Will AI replace web designers?

Both of the usual answers are useless. The first is panic: “it is over, the profession is dead.” The second is denial: “nothing changes, AI cannot design.” Both are comforting because both end the thinking. The truth sits between them and is more interesting: the profession is not dying, but the place where it earns money is moving.

The work AI genuinely took

Let us be honest — some jobs really have gone, and they are not coming back:

• “Web design” work that amounted to installing a theme and entering content
• Building the same five-page brochure site by hand for the hundredth time
• Placing text, resizing images, repetitive layout work
• Boilerplate written from scratch: grids, button states, form validation
• Mechanical revisions of the “make this section one column on mobile” kind

Do not mourn that list. None of it was design — it was production. And production has been automated throughout history; typesetters, film editors and technical draughtsmen all lived through the same wave. The professions did not disappear; the handwork moved to the machine.

What it did not take

AI answers the question it is asked well. It does not know how to ask the right question. When a hairdresser comes to you, your actual job is not building a site: it is working out why they are not getting bookings, seeing what their competitors do, deciding which service should lead. The site is the output of that, not the thing itself.

The model does not pick up your client’s phone, does not sense the unease in a meeting, does not dig for the real want behind “that is not quite what I meant”. It does not know the budget constraint, the disagreement with the business partner, or the unwritten rules of the trade. None of this was ever the easy part — it was always the hard part. It has just become visible.

AI made production cheap, not judgement. Your value was never pushing pixels; pushing pixels just took so much of your time that it looked that way.

“It cannot design” is also wrong

The most common line on the denial side is “AI cannot design, it only copies”. Comforting, but not defensible. Today’s tools produce a reasonable layout, a typographic hierarchy, a consistent palette. “It cannot” is a sentence that becomes slightly more wrong every year — and whoever is leaning on it gets caught unprepared.

The accurate sentence is: it does design, but it does not know who for or why. The layout it produces is averagely good; it is not good for this particular business. The difference is not aesthetics, it is context. And context is created in the forty minutes you spend with the client.

The typesetters’ story

An analogy, because this film has played before. When desktop publishing arrived, typesetters lost their jobs — they genuinely did, and there is no point softening that. But graphic design did not die; it grew. Because what disappeared was setting the type, not deciding what should be set.

The ones who moved from typesetting to design won; the ones who said “I set type” lost. The difference was not talent. It was what they thought their profession was. The same split is happening in web design now, and you still get to choose your side.

Who is actually at risk?

Here is the real answer. The risk is not to “designers”, it is to designers who sell only execution. Anyone whose pitch is “tell me what you want and I will build it” is exposed — because the client can now put that question straight to a tool. Anyone whose pitch is “let us work out what you should do” is not.

That split existed before AI; it was just possible to hide it. Execution was slow, so selling execution alone was a viable business. Once execution got fast, the model collapsed. Whoever lost did not lose to AI — they were standing on a position that was already fragile.

Why does it feel so threatening?

There is a reason, and it is not technical: the most visible part of our work is the part being automated. What a designer shows a client has always been output — the screen, the layout, the colour. The judgement was invisible, so it never got discussed. Now that the visible part has become cheap, we have to name what is left, and we have never had to do that.

So the crisis is narrative rather than professional. For years we said “I build websites”, when what we did was never only that. Now we have to find the accurate sentence — which is less a threat than an overdue clarification.

What about junior designers?

This is the hardest question and the least discussed. The classic path was: start with simple work, learn by doing, move to the complex over time. But simple work is what is automating — so the first rung of the ladder is rising.

The pessimistic reading: newcomers will not find a way in. The optimistic reading: newcomers skip the production skill that used to take three years and start learning judgement directly. Both are partly true. The difference is that “I learned to code” is no longer a career. A junior who can answer who this is for, why, and what outcome it has to produce will rise faster than ever.

Practical advice: for every hour you spend learning the tool, spend an hour understanding your client’s business. The first goes stale in six months; the second lasts your whole career. And there is a consolation — in that second part AI has no advantage over you at all. What you are competing with there is not a machine, only less curious colleagues.

So what should you do?

Three moves, all of which can start today:

• Be the side using the tool, not the side resisting it. A person resisting cannot compete on price with a person using.
• Invest the freed-up time in understanding the client’s business; a designer who can show that the work worked does not end up in a price argument.
• Stop selling hours. If getting faster is making you poorer, your model is wrong, not your tool.

There is also an overlooked opportunity: AI raised the solo designer to something close to small-agency capacity. Volumes that used to need a team can now be delivered alone. That is leverage rather than threat — but only for the people who choose to use it.

Will clients skip designers entirely?

The most common fear: “the hairdresser will build their own site now, so what do they need me for?” Some will — and honestly, that client was already your least profitable. The lowest budget, the most negotiation, the most revisions. Losing them is not a disaster.

But most will not. The reason is not the tool, it is time and attention: the hairdresser’s job is cutting hair. A tool being free does not make learning it and fiddling with it free. People still eat out despite owning a dishwasher.

What changed is that “I can build a website” is no longer a sales argument. The client knows they could do it too. The reason they choose you has to be that they expect a better outcome — and explaining that is now part of the job.

Will prices fall?

This is the version of the anxiety that is actually about money, and the honest answer is: at the bottom of the market, yes. A standard five-page brochure site will keep getting cheaper, because producing it got cheaper. That is inevitable and resisting it wastes your time.

The top end is going the other way. A designer who can guide a business that does not know what it should do, and who can show that the result works, becomes more valuable — because what is scarce now is not production capacity but judgement. The market is splitting down the middle: execution gets cheaper, decisions get dearer.

So the question is not “will prices fall” but “which end are you at”. And that is a position you can change within six months.

To bring it together: no, it will not take your job. But it will change the definition of your job — and if you do not update that definition yourself, the market will do it for you. The difference is only which one hurts more.