If you were building a business website five or ten years ago, the process was fairly predictable.
Choose a web designer. Decide how you want the site to look. Write some content. Add your services and contact details. Then, toward the end of the project, someone would probably mention SEO.
That approach was never really all that good for search performance and by 2026 it really does seem terribly out of date.
The way people go about finding businesses online is changing. Google is still a top resource, but people are increasingly using ChatGPT, Gemini, Perplexity and the various other AI tools to ask their questions. Google is itself increasingly providing answers to searches via AI Overviews and AI Mode.
For businesses looking at a new website there’s another question to consider in addition to design, usability and the usual SEO questions:
Will this site provide search engines and AI systems with enough useful information to work out who we are, what we do and why they should trust us?
That’s now become a fundamental part of modern web development.
Search isn’t disappearing, but it is changing
Over the last couple of years, people have tended to present AI search as if Google is going to vanish overnight.
That isn’t particularly helpful.
People are still Googling. People are still visiting websites. They’re still looking at reviews, they’re still comparing companies and they still want some reassurance before they spend their money.
What’s changing is the journey they take to get there.
A potential customer might previously have searched:
“office cleaning Cambridge”
They could now ask an AI assistant:
“Can you recommend a reliable commercial cleaning company in Cambridge that works with small offices, can clean outside working hours and has good local reviews?”
That is a very different search.
The BBC picked this up in April 2026 as part of their piece on how businesses are responding to AI search. HubSpot were cited as an example of a company which has seen a dramatic drop in website visits as a result of the change in search behaviour. Their CMO went on to point out the sheer level of detail you can now get in a conversation with an AI as opposed to a standard Google search.
This doesn’t necessarily mean the website has become less important.
Arguably, it means the information on the website has become more important.
AI systems need sources. The customer still needs a source to verify that a business is legitimate. They need to understand what the service is, they need to see examples of it, they need to read about the experience and ultimately decide whether they trust the company sufficiently to make contact.
Your website remains one of the strongest sources of information about your business. The difference is that a human visitor may no longer be the first thing reading it.
SEO still matters — probably more than ever
With new terms such as GEO (Generative Engine Optimisation), AEO (Answer Engine Optimisation) and LLM optimisation appearing everywhere, it is easy to assume businesses now need a completely separate replacement for SEO in 2026.
Google’s own guidance is far less dramatic.
Google says its generative AI search features still rely on its core Search ranking and quality systems. So much of what makes a site good from an SEO point of view is also what makes it a good fit for Google’s AI search experiences.
A website still needs to be:
technically accessible and easy to crawl
logically structured
fast and mobile friendly
clear about what the business does
supported by useful, original content
connected internally in a sensible way
accurate and regularly maintained
backed by genuine experience, expertise and trust signals.
What has changed is the way we need to think about that information.
Ranking a service page for a keyword is still worth doing. We also want the business to be a useful source when someone actually comes to us with a more specific question.
That is where SEO and AI search optimisation start to overlap.
This should happen before the website is designed
This is probably the biggest change I would encourage business owners to think about when commissioning a new website.
Don’t build the website first and “do the SEO” afterwards.
A search strategy should help determine the structure of the website from the beginning.
Before you go all ‘righty-o’ and decide you need a five page website because five pages is about right, we need to establish what the potential customer is actually looking for.
What are your core services? What questions do your customers typically ask before they get in touch? Are the various services significant enough to warrant their own pages? Does location matter at all? What is it that sets you apart from the competition? What evidence do you have that you actually know your subject?
These questions influence the sitemap, page layouts, content and navigation.
They also matter to AI.
Google explains that its AI systems can use a process known as query fan-out, where one complicated question generates a number of related searches in order to build a more useful answer.
You don’t need to understand the technology behind it. The important part for a business owner is simpler.
Someone researching your service may be asking far more than one question.
They might well be asking themselves about cost, suitability, location, timescales, alternatives, experience, reviews and potential problems before they even get to your contact page.
A well-planned website should answer those questions naturally.
The best AI search content isn’t written for robots
One of the more encouraging things about this change is that it should result in better websites.
For years poor SEO advice resulted in pages being written primarily to appease the search engines. Keywords were stuffed in everywhere, towns were used in place of each other on essentially identical pages and businesses would publish loads of generic articles just because someone had suggested that blogging was good for Google.
That isn’t where search is heading.
Google’s 2026 guidance specifically encourages websites to provide unique, non-commodity content — information based on real knowledge, experience or perspective rather than another version of something already written hundreds of times online.
That distinction is important now that anyone can ask an AI tool to produce “10 tips for choosing an accountant” or “five reasons to use a professional cleaning company” in seconds.
Why would another AI system need to reference your version?
Far more useful content might explain something your business genuinely encounters.
A builder could explain the problems commonly discovered when extending a particular style of property.
An accountant might well be in a position to talk about the mistakes that small businesses make all the time when they’re first registering for VAT.
A commercial cleaner could explain what happens when they take over a contract from an existing cleaning company and how they manage the transition.
A specialist manufacturer could answer the questions its technical team repeatedly receives from customers.
That is real business knowledge.
It is useful to potential customers, gives Google greater context and provides AI systems with information that isn’t simply a rewording of everything else already online.
Harvard’s own web publishing guidelines make the same point from a practical standpoint: use descriptive headings, answer the questions directly, use meaningful links, keep the information up to date and write in a natural way rather than forcing your keywords into the copy.
Good AI optimisation can look remarkably similar to simply building a very useful website.
Your business also needs to be easy to understand
There is another aspect of AI visibility that doesn’t get as much attention.
AI systems need to establish context.
Imagine landing on a website where the homepage says:
“Helping you achieve more.”
It might sound impressive to the company that wrote it.
But what does the business actually do?
Compare that to a heading which says the company provides commercial office cleaning for businesses throughout Cambridge and Cambridgeshire.
There’s a lot less ambiguity.
That doesn’t mean every website needs to become boring or stuffed with keywords. Good design and good copywriting still matter.
But businesses should be clear about who they are, what they provide, where they operate and who they help.
Supporting information is important too: an informative About page (with named experts or team members where appropriate), genuine testimonials, case studies, accreditations, business details, service areas and links between relevant pages all go a long way to providing a more rounded picture.
Structured data or Schema markup can provide search engines with added context as well. It remains useful as part of a wider SEO strategy, although Google specifically warns against treating Schema — or any other piece of technical markup — as a magic solution for AI visibility.
The objective is not to trick an AI.
It is to remove unnecessary uncertainty about your business.
Don’t get distracted by every new AI optimisation trick
This is an area where business owners need to be slightly careful.
A whole new industry is emerging around AI search and of course there are already a multitude of ‘shortcuts’.
You may hear that every website suddenly needs special AI files. You may be told content needs to be chopped into tiny blocks so an LLM can understand it. Somebody will almost certainly offer a guaranteed way to get mentioned by ChatGPT.
Treat those claims cautiously.
Google’s guidance is very clear on this point. There is no special Schema required for its generative AI results. There is no requirement to rewrite your pages just for AI. There is no need for an llms.txt file to be present in the AI search results. Google explicitly states that it doesn’t use that file at all for ranking or for visibility in its generative search.
That doesn’t mean experimentation is pointless. Different AI platforms work in different ways and this area will continue developing.
But if the choice is between investing in an experimental AI file or investing in better service pages, original expertise, stronger business information, useful case studies and a technically sound website, I know where I would put the budget.
What should you ask a web designer or agency in 2026?
If you’re talking to companies about building a new website the conversation should now be moving on from colours, number of pages and which WordPress theme they’re going to use.
Ask them:
How will SEO influence the structure of the website before design starts?
Will you research what our customers actually search and ask about?
How will the site communicate our experience and expertise?
Are service and location pages being planned around genuine customer needs rather than simply filling a menu?
Will technical SEO (on-page SEO, structured data, internal linking, crawlability etc) be taken into account during the development phase?
How are you getting on with the Google AI Overviews, ChatGPT and the various other AI search tools?
What happens after launch — is there a strategy for building and improving useful content and will you run an SEO audit?
You don’t necessarily need the web designer themselves to be an AI search specialist.
But somebody involved in the project should understand search strategy.
Otherwise it’s extremely easy to spend thousands of pounds on a beautiful website and six months down the line find you’ve got to go and restructure it all because no one thought about how the customers were going to find it.
A website in 2026 needs to do more than look good
Design still matters.
Your website needs to look professional, work brilliantly on a phone, load quickly and make it easy for somebody to contact you.
But those things are now the starting point.
A modern business website also needs to explain your company exceptionally well.
It should demonstrate your knowledge. Answer the questions customers genuinely ask. Give search engines a clear understanding of your services. Provide AI systems with accurate information they can reference. And, most importantly, give a potential customer enough confidence to take the next step.
SEO and AI search shouldn’t be two separate projects bolted onto a finished website.
Consider them alongside the content, structure, and design from the moment the project begins.
Because the question facing businesses in 2026 is no longer simply:
“Can people find our website on Google?”
It is increasingly:
“When someone asks the internet for a business like ours, are we part of the answer?”
Contact RAD for a free website consultation and to discuss integrating SEO/AI search services.
