If AI makes your team meaningfully faster and you bill hourly, you have a pricing problem before you have a technology problem. Every hour you save is revenue you no longer collect. That is the uncomfortable arithmetic sitting underneath every efficiency claim in our industry, and most firms have not said it out loud yet.
It does not mean adoption is a mistake. It means the benefit shows up somewhere other than where firms expect, and if you do not decide where, the market decides for you.
The arithmetic nobody puts on a slide
Take a task your team bills at forty hours. Suppose better tooling takes it to thirty. On a time-and-materials contract you now invoice thirty. Your effective rate is unchanged, your revenue on that task dropped by a quarter, and your cost base did not move, because you did not lay anyone off and the software has its own price.
The standard response is that you will do more work with the same people. That is true eventually and not automatically. It requires demand to be there, requires your business development to keep pace, and requires the freed hours to land on billable work rather than dissipating into the day. In the near term, on a fixed backlog, efficiency on hourly work is a revenue reduction that you are funding.
This is why the honest version of the ROI conversation is more complicated than a time-savings number, and why I argue in the four numbers that convince a partner group that hours saved on their own is the weakest of them.
Where the benefit actually lands
There are four places, and they are not equally available to every firm.
Fixed fee and lump sum work. Here the gain is immediate and entirely yours. You quoted a number, you delivered for less effort, the margin is real. Any firm with a meaningful share of lump sum work has the cleanest path to capturing value, and this is the first thing I look at when a firm asks whether adoption will pay.
Capacity you were turning away. If you have been declining work or extending schedules because you could not staff it, the freed hours convert directly into revenue you were previously refusing. This is the best case and it is real for a lot of firms right now, particularly anyone touching the data center buildout, where demand has been outrunning staffing for a while.
Work you were absorbing. Every firm eats hours: proposal writing, the fourth revision that was really the client's fault, QA that never got billed, the report rewrite. Efficiency here does not reduce an invoice because there was never an invoice. It improves realization, which is the metric your CFO already cares about and the least disruptive place to book a gain.
Higher value work with the same headcount. The longest path and the most durable one. If the mechanical hours shrink, the mix shifts toward judgment, and judgment is what commands a premium. This only works if you actually reprice for it rather than quietly delivering more thinking at the same rate.
Efficiency on hourly work is a discount you give the client without deciding to.
What I would not do yet
I would not lower rates to reflect the tooling, and I would not volunteer a discount because the work took less time. Not out of gamesmanship, but because the value delivered did not decrease. The client is buying a stamped deliverable and the judgment behind it, not a quantity of labor, and being first to reprice downward in a market that has not moved is a poor trade.
I would also not build a fee model around efficiency you have not measured. Firms are quoting more aggressively on the assumption of gains they have not verified on their own projects. If the gain turns out to be twelve percent rather than thirty, you have priced a job you now have to deliver at a loss. Measure first on real work, then let the number affect a proposal.
What clients will eventually ask
Sophisticated owners are going to start asking why the fee did not move. Institutional clients, large developers, and public agencies have procurement people whose job is to notice exactly this.
The answer that holds up is about value and risk, not hours. You are pricing a sealed deliverable, professional liability that sits with you for years, and the judgment that decides what the model got wrong. None of that got cheaper. What got faster was assembly, and assembly was never the thing they were actually buying.
That is a defensible position and it is easier to hold if you have said something about your methods early rather than being caught out later, which is the practical case for telling clients you use AI on your own terms.
The decision worth making now
Look at your revenue split between hourly and fixed fee, and look at whether you are currently capacity constrained. Those two facts determine your entire strategy and most principals can answer them from memory.
Mostly fixed fee, or turning work away: adopt aggressively, the benefit is yours and it arrives fast. Mostly hourly, with a soft backlog: adopt anyway, but direct the gains at the work you were absorbing rather than at billable tasks, and do not expect the first year to show up as revenue growth. Somewhere in between, which is most firms: sequence it, and start where the work is fixed fee.
What does not work is adopting the tools, generating real efficiency, and never deciding where the benefit goes. In that version it leaks out as lower invoices on hourly work, nobody can point to a gain, and the partner group concludes the whole thing was oversold. That is one of the quieter reasons pilots get judged a failure even when the technology worked exactly as promised.
Working out where the benefit should land in your specific revenue mix is part of what our readiness and roadmap engagement does. Start a conversation.
← Back to insights