AI Can Make You Faster. It Can't Make You Creative.
We put AI in the lead for a week.
Not as a sketch tool. Not as a moodboard generator. As the primary creative decision-maker: what colours, what composition, what angle, what copy lands.
It was confident. It was fast. It was completely, dangerously wrong about what made the work any good.
Most of what AI generates looks like what a designer would make if they'd never looked at anything real.
That's the tell. Work that doesn't stick in the memory is work that doesn't have a point of view. AI is very good at making the thing that looks like the right answer. It is not good at making the thing that feels like your answer.
We threw it out and started from the brief again.
The mistake that was actually the direction
One of our early directions had a colour combination that made everyone in the room uncomfortable. Not bad uncomfortable — the kind where you lean forward instead of leaning back.
AI flagged it as an accessibility risk and corrected it immediately.
We put it back. That discomfort was the work telling us something the brief hadn't said yet.
The thing you can't prompt
We asked AI to write a piece of copy that opened with a question, went through three beats of tension, and landed on a specific feeling.
It did exactly that. Sentence by sentence, structurally perfect. Cold.
Human copy isn't better because it's more correct. It's better because it's written by someone who has ever been unsure about something, has ever second-guessed a decision, has ever felt the specific weight of wanting to get something right.
AI has never wanted anything. It doesn't know what it feels like to make something and worry it isn't good enough. That's not a metaphor — that anxiety is part of what makes the work land.
What we kept
We use AI every day now. For research, for production, for the stuff where the answer genuinely doesn't matter and you just need to get to the next thing faster.
What we stopped letting it do: decide what the work should feel like.
Taste is the one thing that doesn't scale. You can't copy it. You can't prompt it. You can only develop it by doing the work long enough to have an opinion — and then trusting that opinion when it counts.
That's not a craft philosophy. That's a competitive advantage.
Got a project that needs to feel alive? Tell us about it.