Yes.
Today you can ask an AI to help you plan a site structure, design an interface, write the copy, code it and check part of the result.
For some projects, that may be all you need.
The interesting question is no longer whether AI can build a website.
It's what information it needs to build the right website, and how you'll know whether what it produced is any good.
AI can do far more than write code
Thinking of it as "a tool that codes fast" doesn't really hold up anymore.
It can help you:
- organise your services;
- propose a site structure;
- research how people search;
- try out different messages;
- create designs;
- write and review code;
- spot problems;
- compare options;
- set up tests before you publish.
We use it in our work too.
But there's a difference between being able to produce an answer and having enough context for that answer to fit a real business.
The problem is usually context
Say you ask it:
"Build me a website for a nursery."
It can produce a perfectly competent site.
But there are still questions left:
Who actually makes the decision?
What worries those people?
What makes this nursery different?
What do you not want to come across as?
How do families find you today?
What do they need to know before getting in touch?
What can you claim, and what can't you?
AI can help you answer them.
What it can't do is know those answers if nobody gives them to it or goes and finds them out.
And if it starts from a wrong assumption, it can build something very convincing around it.
El Niu is a good example
When we worked on the website for El Niu, one of the important decisions came from understanding a fairly simple tension:
the space is designed for children, but it's their parents who decide.
If the site became too childlike, it could lose credibility with the people making the decision.
If it became too corporate, it stopped feeling like the place itself.
An AI could have helped design either direction.
What mattered was spotting first that the tension existed.
From there, the design, photography, content and structure all followed the same thinking.
When I'd just try building it with AI
If the cost of getting it wrong is low, I'd give it a go without overthinking it.
For example:
- a landing page to test an idea;
- a personal website;
- a page for an event;
- a simple portfolio;
- a first version of a project;
- an internal tool;
- a small site you know how to check yourself.
You can get a long way before you need outside help.
Even further if you have some technical or design experience.
When I'd check what it produces much more carefully
Things change when the website carries more weight.
For example, if it:
- is an important source of customers;
- has lots of pages and searches to work on;
- takes payments;
- stores data;
- has user accounts or private areas;
- connects to other tools;
- automates part of the business;
- or a mistake could cost you money or information.
The problem here isn't "using AI".
I'd use it too.
The problem is publishing something important without anyone who knows how to properly check what's been built.
The part that's easiest to underestimate: checking
Generating a first version can be incredibly fast.
Checking it properly takes more work.
You need to look at things like:
- what happens on mobile;
- whether form submissions actually arrive;
- what happens when someone does something unexpected;
- whether Google can understand the important pages;
- whether analytics is measuring what it should;
- whether the copy says things you can actually stand behind;
- what happens to sensitive data;
- what happens when an integration fails.
The simpler the website, the shorter this list.
The more the website does, the more it matters.
So do you need to pay someone to build it?
Not necessarily.
If you know what you need, can give the AI good context and know how to check the result, you can do a lot on your own.
And if you're still testing an idea, it probably makes sense to start that way before investing more.
Professional help starts to add value when the problem stops being "making pages" and becomes deciding:
- what to build;
- how to structure it;
- what's worth doing now;
- what can wait;
- and how to check that everything works as it should.
AI can be part of that whole process.
In fact, it should be.
But using AI and having good judgement about what you make with it are two different things.
If what you need is a website where those decisions are already part of the project, you can see how we approach a Brand Website.
We write about decisions that come up in real projects: how to structure a website, how to capture demand and when technology can simplify a business.