AI Won't Kill UX in Finance — It'll Expose the Designers Who Were Never Strategic
Everyone's asking whether AI will replace designers. Wrong question. The real one is quieter and more uncomfortable: what were you actually being paid for?
Strahil Hadzhiev
AUTHOR
The request that should have embarrassed our whole industry for twenty years
"Can you make it look nice?"
I've heard that sentence, in one form or another, for two decades. Make it pretty. Polish it up. Put a nice skin on the thing we already built. The product decisions were made in a room I wasn't invited to, the engineering was underway, and I was handed the last stage — the packaging — and asked to make it look like somebody cared.
For most of my career, that request was the enemy. Not because it was rude. Because it was built on a lie so common that people mistook it for common sense: that design is a decorative layer you apply after the real thinking is done.
Here's what took me years to be able to say out loud. Early on, I half-believed it too. I thought more features meant a better product, and that my job was to arrange them attractively. Then I actually learned what UX is, and it turned my thinking a hundred and eighty degrees. Users don't exist to adapt to your system. Your system exists to adapt to them — to their fears, their moments, the exact place in a flow where trust either forms or breaks. The screen was never the point. The screen was just where the thinking became visible.
I spent twenty years making that argument to financial companies. Slowly, it landed. UX clawed its way from "the make-it-pretty stage" up to a seat where it could actually reduce risk and change outcomes. We more or less won the argument.
And now AI has walked in and handed the old lie its megaphone back.
Because suddenly anyone can generate a beautiful interface in thirty minutes. And a lot of executives are looking at that and quietly concluding what they always suspected: that design was decoration all along, and now the machine can do the decorating for free. That conclusion is wrong. But it's about to be a very expensive kind of wrong, and the designers who spent their careers only decorating are about to find out why.
AI is a mirror. A lot of portfolios are about to fail the reflection test.
Let me tell you what actually terrifies me about AI in our field, because it isn't the thing people expect.
It's not that AI produces bad work. It's that AI produces gorgeous work, instantly, and gorgeous work is the easiest thing in the world to mistake for good work.
Watch what happens in a modern product team. Someone types a prompt and thirty minutes later there's a polished prototype on the screen. It hits every requirement on the brief. It looks like a finished product. The room feels a rush of momentum — look what we built before lunch. And there is exactly one problem, invisible under all that polish: it was designed for a customer who doesn't exist.
Nobody asked who it's really for. What behavior it's supposed to change. What risk it removes. Which assumption was ever tested. Whether a single real human would trust it with their money. The prototype answered none of those questions. It just made it easy to stop asking them.
That's the trap. Humans read visible progress as real progress. A generated interface feels like an achievement, and that feeling is a lullaby. In finance especially, a fast-but-wrong product isn't a small mistake — it's millions in lost revenue and a hit to trust you don't easily win back.
Because the real constraints in building financial products were never the screens. I could always draw screens. The constraints were the slow, uncomfortable things: uncertainty, false assumptions, the gap between what a business believes users want and what they actually need, the org's own blind spots, the places trust quietly breaks. AI doesn't remove any of those. It just makes it easier to paper over them with something that looks done.
And here's the part that should worry every designer who built their identity on craft alone. When everyone can generate a competent banking interface in minutes, the interface stops being a differentiator. It becomes a commodity. Look at the data — the gap between the best and worst banking apps has narrowed to almost nothing; brand to brand, the experience is consistent and completely forgettable. Sameness is the default outcome of easy execution. If your entire value was producing that now-commoditized layer, AI didn't kill your job. It revealed that the market had already stopped paying for the thing you were best at.
The cruelest version of this is when companies go one step further and ask AI to do the understanding too — generate the personas, invent the pain points, simulate the users. It looks efficient. It's a closed loop of assumptions stacked on assumptions, a synthetic customer built from a guess, validated by nothing, designed for by no one who ever met a real person. It will feel strategic. It will be hollow. And you won't find out until the product ships to humans who don't behave like the ones the machine imagined.
AI is a mirror held up to our profession. It reflects, mercilessly, whether there was ever any strategy underneath your surface — or just surface.
What gets scarce when everything gets easy
Here's the reframe that changed how I think about my own next twenty years.
When execution becomes abundant, execution becomes cheap. That's just economics. The moment anyone can build anything quickly, the ability to build stops being valuable. And whatever's left over — the things the machine can't do — becomes the scarce, expensive, sought-after part.
So what's left over?
Judgment. Taste. The ability to look at ten beautiful AI-generated options and know which one is actually right for this business, this customer, this moment — and why the other nine are traps. Empathy deep enough to model how a frightened person behaves at the exact second their money is on the line. The strategic clarity to know what deserves to be built before a single pixel exists. These don't get automated, because they were never production tasks. They're the human work that production was always hiding.
This is why I'm not anxious about AI. Not because I'm precious about my craft, but because my value was never the wireframe. It was knowing which experiment to run, which friction actually matters, which trust mechanism turns a hesitant user into a confident one. AI makes me faster at all of that. It's the best junior designer I've ever worked with — tireless, quick, occasionally brilliant, and absolutely not to be trusted with a final decision. You take its output the way you'd take a talented junior's: gratefully, and skeptically, and never straight to the customer without a senior mind in between.
That's the honest test now, for anyone hiring a designer or being one. If a designer panics that AI can do their job, they were probably doing the part of the job that deserved to be automated. If a designer gets more valuable the moment execution goes free — because now their judgment isn't bottlenecked by production — that's the one worth having in the room. The machine didn't change what great design was. It just made it impossible to keep faking it.
The future of our work doesn't move down into automation. It moves up — into decision architecture, into the why, into the strategic layer that AI can accelerate but never originate. Fewer screens to polish. More systems to steer. And in an industry about to drown in infinite, competent, forgettable, machine-made interfaces, the only signal that will cut through the noise is the one thing you can't generate: a real understanding of a real human, and the judgment to design around it.
So if you're a company betting your product on how fast you can build now — everyone can build fast now. That's not your edge anymore. Your edge is knowing what's worth building, and that still takes a human who's spent enough years watching people flinch around money to tell the difference between a prototype that looks right and a product that is.
That's the work I've been doing for twenty years. AI didn't come to take it from me. It came to finally make the rest of the industry admit that this — not the polishing — was the job all along.
AI won't kill UX in finance. It'll just make sure the only UX left standing is the kind that was ever really UX to begin with.
Strahil Hadzhiev
AUTHOR
