaque billing and undisclosed automation are a trust problem before they become a disciplinary one. The safer answer isn’t to pretend the machine took longer. It’s to agree the scope, method, assumptions, exclusions and price before the work begins.
Value-Based Legal Pricing: Stop Selling Minutes and Start Selling Judgment
The alternative’s commonly called value-based pricing, although the phrase is occasionally abused to mean “the same fee, but with a brochure”. Properly used, it prices the client’s problem and the value of resolving it, not the provider’s internal production time. That value may include avoided loss, preserved deal certainty, regulatory survival, speed, confidentiality, reputational containment and the ability to sleep without imagining the Companies and Intellectual Property Commission personally setting fire to the boardroom.
David Maister’s enduring point, quoted and applied in a South African LexisNexis analysis of outcome-based pricing, is that professional firms don’t really sell time. They sell skill. AI makes the distinction impossible to ignore. Routine production is drifting towards commodity status. Judgement under uncertainty isn’t. Neither is persuading a hostile regulator, spotting the clause that technically works but commercially detonates the transaction, or telling a CEO that their preferred strategy is lawful, clever and still spectacularly unwise.
That’s where the premium moves – from production capacity to decision quality. A huge associate pyramid once signalled scale. Increasingly, it may signal that the firm owns an expensive human photocopier. A smaller practice with excellent systems, deep specialist knowledge and a few genuinely formidable thinkers can now compete for work previously protected by headcount. The moat is no longer how many people can review the documents. It’s whether anyone knows what the documents mean when the facts turn feral.
When the Algorithm Meets the Invoice
In 2025, some elite United States partners crossed the US$3,000-an-hour line. Reuters reported that some Quinn Emanuel partners, as well as Neal Manne and Bill Carmody at Susman Godfrey, had adopted US$3,000 hourly rates. Court filings cited in the same report showed Kirkland & Ellis partner rates of up to US$2,675 and Latham & Watkins partner rates of up to US$2,745 (Reuters, 27 February 2025).
That isn’t, by itself, evidence of collective madness. In bet-the-company litigation, clients may willingly pay startling sums for scar tissue, pattern recognition and the ability to stop a board from making a headline-sized mistake. The absurdity begins when the same high-status pricing logic is applied to routine production that machines can complete in minutes. Once speed rises and cost refuses to fall, innovation starts looking less like progress and more like a magician’s handkerchief trick performed over the client’s budget.
The vendor’s promise of “80% faster” therefore has a faintly homicidal undertone. It sells the firm an efficiency engine while quietly feeding the old revenue model into the blades. The firm then faces three unattractive choices – hide the saving, inflate the hourly rate or redesign the deal. Only the third is sustainable. Value pricing shouldn’t mean a perfumed version of the old invoice. It means agreeing what the client’s buying, what uncertainty the firm’s carrying and what success looks like before the work begins.
A workable model can be built in four moves. First, set a risk-adjusted base fee according to exposure, urgency and legal complexity. Secondly, add a strategy premium for genuinely novel judgement – scenario planning, negotiation design, evidence architecture or the call that prevents a clever legal position becoming a commercial bonfire. Thirdly, define an outcome component tied only to measurable success and only where ethically and legally permissible. Finally, use a crisis modifier for reserved capacity, compressed response times and senior availability. The equation isn’t mystical – price follows consequence, scarcity and judgement, not the number of people available to type.
That still requires discipline. Firms must define scope, assumptions, exclusions, review limits and decision gates. They must track the true cost of technology and senior quality control. And they must use matter data to test whether the price produced a fair margin. Fixed fees suit repeatable documents. Phased fees suit litigation and transactions. Readiness retainers, activation fees and milestone pricing suit crises. The principle is stable even where the format changes – the client receives certainty, the firm receives a return for expertise, and nobody invents ten hours because the algorithm was inconsiderately competent.
In Africa, the shift must be practical rather than theatrical. A South African LexisNexis analysis makes the central point plainly – clients buy resolution, not minutes. Firms need systems that reveal matter economics, distinguish routine workflows from strategic interventions and support consistent budgeting before a client asks why an efficiency tool has made the invoice more expensive.
Platforms such as AJS become useful at precisely that operational level – matter scoping, time and cost data, workflow standardisation, billing intelligence and better-informed pricing conversations. The technology’s job is to show what work costs, where judgement enters and which outcomes clients will pay to secure.
The Human Premium Isn’t More Hours
The prize isn’t merely margin. It’s a defensible commercial model. Thomson Reuters’ 2026 research says professionals increasingly expect AI to disrupt workflows, billing models and traditional roles, while most organisations still don’t track return on investment. Used sensibly, saved hours can become sharper strategy, better supervision, faster decisions and more useful contact with clients. That isn’t softness. It’s execution. AI can accelerate the first draft. It can’t own the consequence, read the room, negotiate the commercial trade-off or tell a CEO that the lawful option isn’t the wise one. The future firm won’t win by recording the most hours. It’ll win by knowing which minutes matter – and by proving that value to clients, boards and the legal-tech suppliers helping to deliver it.
Build a Pricing Model That Rewards Value, Not Delay
Firm leaders, business owners and legal-tech suppliers should stop measuring transformation through licences purchased and hours allegedly saved. Choose three repeatable matter types, calculate their real delivery cost, identify the client outcome and pilot a fixed, phased or risk-adjusted price. Then measure margin, turnaround, quality and client confidence. If the technology genuinely removes work, let the commercial model admit it.
AJS can help your firm connect matter scoping, workflow, time, cost and billing data so that you can price work with greater clarity and protect margin without preserving inefficiency. Get in touch with the AJS team about building a value-pricing model that fits the way your firm actually works. The alternative is to automate the factory while preserving the invoice for the assembly line – an impressive way to become efficient at going out of date.
(Sources used and to whom we owe thanks: Thomson Reuters Institute, 2026 AI in Professional Services Report; Thomson Reuters Institute and Ragunath Ramanathan, “Future of Professionals” report analysis: Why AI will flip law firm economics, 16 October 2025; Thomson Reuters, Legal Insight Australia, AI’s impact on law firm pricing and client service, 24 October 2024; American Bar Association Standing Committee on Ethics and Professional Responsibility, Formal Opinion 512: Generative Artificial Intelligence Tools, 29 July 2024; Reuters, David Thomas and Mike Scarcella, More lawyers join the US$3,000-an-hour club, as other firms close in, 27 February 2025; and LexisNexis South Africa, Beyond the billable hour: How AI is accelerating the shift to outcome-based pricing, 17 September 2025).
