I've been watching a lot of videos recently about AI-assisted website building. You've probably seen similar ones: someone types in a prompt, an agent gets to work and a few minutes later there's a polished, responsive website sitting on the screen. Sometimes they're described as ‘£10,000 websites built in five or ten minutes’, statements designed to grab people's attention.
As someone who designs websites for a living, you might expect this to worry me. It doesn't. I think it's genuinely exciting.
The ability to create something quickly, something that would once have required considerable time and technical knowledge, opens up all sorts of possibilities. If you're experimenting, learning, testing an idea or simply trying to get something online, these tools are extraordinary. For designers and developers, AI can remove repetitive work, help explore ideas and speed up parts of production.
But the more I've thought about it, the less this feels like a question about websites. The same thing is happening across commercial design. We can generate imagery, copy, presentations, campaigns, social content and countless variations faster than ever before. The website was simply what made me stop and think about a bigger question:
When making becomes this easy, how do we decide what deserves to be made?
I'm not suggesting everything we create needs a strategic justification. There's enormous value in being able to experiment, play, learn or simply make something because you want to. Nor do I think every small business needs an extensive strategic process before it can put something online.
My questions begin when what we're creating has a commercial job to do. It might be for an established business with customers, a reputation and years of brand equity behind it, or it might be for a new business trying to establish those things properly from the start. Either way, once something becomes an important part of how that business wants to be understood, trusted and chosen, the thinking behind it begins to matter.
So yes, you can build a website in five minutes. But how did you decide what website to build?
That's something I'm increasingly asking myself when I work with AI. If I ask it to ‘create a modern professional website for a growing service business’, for example, it has more than enough knowledge to produce something visually convincing. There will probably be a strong hero section, some introductory copy, perhaps three cards explaining the services, testimonials, FAQs and a call to action. It might look great at first glance.
But why are there three cards? Why is the testimonial there? Why are there two buttons in the hero section? Why does the page need an FAQ? Why is that particular message the first thing somebody sees?
None of those elements are inherently wrong. I use similar patterns in my own work because conventions can be useful. People understand them, they feel familiar and they make websites easier to navigate. The problem isn't the pattern itself. It's when we stop asking why the pattern is there.
And that's not exclusively an AI problem. Designers can follow trends without questioning them too. We've all seen work from completely different businesses that somehow feels remarkably similar once you strip away the colours, photography and words. AI hasn't invented arbitrary design decisions. It simply gives us the ability to make them much more quickly.
For a long time, a significant part of the challenge was simply making the thing. If a business wanted another landing page, somebody had to brief it, design it, build it, test it and publish it. The effort involved naturally created a human sense-check where somebody was likely to ask whether the page was actually necessary.
Now we can potentially say, ‘Make me another landing page’, and have one moments later. That's brilliant, but that natural sense-check can start to disappear.
If making another page, image, headline, social post, presentation or campaign variation becomes almost effortless, we can quickly move from having too few options to having too many. AI can give me ten headline ideas in seconds, then another ten, then another twenty. Soon I'm not suffering from a lack of ideas. I have the opposite problem: I need to decide which one is right.
It's a little like opening Netflix and being offered thousands of things to watch, then spending half an hour scrolling because you can't decide what you actually want. More choice doesn't automatically make the decision easier.
And it's not only my ability to create that's increasing. Everyone else has access to the same tools. If businesses can suddenly produce more pages, articles, emails, imagery, video, presentations and social content, the amount of communication competing for our attention increases dramatically. Our ability to produce might be growing, but our customers haven't suddenly gained more time or attention to give us.
When everyone can make more noise, the answer isn't necessarily to shout louder. It's to make the signal clearer.
That's where restraint matters. Good design has never only been about what you add. A significant part of it is deciding what to remove. You might begin with three calls to action and realise there's really only one thing you want somebody to do, or plan eight sections for a webpage and realise five tell the story more clearly. Removing something isn't losing work. Often, arriving at the clearest version takes more thought than simply adding more.
Everything should earn its place.
In commercial design, that isn't restraint for the sake of minimalism. Everything we create communicates something about the business. A website, brochure, campaign, presentation, social post or piece of imagery might be doing a different job, but each contributes to how that business is understood. The question isn't simply whether we can make it. It's whether it serves the brand, helps the customer and supports what the business is trying to achieve.
Every decision sends a signal: the words, typography, colour, imagery, hierarchy, tone of voice and even what we choose not to say. Consistency doesn't mean everything a business creates has to look the same. It means the decisions should feel as though they come from the same understanding of the business.
Who are we? Who are we talking to? What matters to them? What are we trying to communicate? What do we want somebody to understand, feel or do?
Those questions give us something against which to judge all the possibilities AI can create. If AI is given very little context, it has to fill in the gaps somehow, and familiar patterns are a perfectly reasonable place to start. The more useful question is what happens next. Does somebody understand the business well enough to recognise what fits and what doesn't?
When I use AI in my own work, I rarely accept the first thing it gives me. I'll question a headline because it feels too corporate, remove something because we're repeating ourselves, reject an image because it doesn't feel right for the brand, or change the hierarchy because something relatively unimportant has been given too much prominence. Sometimes the output is technically correct and visually polished, but it still doesn't feel right for the business I'm working with.
That judgement comes from understanding the business first: what it does, why its customers choose it, what matters to the people inside it, what its customers need to understand and how the brand should make them feel. Without that context, it's difficult to decide whether something belongs.
AI can produce options. Someone still needs to know what belongs, and why.
And that responsibility doesn't end once something has been created. There's another part of the five-minute website showreel that doesn't tend to make the cut: what happens six months later.
Somebody wants to add a case study. A new employee joins. A service changes. Marketing needs a landing page for a campaign. Someone wants to replace an image, publish an article or create a presentation. At that point, the question is no longer simply whether the original design works. It's whether the thinking behind it can survive as the business and its communication evolve.
For me, that's part of brand stewardship. If the thinking behind a brand is sound, it should help guide what happens next, whether that's a new page, campaign, presentation, piece of content or something that hasn't even been thought of yet.
A business needs the freedom to move, respond and create without calling a designer every time it needs a new piece of communication. But complete freedom isn't necessarily the same as empowerment. Sometimes the better system is one that provides carefully considered building blocks and an understanding of why they're there. Enough flexibility for people to do their jobs, with enough structure to protect the thinking, personality and recognition that make the brand distinctive.
A brand shouldn't only work while the person who created it is in the room.
A website makes that particularly tangible. Where does the content live? What can people safely change? Who owns the hosting account? Where are the source files and assets? What external services does it depend on? If something breaks, who fixes it? If the person who created it disappears, can somebody else realistically take over?
None of those questions are arguments against AI-assisted development. They're simply part of designing something for long-term commercial use rather than for a five-minute demonstration.
I don't think the answer to faster technology is to deliberately work more slowly. That would miss the opportunity. I want to use the speed. I want AI to remove repetitive work, help me explore possibilities and give businesses access to things that might previously have been beyond their reach.
But when what we're creating has a commercial purpose, being able to make it isn't enough.
What should exist? What doesn't need to? What is the business trying to achieve? Does it belong to the brand? Does it help the person it's intended for? What do we want them to understand, feel or do? And once we've made it, how do we make sure the thinking behind it survives?
AI can give us more possibilities than we've ever had before. I don't think the challenge ahead is going to be whether we can make enough. It might be whether we can understand, question, edit and choose well enough to know what deserves to become real.
That's why I'm excited about where this technology is taking us. Not because it means we can make everything faster, but because perhaps it gives us an opportunity to think more carefully about what is worth making in the first place.
Because the website was never really the point.
The thinking behind it was.
The faster we can make things, the more important it becomes to understand why we're making them.
