Everyone uses AI. That is why it all looks the same.
If you have ever asked an AI for a landing page, a brand guide or a presentation, you will recognise the signature. Why it happens, where the difference lies, and what to say when a client tells you ‘we’ll just do it with AI’.
Ask an AI tool for a landing page, a brand guide or a presentation. Then ask a different tool for the same thing. You will find the same tell-tale signs:
The little label above the heading. That capitalised tag with a small dot beside it, which no designer adds out of habit.
Dashes everywhere. Long dashes scattered through every paragraph. When a text has three per paragraph, you already know who wrote it.
Tiny type across the page. Minimal snippets of text scattered everywhere, impossible to read on a phone.
Too much prose. An explanatory paragraph where a single sentence would have done.
It happened to us on this very website. The first version of this articles section had the little label above the heading and rounded boxes with a green bar. We looked at it and the reaction was immediate: ‘this looks very AI’. We took it out.
Different tools, same result
It makes sense. An AI generates the statistical average of everything it has seen. And the average, by definition, does not stand out.
When everyone publishes with the same tool, the results converge. Landing pages and presentations start to resemble one another, and none of them stays in anyone’s mind.
Where the difference lies
Our projects also use AI in their execution, every day. We use it to write code, review copy and test ideas faster.
The difference lies above the tool: years of context, data on what worked and what did not, the judgement to say ‘this is unnecessary’ and an eye for the detail nobody asks for but everybody notices. A designer’s eye above the model’s mathematics.
(AI + experience) > AIAlways.
So is AI bad? Not at all. It adds a great deal when there is someone who knows what to ask of it and what to discard.
What stands out is the same as it has always been: experience, creativity and strategy. The same differentiator as ten years ago, now delivered faster.
Of course, there are tasks where this does not apply. The more operational ones, where a designer’s or brand strategist’s eye does not move the needle. There, AI alone is enough, and that is perfectly fine.
‘We’ll just do it with AI’
Now for the provocative part. You have already heard that line, or you soon will. And there are two ways to respond:
The defensive one. Explain, justify your process and lower your price ‘just this once’.
The expert’s one. ‘Fine, let’s do both and compare.’
Notice what the second answer does. It does not argue, take offence or try to sell anything. It invites the client to see the difference for themselves, knowing what the outcome will be.
It is the answer of someone who knows they will win the comparison.
And if you are not sure you would win it… Then you have a real problem. But that problem predates AI.
How we apply it at Brova
We sum it up like this: AI gives us speed, and people make the decisions. We use Claude, Codex and Gemini, among other tools, and every deliverable goes through someone who decides what stays and what goes.
If you want to see the difference in your own project, ask us for a proposal and compare it with whatever any AI gives you. Fine, let’s do both and compare.
Ideas we share with Ezequiel Apelbaum
Much of how we think about this comes from Ezequiel Apelbaum, who writes about business development for agencies in his newsletter The Infinity Way. The ‘let’s both do it and compare’ scene and the list of AI tells come from his writing. These are other ideas of his that we fully share:
AI converges on the average. A model predicts the most probable answer, and the most probable is rarely the most brilliant. Tuning with human ratings also rewards what is balanced and penalises what is bold.
What was never written down, AI does not know. Years in a market, conversations with clients and problems solved without documentation are left out of any training data.
Execution has become almost free. If every agency uses the same tools, AI makes them equal to one another, only faster. What gains value is judgement.
Specialise. Deep knowledge of an industry or a type of problem is precisely what no dataset contains.
Charge for judgement. If AI speeds up delivery, the price stays the same and the margin improves. What you sell is knowing what to ask the tool for and what to discard.
AI as infrastructure. It sits inside the process so you can deliver more, and better. The product is still the result for the client.
His conclusion is ours too. AI exposes the agencies that only sold execution. For those that sell judgement, the best decade of their careers is beginning.
Joaquín Martinez Herrera 7 min read
