Strategy
Pricing Creative Work When AI Compresses the Hours
Here is an awkward arithmetic problem the agency industry is mostly avoiding in public. Our teams now use AI tools across research synthesis, motion previz, code scaffolding, resizing, and first-draft everything. Tasks that took a week take an afternoon. If we bill by the hour, our reward for this efficiency is a smaller invoice. The billable hour has quietly turned from a pricing convention into a conflict of interest: the client wants outcomes fast, and the model pays us to be slow.
We have spent this year rebuilding how we price, and since we routinely advise clients to be transparent about hard transitions, here is ours.
Hours were always a proxy, and the proxy broke
Hourly billing survived for decades because effort roughly tracked value. Senior time cost more, complicated problems took longer, invoices approximated worth closely enough that nobody had to think hard. AI breaks the approximation in both directions at once. Some outputs got radically cheaper to produce while remaining exactly as valuable to the client: a brand identity that repositions a company is worth what it is worth, whether the exploration took two hundred hours or sixty. Meanwhile the scarce inputs, judgment, taste, accountability, the knowledge of which of forty options is right, were never hour-shaped to begin with.
Clients feel this too, from the other side. Procurement teams are already asking why timelines have not shortened and rates have not dropped. The honest answer is that some have and some should not, and a pricing model built on hours cannot express that distinction. So we stopped asking it to.
What we do instead
Our current model, refined over a year of live engagements, has three layers:
- Fixed-price outcomes for defined work. Brand platforms, identity systems, site builds, campaign development: priced on the value and complexity of the outcome, with scope defined by decisions delivered rather than hours consumed. If we get faster, we earn the difference. If we misjudge, we eat it. That symmetry keeps us honest and keeps the incentive pointed at the client's finish line.
- Retainers priced on capacity and seniority, not clock time. Ongoing partnerships buy a slice of a senior team's attention and a guaranteed responsiveness, measured in outputs shipped per quarter, reviewed together, adjusted annually.
- Productized discovery. The research and strategy phase is now a fixed-scope, fixed-price unit with named deliverables. It prices the most valuable thinking we do as a product rather than a loosely estimated preamble.
What we killed: rate cards as the primary artifact of a negotiation, and change orders denominated in hours. What we kept: total transparency about who works on what. Clients are not buying opacity; they are buying an incentive structure that finally points the right way.
The conversations this changes
Value pricing forces a discipline hourly billing let everyone skip: agreeing, up front, on what the work is worth and how we will know it worked. Those conversations are harder and better. They surface the real stakes, they filter out projects where the client cannot articulate the value either, and they turn the kickoff from a scoping exercise into a strategy session.
They also change how we talk about AI itself. We tell clients plainly which tools we use and where. Under hourly billing that disclosure was fraught, since every efficiency read as a discount owed. Under outcome pricing it is just craft: here is how we work, here is why the output is better and arrives sooner, here is the senior judgment you are actually paying for.
The next two years
Prediction: agencies that keep selling hours will spend the next two years in margin-compression trench warfare with procurement, negotiating over a unit of measure that no longer measures anything. Agencies that price outcomes will have a different negotiation available: what is this worth to you, and can we prove we delivered it? That second conversation is the only one with a future, and the firms that learn to have it now will not miss the hours. We do not.
Building something this could apply to?
We take on a small number of flagship projects each quarter.
Start a project