What is GEO, and why does it have a new name?
GEO, generative engine optimization, describes the practice of making a website easy for AI systems, ChatGPT, Google's AI Overviews and AI Mode, Perplexity, to find, understand, and cite. It sits next to traditional SEO work rather than replacing it: the target audience is an AI system generating an answer rather than a search engine ranking a list of links, but the underlying requirements overlap more than the new name suggests.
Why this is worth a UK service business's attention now: a growing share of research before a first enquiry happens inside a chat interface rather than a traditional search results page, particularly for considered purchases where someone might ask an AI assistant to compare a handful of options before visiting any website directly. If a business is invisible to that step, it is not losing a ranking position, it is missing the conversation entirely, before a visitor ever reaches the site to be judged on anything else, including how strong the offer actually is.
Why ranking on Google no longer guarantees an AI citation
Google's own AI features documentation sets out the actual eligibility bar plainly: "to be eligible to be shown as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet". That is a necessary condition, not a sufficient one. Plenty of pages that are indexed and rank respectably still never get surfaced as a cited source, because being eligible to appear and being the specific passage an AI system chooses to quote are two different things.
The gap between the two shows up most on pages that bury their actual answer under introductory copy. A human visitor will scroll past a paragraph of scene-setting to get to the point. An AI system summarising a page for a one-line citation is far less patient, and if the clearest statement of what you do or offer is three paragraphs down, a competitor whose homepage states it in the first sentence is the one that gets quoted.
It also shows up on pages where the content that matters is loaded in after the initial page render, common on sites built with heavy client-side JavaScript and no server-rendered fallback. If a crawler fetching the page sees an empty shell before your scripts run, there may be nothing there yet to index, let alone cite. This is an old SEO problem with a new consequence: it now also blocks AI visibility, not just search ranking.
How AI Overviews, ChatGPT, and Perplexity source answers differs
The three systems most people mean when they say "AI search" work slightly differently, though the underlying requirement, a crawlable, well-structured page, holds across all of them. Google's AI Overviews and AI Mode draw on Google's own index, so the eligibility bar above (indexed, snippet-eligible, meeting Search's technical requirements) applies directly. Perplexity is built around live web citation by design, retrieving and quoting sources for most answers rather than treating citation as an occasional feature. ChatGPT cites web sources when browsing or search is enabled for a given answer, and otherwise draws on its training data, which is one reason consistent, long-standing content tends to be represented more reliably than something published last week. None of the three reward keyword density or hidden metadata over an actual, clearly stated answer.
What actually determines whether you get cited
Google is direct that structured data is not required for generative AI search, and there is no special schema.org markup needed to appear in these features. The guidance instead points back to the same foundational practices as ordinary SEO: technical crawlability, clean indexing, and "helpful, reliable, people-first content." In practice, three things matter most for whether an AI system can use your page as a source.
| Factor | Why it matters to an AI system |
|---|---|
| Indexed and crawlable | A page an AI system cannot reliably fetch and index cannot be cited, regardless of content quality. |
| A clear, early answer | Systems favour passages that state a specific answer near the top over pages that require inference to summarise. |
| Descriptive structure | Clear headings and short, self-contained sections are easier to extract a single citable passage from than long, undivided copy. |
None of this requires new markup, a separate "AI-readiness" build, or rewriting a site from scratch. It mostly overlaps with what a well-structured, genuinely useful page already does, which is part of why Google frames it as an extension of existing SEO practice rather than a new discipline.
Common mistakes that block AI visibility
Most of the sites that are indexed but never cited share a small set of avoidable problems, none of which require a rebuild to fix:
- Core content loaded only by client-side JavaScript, with no content present in the initial HTML for a crawler to read.
- A vague homepage opening, brand story or mission statement first, with the actual service or offer stated several paragraphs down.
- One undifferentiated "services" page covering everything the business does, instead of specific pages a system can match to a specific question.
- No FAQ-style content answering the direct questions a prospective client would actually ask, in plain language rather than marketing copy.
- Duplicate or missing meta descriptions, which weakens how confidently a system can summarise what a given page is about before ever reading it in full.
A quick way to check your own visibility
Ask ChatGPT, Perplexity, or Google's AI Mode a question a prospective client might actually ask, something like "who does [your service] in [your area]", and see whether your business is named or cited. Do this for two or three realistic questions, not just your own brand name, since branded queries tend to surface a site regardless of how citable the content actually is. If you are indexed in Google Search but never appear in these answers, the gap is usually structural rather than about whether you exist online at all.
What to fix if you are invisible
This is where AI visibility and the CRISP dimensions overlap directly. The Contextual dimension, whether a visitor (or a system reading on their behalf) can tell what you do within seconds, and the Intelligent dimension, whether the site's structure guides a reader to the right answer, cover much of the same ground as GEO best practice. A site that scores well on both is, by construction, also easier for an AI system to extract a clean citation from. Fixing the underlying clarity and structure issues tends to improve both at once, rather than requiring a separate AI-specific project.
You can see where your own site stands with the free CRISP audit, or read more about what a full audit covers. Where the fix goes beyond a few pages of copy, that is what the rebuild service is for.
Measure the result as a repeated observation, not a one-off screenshot. Choose five realistic prompts that represent different stages of a buyer's research, record the assistant and date, and note whether your business is named, linked, or absent. Repeat the same set monthly. Answers can vary between tools and sessions, so the trend is more useful than any single response.
Keep ordinary search data beside that record. If impressions, indexed pages, and branded searches are growing while AI mentions remain absent, you may have a citation or clarity gap. If both search and AI visibility are weak, the problem is probably more fundamental. This comparison does not prove why an answer changed, but it stops the team mistaking random variation for progress and gives every content change a consistent baseline.
Frequently asked questions
Do I need special schema markup to appear in AI Overviews?
No. Google's own documentation is explicit that structured data is not required for generative AI search and that there is no special schema.org markup needed to appear in these features. What is required is being indexed and eligible to appear in regular Google Search with a snippet.
Does AI search visibility replace normal SEO?
No. Google's guidance treats the same foundational SEO practices, technical crawlability, indexability, and helpful, people-first content, as the basis for AI Overviews as for normal search results. AI visibility is closer to a consequence of solid SEO than a separate discipline replacing it.
Why would ChatGPT or Perplexity cite a competitor instead of me?
AI answer engines tend to favour pages that state a clear, specific answer near the top, use descriptive headings, and are easy to extract a single passage from. A page that requires scrolling through marketing copy before reaching an actual answer is harder for these systems to quote cleanly, even if the underlying business is a good fit.
How do I check if my site is currently visible to AI search?
Ask ChatGPT, Perplexity, or Google AI Mode a question a prospective client might ask about your service and see whether your site is cited. Check indexing status in Google Search Console. And run Fresh & Crispy's free CRISP audit, since the Contextual and Intelligent dimensions catch many of the same structural issues that make a page hard for an AI system to cite.