There is a question I keep hearing from smart founders and marketing leads who are thinking about hiring an AI-enabled agency. It goes something like this:
The short version
AI can make agency delivery faster, but speed is not the same thing as value. A strong AI-enabled agency should use automation to include more monitoring, analysis, iteration, and accountability inside the same commercial relationship.
If the only pricing story is "AI made us cheaper," the agency has probably misunderstood what the client was buying in the first place.
"If AI does it in five hours instead of fifty, why am I paying the same price?"
It is a fair question. And most agencies are terrified of it. So they dodge it, drown it in jargon, or quietly lower their prices and hope nobody notices the margin bleed.
Here is the honest answer. And it is not what either side expects.
The received wisdom is wrong
The received wisdom is that AI lowers cost, so it should lower prices. That is the intuition. Agency does the work faster, passes the saving on, everyone wins. Simple.
It is also how you turn a functioning creative business into a commodities market, where the only competition is who can undercharge the most. The logic feels customer-friendly. In practice it destroys the thing that makes an agency worth hiring in the first place.
You were never paying for hours
The thing is, you were never paying for hours. You were paying for the judgment behind the hours. The strategy. The experience that knows what to build before it starts building. The QA eye that catches the problem before it ships. The relationship that understands your business well enough to push back on a bad brief.
AI compresses delivery time. It does not produce strategy. It does not provide accountability. It does not manage the gap between what a client asks for and what a client actually needs.
At Foundry, our agents handle the mechanical layer: the first drafts, the data pulls, the structural work, the iterations. That compression is real and we are not pretending otherwise. But what that frees up is not time we bill for. It is attention. It goes into the thinking that still requires a human with real skin in the game.
A strategy that saves a client £50,000 in wasted ad spend has the same value whether it took three weeks to build or three hours. The price belongs to the outcome, not the clock.
AI changes what can be included
Here is where the conversation gets more useful. What AI actually changes is not the price of outcomes. It changes what agencies can responsibly include in those outcomes.
A retainer that previously covered "monthly strategy and content" now covers monthly strategy, content, monitoring, exception flags, competitor tracking, and a review loop, because the operational cost of adding those things has dropped to nearly zero. The client gets more. The agency margins improve. And the pricing makes sense because you are now selling a system that works continuously, not a batch of hours that arrives in an invoice.
The agencies that are doing this well are not charging less. They are charging for more things, at the same or higher price point, and proving the value through visibility: what ran, what was flagged, what was improved, and what it moved.
The agencies cutting prices are solving the wrong problem. They are answering "how do we look cheaper" instead of "how do we become more valuable."
The better buying question
If you are evaluating an AI-enabled agency and they have just slashed their retainer price because "AI makes it faster," ask what is inside that retainer. Nine times out of ten, they have cut price and scope together, and what you are buying is a lighter service dressed up in AI language.
The better question to ask any agency, AI-powered or not: what breaks if I do not pay you this month? If they cannot answer that clearly, the price is irrelevant.
At Foundry, the answer is: your monitoring stops, your exceptions do not get caught, your strategy drifts, and nobody with accountability is watching. That is what the retainer buys.
The firms that will win the next five years are not the ones who passed AI savings on as discounts. They are the ones who used AI to do more, and priced it accordingly.
What should an AI-enabled agency retainer include?
A serious retainer should include strategic direction, execution, quality control, monitoring, reporting, competitor awareness, test cycles, exception handling, and a clear owner for decisions. AI makes that operating layer cheaper to run. It does not remove the need for judgment.
What should buyers ask before accepting a lower AI agency fee?
Ask what scope has disappeared, who reviews the work, what gets monitored, how exceptions are handled, how performance is attributed, and what the agency is accountable for after delivery. A lower fee that removes the control layer is not a saving. It is risk in a cheaper box.