You have probably seen the look by now.
Beautiful typography. Tasteful colours. A clever abstract mark. A mockup on textured paper. Maybe a tote bag nobody asked for.
It looks good.
And five minutes later you cannot remember which company it was for.
This problem existed before generative AI. AI has simply made it faster to produce.
The average of good-looking things is still an average
Generative systems are exceptionally good at recognising patterns. That is part of what makes them useful.
Ask for a contemporary premium skincare identity and the system has a vast visual vocabulary to draw from. It understands, in a statistical sense, what that category tends to look like.
Which is precisely the problem if your goal is distinction.
The safest answer is often the most familiar one.
AI can get you to “looks like a brand” very quickly. It cannot guarantee “could only be this brand”.
Generic starts with a generic brief
Consider these prompts:
“Design a premium identity for a modern wellness brand.”
“Write a sophisticated headline for a boutique hotel.”
“Create a minimalist website for a design studio.”
There is almost nothing in those instructions that belongs to a specific business.
So why would the output?
A more useful starting point might describe the contradiction inside the business. A medical practice that needs to feel clinically credible without feeling cold. A jewellery brand rooted in Indian craft that does not want to look traditional. A children’s clothing business that began in a cupboard at home and grew into three warehouses without losing the friendliness that got it there.
Now there is something to design around.
Context is the antidote
Give AI your actual constraints. Existing brand assets. Customer language. Product photographs. Materials. Technical limitations. Cultural context. Competitor patterns you want to avoid. Things the founder dislikes. Things the sales team keeps having to explain.
The goal is not to write a more elaborate prompt for the sake of prompt engineering.
The goal is to put reality back into the process.
Distinctiveness often comes from the awkward details
The details that make a business interesting are rarely the ones that appear in a generic creative brief.
A founder’s dog becoming the mark of a canine wellness company.
A brand name literally being constructed into its own symbol.
A company’s first stock sitting in a cupboard because the founder had noticed other new mothers could not easily find affordable branded children’s clothing.
These are not aesthetic instructions. They are stories, constraints and points of view.
A designer can decide whether they deserve to become visual ideas. AI can then help explore how.
The danger of looking too finished too early
AI has another side effect: rough ideas arrive looking convincing.
That can be dangerous.
A beautifully rendered packaging mockup can make a weak identity feel resolved. A cinematic image can distract from an ordinary campaign idea. A polished homepage can hide the fact that nobody has agreed what the company actually does.
In traditional design processes, roughness was sometimes useful. It allowed everyone to discuss the idea before becoming attached to the finish.
We need to preserve that discipline even when the tool makes everything look presentation-ready on day one.
A brand needs a reason, not just a style
Style matters. Typography matters. Colour matters. Craft matters.
But without a reason behind the choices, style becomes costume.
The question to keep asking is not “Does this look good?”
Ask: what makes this specifically true for this business?
If there is no good answer, the problem is probably not the prompt.
The problem is that we started designing before we had anything meaningful to say.