Search isn’t just Google anymore
For over twenty years, being found online has been about one thing: ranking well on Google. In 2026, that is changing fast. Analysts have predicted that traditional search volume could drop by a quarter this year, as more people turn to AI tools for answers instead of typing a search and clicking through a list of links. Google’s own AI-generated answers (“AI Overviews”) now reach over 2 billion people a month and already show up on roughly 3 in 10 UK Google searches. ChatGPT is used by around 800 million people every week, and UK visits to it alone jumped from 368 million to 1.8 billion in under a year.
What this means for your business: some of the customers who used to find you through a Google search are now getting their answer directly from an AI tool, sometimes without ever clicking through to your website at all. And when an AI tool does name its sources, it usually only mentions a handful — typically somewhere between two and seven. Being one of those few, not just appearing somewhere on page one of Google, is the new challenge, and its name is Generative Engine Optimisation, or GEO.
Traditional SEO vs GEO: what’s actually different
Traditional SEO and this newer idea of GEO both aim for the same thing — being found by the right customers — but they’re playing to a different audience. SEO is written for a search engine (usually Google) that ultimately hands a person a list of links to choose from. GEO is written for an AI tool that reads your page itself, decides whether to trust it, and then folds your information into its own answer — sometimes crediting you, sometimes not.
That shifts what “doing well” looks like. The old optimisation tools such as having the right keywords, plenty of other sites linking to you etc. still count for classic Google rankings. But AI tools seem to lean more heavily on independent, third-party evidence (reviews, press coverage, other people talking about you) than on what you say about yourself. And because an AI answer usually only draws on a small handful of sources, being clear, accurate and well-organised matters just as much as being popular.
Importantly, none of this replaces good old-fashioned SEO — it builds on it. A slow, disorganised, poorly indexed site was never going to rank well, and it’s not going to get picked up by AI tools either. The foundations haven’t changed; what’s new is an extra layer of clarity built specifically for these AI “readers”.
How AI actually “reads” your website
This is the bit that’s easy to miss, and it’s the part that’s really about design rather than writing.
An AI tool doesn’t read your whole page the way a person would, taking in the design and the layout and getting a feel for it. It breaks the page up into small chunks — a section here, an answer to a question there — and judges each chunk on its own. So a great answer buried halfway down a page, surrounded by clutter, can easily get skipped over, even if the writing itself is excellent.
AI can’t “see” your design the way a human can. A person glances at a page and instantly knows what’s a price, who wrote it, or what the heading is, just from how it looks. AI tools need that spelled out clearly in the underlying code structure of the page, not just implied by colour and layout or a larger line of text. A page that looks like a great design, but that is built as one big block of styled text is effectively invisible to an AI tool reading it.
The upshot: it’s no longer enough for a page to look good. It also has to be built so a machine can tell, section by section, what each part actually is.
What that means for how Instilled design your pages
While our websites have always been designed and built with traditional SEO and Accessibility in mind, in practical terms, this changes a handful of everyday design decisions:
- Clear, properly structured headings. Not just headings that look bold and big — headings that are genuinely marked as headings, in a sensible order. This is what lets an AI tool understand what each section of the page is about.
- Answer the question up front, then explain. Start each section with a plain, direct statement of the point, then add the detail underneath. That first line then works well whether a human skims it or an AI tool quotes it.
- Build real FAQ sections. The questions your customers actually ask, with clear direct answers. AI tools lean heavily on this kind of question-and-answer format when pulling together a response.
- Avoid “pretend” formatting. A price, a list, a table should be built as an actual price, list or table in the page — not just styled to look like one. This is exactly the sort of thing we already include as standard.
- Don’t hide your key content. If your most important text only appears after fancy animations or scripts finish loading, some tools reading your site may never see it at all.
None of this is exotic or new — it’s mostly what we have always produced as good website design. AI search has simply raised the cost of getting it wrong.
“Schema markup”: a label, not magic
You’ll hear the term “schema markup” or “structured data” a lot in this space. Behind the jargon, it’s a simple idea: it’s a way of labelling information on your page so a computer doesn’t have to guess what it’s looking at — this is the price, this is the opening hours, this is who wrote the article, this is a genuine customer review.
It’s worth being realistic about what it does and doesn’t do. Labelling your content this way does not guarantee an AI tool will quote you or that you’ll appear in an AI-generated answer. What it does is make your site easier to interpret correctly once it’s already being considered — think of it as removing ambiguity, not buying a place in the results.
For most small business sites, the labels worth having cover: who you are and how to contact you, your opening hours and service area if that applies, your articles or guides (with author and date), your products and prices if you sell online, and genuine customer reviews. One thing worth flagging: FAQ-style labelling used to earn a nice-looking snippet directly in Google search, but Google removed that particular perk in 2026. It’s still worth building genuine FAQ sections for AI tools to read — you just won’t get the same visual reward in classic search results.
The practical takeaway for a client site: this is a technical setup task best handled once, properly, by whoever builds or maintains your site — it’s not something you need to manage yourself, but it is worth asking your web team whether it’s in place.
The basics still matter — there’s just more of them now
Everything that’s always made a website work well — fast loading, sensible navigation, working properly on a phone, being easy for search engines to find in the first place — still comes first. None of the AI-search advice above matters if a site is slow, disorganised, or effectively invisible to search engines already.
What’s changed is that there’s now a second set of automated “visitors” to think about, alongside the familiar ones from Google and Bing: the tools that crawl the web on behalf of ChatGPT, Claude and others. A site can accidentally block these without realising it, especially if security settings were set up to block unfamiliar bots by default.
Freshness and authority: the signals AI engines lean on
Two signals matter more here than they typically did for classic SEO. The first is recency: AI engines weigh how up-to-date a source is when choosing what to cite, so a strong guide from 2024 that’s never been touched since will keep losing ground to a 2026 article covering the same topic. Keeping cornerstone pages (services, key guides, pricing) refreshed with current information and a visible “last updated” date is a small edit with an outsized effect.
The second is earned authority. AI engines appear to favour independent, third-party mentions — reviews, press coverage, industry write-ups — over what a brand says about itself on its own site. In practice, that means a clear About page, a real author bio on content pages, consistent business details across the web, and active pursuit of genuine reviews and press mentions all feed into how much an AI engine trusts what your site says, not just what’s written on the page itself.
A quick checklist for your own site
If you want a simple gut-check for your own website, ask your web team (or us) these questions:
- Are our headings and page structure genuinely clear, not just styled to look tidy?
- Do our key pages answer the obvious question in the first line, before going into detail?
- Do we have real FAQ sections answering the questions customers actually ask?
- Is our contact information, opening hours and business details clearly labelled for search engines and AI tools to pick up correctly?
- Are our best pages kept up to date, with a visible “last updated” date?
- Do we have a proper About page and real author information, not placeholder text?
- Are we actively gathering genuine reviews and press mentions, rather than relying only on what we say about ourselves?
- Is the site itself still fast, mobile-friendly and easy to navigate?
At Instilled, we design for people, and we structure for machines
The encouraging news buried in all of this is that GEO isn’t really a bolt-on discipline — it’s a stricter, more literal version of good web design and content practice that’s been sound advice for years: clear structure, honest content, real technical performance. The sites that will do well in an AI-search world are, for the most part, the same sites that were already well built.
At Instilled we build every client site with this structure baked in from the start — semantic Kadence layouts, proper schema, and content that’s genuinely organised rather than just visually tidy — rather than treating it as an SEO add-on after launch. If you’re not sure whether your current site is set up to be read (and cited) by AI search as well as it ranks on Google, that’s a conversation worth having before your next rebuild or refresh.
Sources: Search Engine Land – Mastering GEO in 2026; Opace – Structured Data & Schema Markup for SEO, AI Search and LLMs (2026)
