If Your Design Can Be Replaced by a Prompt, It Deserved To Be
Every week another designer posts the same panicked thread.
AI is coming for our jobs. Figma Make is a threat. Claude can build a landing page now. Save the craft. Protect the profession. Sign this petition to feel something.
I design and build websites for a living, so by the rules of the internet I’m supposed to be scared too. I’m supposed to circle the wagons and post about how a machine can never replicate the soul of real design.
I’m not going to do that. Because most of what’s getting replaced right now should be replaced. And the sooner designers admit that, the sooner we stop being scared.
Here’s the uncomfortable version: if a prompt can replace your design, your design was never the valuable part. The prompt didn’t lower your worth. It exposed it.
The work AI is eating was never worth much
Let’s be honest about what these tools actually do well.
They generate a clean hero section. A pricing table with three tiers and a popular badge. A features grid with icons. A testimonial slider. A footer with the same six links everyone has. A gradient. A little scroll animation to make it feel alive.
And they do it in about four minutes.
Now ask yourself the question no designer wants to ask out loud: how much of your last project was exactly that?
Because if the answer is “most of it,” the AI isn’t threatening your job. It’s doing an honest audit of it. For years a huge chunk of design work has been assembling the same known patterns into a slightly different arrangement and charging for the hours. We called it craft. A lot of it was just muscle memory with a moodboard.
The machine learned the muscle memory. Of course it did. It was the easiest part to learn, because it was the least thoughtful part of the job. We just never had to admit that while we were the only ones who could do it.
We spent a decade optimising for the exact thing AI is best at
This is the part that stings, and it’s our own fault.
For ten years the design industry rewarded surface. Dribbble shots that looked incredible and explained nothing. Awwwards sites that won Site of the Day and generated zero leads. Portfolios full of screens that had never survived contact with a real user. We trained an entire generation of designers to chase what looks good in a grid of thumbnails, not what works in someone’s hands at 11pm on a bad connection.
We optimized for aesthetics divorced from outcomes. We made “looks designed” the whole game.
Then we handed a machine ten years of that exact output and acted shocked when it got good at producing it.
AI didn’t kill design. It ate the specific, hollow version of design we spent a decade celebrating. It’s very good at making things that look designed, because that’s precisely what we taught the internet to value. The bill for chasing likes over outcomes finally came due, and the invoice is addressed to us.
What a prompt still can’t do (and it’s the whole job)
Here’s where the scared designers and I actually agree, they just bury the lede.
A prompt can generate a beautiful landing page. It cannot tell you that your entire homepage is answering the wrong question. It can’t sit in a call with a founder who can’t articulate what their product does, and pull the actual value out of them by asking the third and fourth uncomfortable question. It can’t decide that the pricing page should come before the features because this specific company sells to CFOs who care about cost before capability. It can’t look at a flow and feel that something is off two screens before the drop-off happens.
It can’t make the ugly, correct call. It can’t kill the beautiful section that’s hurting conversions. It has no taste, only an average of everyone else’s.
That gap, between “looks designed” and “makes a stranger understand and act”, is not a small part of the job. It is the job. It always was. AI just made the first half free and left the hard half sitting exactly where it always was, waiting for someone with judgment to show up.
The designers panicking are the ones who lived entirely in the first half. The designers who are calm are the ones who were already doing the second.
This is going to be brutal for a specific group, and pretending otherwise is cruel
I’m not going to sugarcoat this into a LinkedIn-friendly “everything will be fine, just upskill.” It won’t be fine for everyone.
If your value was speed and volume, producing clean, competent, unremarkable screens faster than the next person, that value is gone. Not shrinking. Gone. A tool does it now for the price of a subscription, at 3am, without complaining.
If you’re a junior who was planning to spend three years doing exactly that kind of production work to earn your stripes, the ladder you were going to climb is missing its bottom rungs. That’s a real problem and I don’t have a tidy answer for it. The path that turned juniors into seniors ran straight through the work AI now does for free.
But here’s the thing I keep coming back to: none of that is the AI’s fault, and blaming it is a waste of the anger. The tool didn’t decide your work was commoditizable. The market did, years ago, and the tool just enforced the verdict. We had a decade of warning signs, cheaper and cheaper template marketplaces, Fiverr gigs, “just use a Framer template”, and we called all of it a threat to real design while quietly doing work that wasn’t much better than the template.
Stop defending the craft. Start being worth the money.
So here’s my actual argument, and it’s not a doom post.
Stop trying to protect design from AI. You can’t, and the parts you’re trying to protect mostly don’t deserve protecting. Every hour spent writing “a machine can’t replicate human creativity” is an hour not spent becoming the kind of designer a machine genuinely can’t replace.
Become worth the money. Get so good at the judgment, the strategy, the uncomfortable questions, the taste, the outcome, that no prompt gets within a mile of it. Let the machine have the hero sections and the pricing grids. Take the part that was always the actual value and lean into it so hard that clients would feel stupid handing it to a tool.
The designers who survive this aren’t the ones who fight the prompt. They’re the ones who were never doing the prompt’s job in the first place.
If your design can be replaced by a prompt, it deserved to be. That’s not a threat. It’s the most useful thing anyone will tell you this year. The only question that matters now is whether the part of your work that can’t be replaced is big enough to build a career on.
For a lot of designers, honestly, it isn’t yet.
Make it bigger.
Hey, I’m Vishal, a product designer, vibe coder, and someone who’s been deep in the trenches with tools like Webflow and Framer for years. I run OhhMyDesign, a small studio where we help brands go from messy ideas to clean, high-performing websites that actually do the job.
Most of what I write here comes straight from real client work, late-night rebuilds, and way too many “quick fixes” that turned into full-blown redesigns. If you’re into design that looks good but also makes sense, stick around.
