
Brand lives at the edges, not the average.
AI, where’s the personality? Using AI design tools lately has reminded me of something from my early days at college. We were taught the fundamentals first. Grids, spacing, alignment, structure. The discipline of design. How to create something clear, simple, and well-balanced. And then, at some point, the message flipped. The real work wasn’t in following the rules. It was in knowing when - and how - to break them.
That idea has stuck with me throughout my career. Because that’s where design stops being correct… and starts becoming interesting. What I’m seeing in AI-generated design right now feels very familiar.
It’s logical. Structured. Based on best practice. Easy to defend. In many ways, it’s exactly what we were all trained to do in the early stages of learning design. And at first glance, it looks good. But when you spend a bit more time with it - when you actually start to critique it properly - it begins to feel a bit off. Not in a way that’s immediately obvious. Much more subtle than that.
It works. It’s usable. It makes sense. But it’s also a bit flat. A bit safe. Slightly forgettable. That’s the part I keep coming back to. What’s missing isn’t usability. It’s personality.
Good design gets you clarity. It helps people move through a journey without friction. It removes confusion and makes things accessible. But once you’ve achieved that - and increasingly, AI can get you there quite quickly - you start to look for something more. You want to feel something about the product. You want a sense of character. A tone. A point of view. And that’s where things start to diverge.
This is also where brand really comes into play. Not in the obvious sense - logos, colours, marketing campaigns - but in the quieter, more subtle decisions that shape how a product feels. It shows up in how bold or restrained something is. In the way content is written. In whether an interaction feels efficient or slightly playful. In how far you’re willing to push something beyond what’s expected.
Two products can follow the same patterns, use the same structure, even solve the same problem - and still feel completely different. That difference is brand. It’s what makes something recognisable. What makes it stick. What makes someone choose it, and come back to it. It’s how you make your user’s feel. And it’s also the part that’s hardest to systemise. AI, by its nature, leans in the opposite direction. It pulls from what already exists. It aggregates what works, what converts, what aligns with best practice, and gives you a clean version of that. It gets you to something solid, quickly. But it also pulls everything toward the middle. Because brand isn’t the average. It lives in the edges. In the small, deliberate decisions to push something further, or to hold something back. In choices that don’t always look “optimal” on paper, but feel right in context.
That kind of judgement doesn’t come from patterns alone. It comes from understanding what a product is trying to be, who it’s for, and how it wants to show up in the world.
I saw this play out quite clearly the other day. A product manager had used AI tools to generate a full journey. They’d done the right things - used the design system, structured the flow well, created something that looked coherent. At a glance, it was compelling.
But when we started to dig into it, the cracks appeared. It hadn’t accounted for the wider system. It made assumptions about data and components that didn’t quite hold up. Small inconsistencies started to appear, and fixing them created new ones elsewhere.
We spent time working through it together, adjusting and reshaping. In the end, it wasn’t faster. It didn’t lead to a better outcome. It just got us to something that looked ok, but still needing conversations to validate assumptions. In essence we had to go back a step in the process to get the context.
This is where I think the role of the designer is becoming clearer. If AI can get us to a solid baseline - something usable, structured, and aligned to best practice - then the value shifts. It’s no longer just about producing the output.
It’s about knowing whether that output is actually any good. Understanding what’s missing. Bringing in the nuance, the context, the sense of what this product should feel like. It’s about making decisions that go beyond correctness.
And this isn’t just a design problem.
I saw a job listing recently where the entire application process was handled by AI. You use AI to craft your CV, AI reviews it, AI runs the initial conversation. At every stage, you’re shaping your response to match the system. Presenting the right signals, the right structure, the right information. But somewhere along the way, something gets lost.
Where’s the personality? The individuality? The slightly rough edges that tell you what someone is actually like to work with?
It feels like the same pattern. We’re getting very good at producing things that are correct. Optimised. Aligned. But we risk losing the parts that make them human.
AI is incredibly good at getting us to “good”. And “good” is quickly becoming the baseline.
But the things that really matter - the things that make a product stand out, that people remember, that create any kind of connection - don’t come from best practice alone. They come from opinion. From intent. From small, deliberate decisions that shape how something feels.
They come from people.
And if we’re not careful, that’s the bit we slowly optimise away.
