If artificial intelligence changes who does the work and how quickly, then agency fee models built on hours and headcount are quietly obsolete. Remuneration is where the AI conversation gets real.
The economics have moved; the contracts haven’t
Agentic tools compress tasks once priced by time. McKinsey research suggests agentic AI could power up to two-thirds of marketing tasks. Yet most agency contracts still price inputs — full-time equivalents, hours, seniority mixes — as though nothing had changed.
When the cost of production falls but the fee is anchored to effort, the model and the economics diverge. Someone captures the productivity gain; the contract decides who.
The questions clients must ask now
What are we actually paying for — people, or outcomes? Who keeps the efficiency the AI creates? How do we value work an agent produces in minutes but that still depends on human judgement to be any good?
These are not annual-negotiation questions. They are design questions, and they belong in a deliberate remuneration and contracting exercise, not in the margins of a renewal.
Remuneration as a strategic step
Treated properly, remuneration design realigns incentives across the ecosystem: it rewards value rather than volume, builds in transparency, and makes the AI productivity dividend a shared, governed outcome rather than a silent transfer.
Common questions
How is AI changing agency fees?
Agentic tools compress time-priced tasks, so input-based fees built on hours and FTEs no longer match the economics of the work.
What should replace hourly agency fees?
Value-based remuneration, designed deliberately and transparently, with an explicit view of who captures the AI productivity gain.