The pitch that landed last month. You were cheaper. Not sharper, not better positioned, not a clearer fit. Cheaper.
The client who came back to renew and opened with: "Given how much AI is doing now, we'd like to revisit the rate."
The brief that arrived lighter than it used to. Less context, less background. The assumption sitting underneath it: your tools can fill the gap, so why send us the full thing?
None of this is catastrophic on its own, but there is a pattern underneath it, and the pattern is worth naming clearly before it becomes the new normal.
Cheap production does not lower your value. It exposes whether your value was ever clearly stated.
Most agencies have never had to answer that question precisely. Scarcity of output did the work for them. When the same output took days of senior time and a production schedule, the cost was visible, the effort was legible, and the price felt proportionate. When the same output takes one person and one afternoon, the client's mental model of what they are paying for starts to shift. Not always, and not all at once. But directionally.
The question underneath value compression is not "are we using AI well." It is more exposed than that.
If production is getting cheaper everywhere, what are clients still paying us for?
Here is what the data actually shows, because this is where the fear often outruns the facts.
27% of agencies have been asked by clients to cut rates because of AI. 47% expect the request but have not had it yet. Only 13% have actually lowered their prices. Six months on from the first survey, the rate of discount requests had not increased. The predicted wave had stalled.
That is not a reason to relax. It is a reason to be precise.
The agencies holding rate are not holding because their clients forgot to ask. They are holding because they can name and show what the client is paying for that the tool does not provide. The divide between them and the agencies stuck on price is not about tools or speed. It is about clarity.
(Source: Productive, Agencies in the AI Era: Between Hype and Reality, September 2025, 181 agencies; and Agencies in the AI Era 2.0, March 2026, 174 agencies.)
I ran two agencies simultaneously in South Africa for nearly fifteen years. When a major client froze payments during an internal investigation, both were caught in the blast radius. One closed. One held.
Same crisis. Same market. Same leadership.
What I did not fully understand at the time was why the agency that held up did. The governance and process structures had been put in place because pharmaceutical clients required them. I had not designed them for resilience. I built them to keep clients. The documented approval chains, the financial controls, the clear lines of accountability. All of it had been installed to satisfy external requirements. None of it was strategic. All of it turned out to matter when the pressure came.
The other agency had far less of that. The hustle reflex had kept attention on delivery and growth. The structural work felt unnecessary when things were going well. It became critical when they were not. By then it was too late to build it.
Took me two agencies to learn that. I did not earn the lesson through foresight. I arrived at it accidentally, and understood it only afterwards.
The agencies that are holding rate now are not doing so by accident either. They can point to where judgement, context, and trust carry the work. Not as a positioning slide. In the actual work, at the moment it matters.
Strategy shaped by a client relationship the tool does not have access to. Copy reviewed by someone who knows what the legal team will flag before it goes anywhere near them. Creative direction that reflects two years of understanding what this brand will not do, and why. The AI is in the room. It is not the reason the client trusts you. The agencies who survive can show that distinction. The ones who cannot will end up arguing on price, repeatedly, until the argument is no longer winnable.
Here is the practical problem.
Most agency founders cannot currently answer the value question with any precision. The value is there. It has never been located. It lives in the heads of senior people, in institutional memory, in the way the work gets done. It has never been mapped against where AI is now touching the process.
You cannot defend value you cannot locate.
The first move is not a positioning exercise. It is not a document about what makes you different from the agency down the road. It is a clear-eyed look at where AI is actually touching the work right now, and where human judgement is still carrying it. Not as policy. As operating fact.
Once you can see it, you can say it. To a client who asks. To a procurement team that wants to understand the review process. To a new hire who needs to know what the standard is.
That is what the AI Workflow Clarity Audit produces. Two weeks. A map of where AI touches your work, where the workflow and data and review boundaries are unclear, and what needs tightening first. From £500.
If this is the question your agency is quietly sitting with, the audit is the door I would point you at.
Craft with Command goes out weekly. If you are building a UK agency and want to think more clearly about how AI fits the work, you are in the right place. Subscribe.