AI Strategy • 2 July 2026 • By Stanislaus Martins
A Firm That Bills by the Hour Has an Awkward Relationship With AI
A firm that bills by the hour has an awkward relationship with anything that makes the hour shorter. That tension explains most of what happens next.
A firm that bills by the hour has an awkward relationship with anything that makes the hour shorter.
Nobody says this out loud in the partner meeting. The conversation stays on quality and risk, which are genuine concerns. But underneath sits a commercial fact that shapes every decision: if a task that took four hours now takes forty minutes, somebody has to decide what happens to the other three hours and twenty minutes, and who they were billed to.
I buy professional services and I sell advisory work, so I have watched this from both ends. The firms handling it well decided the commercial question first and the tooling second. The firms struggling did it in the other order.
## Two things make this sector different
**The material is privileged or confidential by default.** A law firm holds matters under privilege. An accounting firm holds financial records before they are public. A consultancy holds strategy that has not been announced. Unlike most businesses, the sensitivity is not an edge case, it is the entire inventory.
So when an associate pastes a contract into a free tool to summarise it, or an analyst drops a client's management accounts into a chatbot to build a model faster, the exposure is not embarrassment. In legal practice it raises a question about privilege that nobody wants to have to answer.
I call it BROAI. Bring Your Own AI. In professional services it is common, and it is driven by the most conscientious people in the building, because they are the ones working at eleven at night trying to turn something around.
**The billing model resists the benefit.** This is the part that stalls adoption more often than risk does. If efficiency reduces billable hours and nothing else changes, the firm has made itself smaller. That is a pricing problem, not a technology problem, and pretending otherwise is why so many initiatives quietly stop after the pilot.
## Where it genuinely pays
**First pass document review.** Not the judgement. The reading. Identifying which clauses in a stack of contracts deviate from standard, so a senior person reviews forty pages rather than four hundred.
**Research and precedent.** Finding the relevant material faster. The verification stays human, always, and firms that skip that step will eventually be publicly embarrassed by a citation that does not exist.
**Drafting the routine.** Engagement letters, standard sections, status updates, the administrative correspondence that consumes junior time without teaching anybody anything.
**Proposal and pitch work.** Most firms rebuild the same credentials and methodology sections repeatedly. This is pure overhead and an obvious candidate.
Notice none of these replace the advice. They compress the work around the advice.
## The pricing conversation you cannot avoid
If a firm gets meaningfully faster and keeps billing by the hour, it earns less for the same output. There are only a few honest responses. Move toward fixed fees or outcome pricing on the work that is now predictable. Redeploy capacity into work you previously turned away. Or take the efficiency as margin on fixed fee work you already do.
What does not work is buying the tools and hoping the model sorts itself out. It will sort itself out in the direction of lower revenue.
## What good looks like
An approved tool set, chosen with the confidentiality obligations in front of you rather than after. A written rule about client material that every fee earner has actually read. Verification as a required step, not a recommended one. And a decision about pricing that the partners have made deliberately.
SWITCH, the framework I use with clients, opens with Scan and Score for this reason. In professional services the scan has to cover the business model, not only the systems.
A question for the next partner meeting. If a client asked whether their matter had been processed by an AI tool this year, who in this firm could answer, and how long would it take them?
## Frequently asked questions
**Is it safe to use AI on client matters?**
Only in tools the firm controls and has assessed against its confidentiality and privilege obligations. Free consumer accounts are not that, regardless of how careful the individual is.
**Will AI reduce fee income?**
It will if the firm keeps billing by the hour and gets faster without changing anything else. That is a pricing decision the partners need to make deliberately.
**Where should a firm start?**
Internal and administrative work first. Proposals, credentials, status reporting. No client exposure, immediate time saved.
**Can AI be trusted for legal research?**
As a starting point, never as a final answer. Every citation and authority must be verified by a person. Firms that treat this as optional will find out why it is not.
**Does this change how juniors are trained?**
Yes, and it is the unresolved question in the sector. A great deal of professional judgement was historically built by doing the routine work that is now being automated.
*Stanislaus Martins advises and trains enterprise teams across Sub Saharan Africa on the practical application of technology and AI in professional practice. Formats are on the [speaking and training page](/speaking), or start a conversation on the [work with me page](/work-with-me).*