AI Can Build You A Website Now. That's Exactly The Problem.

Smiling man with curly hair and tattooed arm sitting at a table with a laptop, coffee cup, and artistic framed pictures on a gray wall behind him.
Tom Declat
July 31, 2026

Anyone can have a website by tonight. All you need to do is type a prompt, pick a template and AI fills in the rest. It'll be live in an hour.

I've been doing this for 15 years and it still blows my mind, years ago that would have been WEEKS of work.

But what I think... that might be a bit of a shocker, I think that's genuinely good news. For a lot of people.

The Genuinely Good News

If you're starting out, you know, testing an idea or running a side project with no revenue yet or something, AI-built websites are absolutely brilliant.

They're super cheap, very fast and easy enough to use... plus good enough to look legitimate while you figure out if the idea even works.

I'd never tell someone bootstrapping their first business to spend five figures on a website. That's a terrible use of money at that stage! We even have a product that's specifically for start-ups but unsurprisingly that's bought less and less these days.

Where The Conversation Changes

But somewhere along the way (and actually something I see fairly often on LinkedIn) is the distorted message,"AI can build websites now, so agencies are dead." But that's not what's actually happening...

What I Keep Seeing With Established Businesses

Here's the pattern with the (and I mean this with all respect) the more serious, more established businesses we tend to work with. They tried the fast route years ago, back then it was a templated builder with Squarespace or something, which was fine for launch.

One consultancy I spoke with had their AI-built site live for four months before anyone on the team realised the enquiry form had never actually been connected to an inbox... which is a bit of a howler I'd say. But nobody had checked, because nobody could tell if it was working, the site just looked finished.

That's the trap to be aware of. Three months in, nobody could edit it without breaking something and they sounded like every competitor's copy (spoiler: because it was trained on the same data everyone else's was). And worst of all, it wasn't built around how their actual buyers think and decide, it was built around what they DID. Again, not ideal.

This Is A Clarify Problem

We call this a Clarify problem, the site exists, it's "live" and it technically works. But it isn't doing the packaging and positioning job your business actually needs from it. Again, this isn't a design problem, it's a misunderstanding of WHY websites are needed in the first place.

Five Questions To Ask Before You Commit

Before any established business goes down the AI-built route, here's what I'd tell them to check first:

  1. Who's actually going to update this in six months, and can they?
  2. Does the copy sound like your business, or like a Wikipedia summary of your industry?
  3. Is there an actual path from visitor to enquiry, or just a page that looks nice?
  4. What happens when you need something the template can't do?
  5. Who do you call when it breaks?

If you can answer all five without hesitating, you're probably fine, crack on. If two or three make you pause, THEN that's worth sitting with before you launch.

"You Would Say That"

Obviously, I get the pushback already: "You would say that, you sell websites."

Which you know, is totally fair. But this isn't really about AI versus agencies, it's about matching the tool to the stakes.

A £2,000-a-month consultancy relying on referrals can't really afford a website that stops potential buyers because the copy feels generic and nobody can touch it without a developer. A side project testing an idea can ABSOLUTELY afford that risk, there's nothing to lose yet.

We have clients that their target market need to have 6-figure bank balances before they work with them. They unsurprisingly, do not have an AI generated website. When you NEED to get this right, when it's the face of your business and usually, the first touch point you client base will have with you, it's worth thinking about what you want that experience to be like.

The Real Question

It's not "AI bad, man must smash". Instead, just know which business you're running.

If you're small, scrappy, and just need to exist online, AI-built is a really smart move. I'd recommend that every time, use it and don't overthink it.

If you're an established business where your website is doing things like bringing in leads, building trust, carrying your reputation... it's worth treating it like infrastructure, not a template.

Which one are you? The one testing an idea, or the one with something to protect?

I turned the five questions above into a one-page checklist "Is Your Website Infrastructure Or A Liability?" Grab it here, and run it against your own site in five minutes.

If you're not sure which camp you're in either way, email me and tell me what your site's actually meant to be doing for you. Happy to give you a straight answer.

Subscribe to receive weekly tips, advice and knowledge on growing your business online

Before you commit to anything

Not sure where to start? The 4C Growth Diagnostic takes eight minutes.

Twenty questions. A scored breakdown across all four pillars, Clarify, Craft, Convert, Continue. You will finish with a clear view of which part of your business is holding the next stage back, a priority for where to focus first, and a free resource that addresses your specific gap directly.

There is no pitch at the end. If the output is useful, you have something to act on today. If it is not, you have lost eight minutes.