What Happens When Clients Benchmark Your Work Against AI
A marketing manager can now put the brief into a model before the agency walks in, and arrive holding something. That benchmark changes the room whether or not anyone mentions it.
The comparison is not fair. What the client generated took four minutes and no context, and what the agency brings took three weeks and a category audit. But fairness is not the point. The client now has a reference that did not exist before, and it is anchored at zero cost.
Why The Benchmark Is Hard To Argue With
Anchoring works on the person doing it, not on the person being judged. Once someone has seen a serviceable version of the idea for nothing, every subsequent version is being priced against that, silently, whatever they say aloud.
Worse, the generated version is usually competent. Not good, but not embarrassing, and competent is exactly the level at which a busy client stops being able to tell the difference without help.
The problem is not that AI produced something better. It is that it produced something adequate before the meeting started.
The Answer That Does Not Work
Pretending it did not happen. An agency that presents as though the client walked in empty-handed is arguing against a comparison it has refused to look at, and the client makes it silently anyway.
The Only Answer That Works
Ask to see it. An agency that opens by requesting the client's generated version, then walks through what it assumed, what it missed and why the category makes those assumptions wrong, has turned a silent benchmark into a shared diagnosis.
That is a harder meeting to run and it requires actually knowing the category better than a general model does. Agencies that do will be fine. Agencies whose advantage was production capacity will find the benchmark keeps getting closer, and there is no version of this where waiting improves the position. Related: When Clients Can Generate Ideas Themselves, What Is The Agency's Real Value?
Advertisement
