A solicitor in Cardiff ran an experiment last month. She typed “best family law solicitor in Cardiff” into Google, read the AI-generated summary at the top of the page, and never scrolled. The eight firms named in that summary received a thousand pounds of free advertising. The dozens of firms ranking below it got nothing.
That is search in 2026. The middleman has moved.
What an AI Overview actually is
An AI Overview is the boxed answer Google now generates at the top of most commercial searches. It pulls from the open web, summarises the relevant points, and lists the sources it considers authoritative. Other AI assistants (ChatGPT, Gemini, Claude, Perplexity) work the same way: they retrieve, summarise, and cite. Google’s announcement of generative AI in Search is where the consumer rollout was formalised.
From the user’s point of view, they ask a question and get an answer. From a business owner’s point of view, the question is suddenly very different. It is no longer “are we on page one?”. It is “are we in the citation pool?”.
Why this is a bigger shift than people think
For twenty years, SEO was a popularity contest with click-through-rate. You ranked, people clicked, you got a chance to convert them. The whole industry was built around that funnel.
AI Overviews break the funnel in two important places.
Most searches no longer produce a click. Search Engine Land and several industry trackers have been showing zero-click rates above sixty percent on commercial queries with AI Overviews present. The user reads the summary and moves on.
The pages that get cited are not always the pages that used to rank. AI engines weight clarity, structured answers, named authors, and verifiable trust signals heavily. A smaller, well-organised firm with a tight, well-structured page can get cited above a much larger competitor with a thin one.
This is the genuine opening for smaller, well-organised businesses. Not because the rules are easier (they are not), but because they are different.
What gets a page cited by an AI Overview
Pattern recognition from the last eighteen months of AI Overview behaviour points to five things.
Direct, structured answers. AI assistants prefer pages that answer the question in the first 150 words. Long pre-amble works against you. Get to the point, then expand. Treat each H2 as a chance to surface a clean, citable answer; the principles behind rich snippets apply directly.
Schema markup. Article schema, FAQ schema, Organisation schema, How-To where relevant. The structured data is how AI engines understand what kind of content they are looking at. Our glossary defines each of these in plain English, and Schema.org has the canonical vocabulary.
Named authors and verifiable experience. A page written by a named person with a real bio outperforms a page written by “the team”. This is what the SEO world calls E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness). AI engines weight it heavily.
Topical authority across a cluster of pages, not one isolated page. If your site has one page about a topic, you look like a casual mention. If your site has ten internally-linked pages forming a hub around a topic, you look like a primary source. This is the core of topical authority and it is built into how we structure every SEO strategy engagement. Strong internal linking ties the cluster together.
External validation. AI engines cross-reference. A page that says “we are experts in X” without anyone else corroborating it gets weighted lower than a page that has reviews, mentions in industry publications, or visible accreditations.
GEO and AEO are the new acronyms worth knowing
Two related disciplines have emerged to deal with this shift, and they are not interchangeable.
Generative Engine Optimisation (GEO) is the practice of optimising for being cited by AI-generated answers. The focus is on becoming a quotable source: clear definitions, structured pages, named authors, topical depth. We covered the practical framework in how to get your business found in ChatGPT, Google AI Overviews, and AI Search.
Answer Engine Optimisation (AEO) is the closely-related practice of structuring content so AI engines can extract clean answers from it. FAQ blocks, short direct answers, and rigorous use of schema markup are the workhorses.
Most SEO work in 2026 should include both. They share most of their foundations with traditional SEO, but the emphasis is different.
What this means for your existing content
Most existing pages on UK business sites were written for a world where the user clicked through. They start slowly, hold information back, and rely on the reader scrolling. AI engines lose interest before that pays off.
Three changes that almost always lift a page’s chance of being cited:
- Lead with the answer. Put the direct response to the page’s title question in the first paragraph. Expand below.
- Add a short FAQ block. Three to five real questions a customer would actually ask, with two or three sentence answers. Mark up the section with FAQ schema.
- Name the author. Replace “by The Team” or “Admin” with the actual person who wrote it. Add a one-paragraph bio at the foot of the article. Link to their author page.
You do not have to rewrite the page. These three changes alone shift the signal.
What the metrics now tell you
Old SEO metrics (rankings, click-through-rate, organic clicks) still matter, but they are now half the picture. The other half is whether you are being cited in AI answers.
Two ways to track that:
- Run sample queries periodically. Pick the ten questions you most want to be cited for. Ask ChatGPT, Gemini, Google AI Overview, Perplexity. Record whether your domain is named. Repeat monthly.
- Watch AI assistant referral traffic in GA4. Sessions from chat.openai.com, perplexity.ai, gemini.google.com and similar. A year ago this was statistical noise. In 2026 it is a real channel. Google’s GA4 setup docs cover the channel grouping config.
If your site is being cited, the citation often does not produce a click. But it produces brand awareness and, increasingly, direct enquiries from people who saw your name in an AI answer and visited later. If you are seeing a traffic drop without a conversion drop, work through our diagnostic on why your website traffic has dropped before assuming the worst.
What you should not do
Two things to be careful of, because both are tempting and both backfire.
Do not stuff AI-generated content onto your site to “feed the AI”. Google’s classifiers spot the pattern almost immediately, and being flagged hits all your pages, not just the new ones.
Do not stop publishing for humans. AI engines weight content that real people engage with. Pages that read well, get shared, attract comments and earn time-on-page perform better in AI citations too. Writing for AI is writing for humans first, with structure that AI can also follow.
Frequently asked questions
Are AI Overviews replacing traditional Google search results?
Not replacing, layering on top of. The blue-link results still appear below the AI summary. But the summary captures most of the attention, so the practical impact is the same: if you are not in or near it, you might as well not be there.
How can I tell if my site is being cited in AI answers?
Run sample queries directly in ChatGPT, Gemini, Google AI Mode, Perplexity. Search your brand name and the questions you want to win. Record the results. Repeat monthly to track movement.
Do AI Overviews favour big brands?
Less than you would think. Big brands have weight, but the AI engines also weight structural clarity, named authors and topical depth. Smaller, well-organised sites that get those right out-cite much larger competitors regularly. The pattern in 2026 is “well-organised mid-size and small sites are punching well above their weight”.
Will my old SEO work still pay off?
Mostly yes. Strong technical foundations, a useful site that loads fast, and pages that genuinely answer real questions are the underpinning for AI Overview citations too. The shift adds new things to do, it does not invalidate the old ones.
How long does it take to start being cited?
Faster than traditional ranking, often. We have seen pages that were freshly published and well-structured cited within four to eight weeks. The bottleneck is usually the quality and clarity of the page itself, not Google’s index speed.
Where to go next
If you want a look at how visible your business currently is in AI answers, and what the realistic shortest path to changing that looks like, an AI visibility check is a useful place to start. AI search is built into every SEO package as standard.
Book a discovery call and we will run sample queries against your site live on the call, no slide deck required.
