AI Did Not Create Your Sameness Problem, It Exposed It
· Updated
- AI did not cause sameness, it removed the cost that used to hide it.
- Making more content does not fix sameness, because differentiation is a judgment problem.
- Put the brand standard in the hands of whoever holds the prompt, before anything is generated.
- In a category that already looks identical, a held standard is the cheapest edge there is.
The world's most influential CMOs met in Cannes in June, and Forbes' account of the event described the condition everyone in a marketing seat is feeling: a cutthroat market saturated by AI, and the work of piercing through the noise. I agree with the diagnosis. Standing out is harder than it was. Undifferentiated output does not just fail to land. It wears down equity a brand spent years building. But the sameness is older than the tool.
I do not work in a category that just discovered sameness. Multifamily marketing looked interchangeable long before anyone typed a prompt. Same amber sunset render, same smiling stock couple at the same kitchen island, same three adjectives about elevated living. That sea of identical was not made by a machine. It was made by people, on deadline, choosing the safe version every time. AI did not invent the sameness problem. It removed the last thing that used to hide it.
Did AI actually make everything look the same?
No. AI made looking the same nearly free, which is a different claim. Generic work always existed, it just used to take budget, time, and a team, so there was less of it. Now anyone can produce average work in bulk in an afternoon. The flood is real. The tool scaled a choice people were already making.
That distinction matters because it changes the fix. If AI caused the problem, you fight it with a better tool or a ban. If AI only exposed the problem, no tool saves you, because the gap was never technical. The gap was that many brands never decided what made them different in a way anyone could act on. When production was expensive, that gap stayed hidden behind the sheer effort of shipping anything. Cheap generation pulled the cover off.
Why doesn't making more content help a brand stand out?
Brands that fail to stand out rarely fail for lack of output. They fail because the work they make looks like everyone else's, and making more of it does not change that. Differentiation is a series of specific choices to not do the obvious thing, and each of those choices is somebody's judgment.
Watch what happens when a team responds to saturation by generating harder. More posts, more variants, more channels. Most of that work drifts toward the same average, because whoever typed the prompt brought no point of view to bend the output away from center. You cannot out-produce sameness. You can only out-decide it, one deliberate call at a time, and a generator has no opinions to lend you.
How does a brand stand out when every competitor looks the same?
By holding a standard most competitors will not. In a category where every property reaches for the same render and the same stock smile, the brand that shows a real building honestly, with restraint and a consistent hand, reads as the different one. The bar in this category is low precisely because everyone chose safe, which makes clearing it cheaper than it sounds.
I spent most of my career building and leading a creative team, and I now lead brand in rooms built around capital, construction, and operations. What I have learned across both is that differentiation does not come from the creative department having better taste in isolation. It comes from a brand willing to make a slightly harder choice, repeatedly, in public, when the safe version was right there and free. That willingness is the whole edge.
How do you keep AI from eroding the brand instead of building it?
Move the brand standard upstream of the tool. A review queue built for a handful of assets a week cannot inspect a full day of generated output, so governance at the end fails by arithmetic. The control that scales is a standard applied at the point of creation, by the person holding the prompt.
That means doing the unglamorous work many brands skipped:
- Show on-brand and off-brand side by side, with the reasoning attached, so the standard survives contact with a deadline.
- Put judgment before generation. The question is never "can the model make this," it is "should this exist and does it sound like us."
- Make the standard a shared object, so a marketer, a model, and a new hire all reach for the same bar.
The tool takes whatever it is pointed at and makes more of it. Point it at a brand with a spine and it multiplies a point of view. Point it at a brand that never had one and it multiplies the sameness faster than you can review it. Standing out is harder now. It is harder because only one thing is left doing the work, and it is the thing that always separated distinctive brands from average ones: judgment held under pressure. The tools are the same for everyone. The difference is what a brand is willing to decide, and how often it is willing to decide it again.
Frequently asked
Does more AI content help a brand stand out?
Usually the opposite. When generation is cheap, more output means more average output, and average is exactly what an audience already ignores. Volume was rarely the constraint on differentiation. Judgment was. AI removes the volume constraint and leaves the judgment one fully exposed, which is why saturated categories are getting harder to break, not easier.
Is brand sameness an AI problem or an older problem?
It is much older. Whole categories looked interchangeable long before generative tools existed, because sameness is safe and safe survives review. AI did not introduce the instinct to blend in, it just made blending in free and instant. The tool exposed a discipline gap that was always there and used to be masked by the cost of producing anything at all.
How do you keep AI output from eroding brand equity?
Move the brand standard upstream of the generator. Decide what on-brand and off-brand actually look like, write it down with real examples and the reasoning attached, and share it with everyone who prompts, so judgment is applied at the point of creation instead of in a review queue at the end. A queue cannot keep pace with what a team can now generate in a day. A shared standard, applied before anything is made, can.