How this is calculated, in full
Effort saved is repeatable share × (1 − 1/L). That saved effort is then treated differently depending on the contract:
On fixed-fee work, the price is agreed for a defined deliverable. Saved effort reduces cost while revenue holds, so the saving converts to margin.
On time-and-materials work, you bill for hours. Saved effort reduces billable hours, so revenue falls with cost. Direct labor cost falls too, but the overhead it was carrying does not — which is why the T&M line can bend downward.
This is a model, not a forecast. It assumes you keep the same headcount and hold price constant, and it ignores the cost of the platform itself, transition time, and any new work the freed capacity wins. Treat it as a way to see which lever actually matters, not as a prediction of next year's P&L.
Why this is the argument, not a marketing point
ACEC and Virginia Tech put it plainly: "If a consulting engineering firm can deliver the same work product 30% to 50% more efficiently in the future but still charges for its services by the hour, it is fundamentally in a race to the bottom."
AECOM's CEO Troy Rudd has described clients themselves raising it — "moving away from something like cost plus to something that looks more like a fixed fee." And ENR identified exactly this as a constraint on AECOM's own $390M acquisition of the AI startup Consigli: "the economics of time-and-materials contracts limit how much efficiency can be monetized."
If your fixed-fee share is low, the highest-return AI project in your firm this year is probably not a technology project. It is a contracting one.
This calculator is Appendix C of The AI-Augmented Engineering Firm, made interactive. The report sets out the four assumptions behind the model and what would break each of them.