What is AI SEO? A plain guide to being found in AI search

AI SEO is the work of making your business easy for AI search tools to understand, cite and recommend. What it covers, where GEO and AEO fit, and where to start.

The Brand Muffin

AI SEO is the work of making a business easy for AI search tools to understand, cite and recommend. It covers the same ground as search engine optimisation, with one change of focus: the reader is now often a machine that will write an answer from your pages before a person ever sees them.

That is the whole idea. The rest of this note explains the words around it, what the work involves and what it cannot promise.

What AI search is

AI search is any search where the result is a written answer. Google does it with AI Overviews and AI Mode. ChatGPT, Perplexity, Copilot and Gemini do it in a conversation. In each case a person asks a question, the tool reads a number of web pages and replies in sentences, usually listing the pages it drew on.

For a business, that changes where you can appear. A results page had ten places. An answer has room for a few sources, and sometimes for one recommendation.

Two meanings of the same phrase

“AI SEO” is used for two different things, and it helps to know which one is meant.

The first is using AI tools to do SEO work: drafting copy, sorting keywords, checking pages. The second is being found in AI search: making sure that when an assistant answers a question about your field, your business is among the sources it reads and names.

We mean the second. It is the service we call AI SEO.

Where GEO and AEO fit

Two acronyms travel with the phrase.

Generative engine optimisation (GEO) is about generated answers, the kind Google’s AI Overviews or ChatGPT write. The term comes from a research paper published in November 2023.

Answer engine optimisation (AEO) is about direct answers to specific questions: the short reply to “what is”, “how much” or “who does”.

You will see the two used for the same work, and that is fine. AI SEO is the plain name for both. If someone offers you GEO or AEO, ask what they will change on your website and how they will show you the result. The answer matters more than the label.

How it differs from SEO, and how it does not

Less than the new vocabulary suggests. Google’s own guidance is direct about it: “The best practices for SEO remain relevant for AI features in Google Search”. The same page says there are no extra technical requirements and no special files or markup to add. A page has to be indexed and eligible to show in ordinary search before it can be a source for an AI answer.

So the foundation is the one we described in why organic SEO matters more than ever: pages that load, read clearly and say true, useful things.

Three things do change.

  • The answer is built from passages. An assistant lifts the two or three sentences that answer the question. A page that states its point plainly, near the top, is easier to quote than one that takes five paragraphs to arrive.
  • One question becomes several searches. Google says its AI features may run a number of related searches across subtopics to build one answer. Pages that answer the narrower questions around a subject get drawn in, not only the page that targets the main phrase.
  • Being named counts, with or without a click. A person may read the answer and never visit a website. If your business is the one the answer names, you were still found.

Other assistants choose their sources in their own ways, and their answers can change from one day to the next. That is a reason to get the basics right everywhere, since no single trick covers them all.

What the work involves

The ingredients are familiar. The care goes into how clearly they are put together.

  1. A clear picture of the business. The same name, services and contact details wherever they appear, and specific descriptions of what you do. Ambiguity is what stops a machine from repeating you.
  2. Pages that answer real questions. One page for each question a customer asks, with the answer first and the detail after it.
  3. Structure a machine can read. Text as text, sensible headings, links between related pages, and structured data that describes only what the page shows.
  4. Observation. A fixed list of questions, asked of the same tools at intervals, with the answers and their sources recorded.

What it cannot promise

Nobody can guarantee that an assistant will recommend a business. Each platform decides its own answers, and the same question can return different sources twice in one afternoon.

Measurement has limits too. Google’s Search Console now has a generative AI performance report, which counts how often links to your site were shown in AI Overviews and AI Mode. It covers Google only, and it counts appearances, not enquiries. Honest reporting says what was observed, when, and how far it can be trusted.

Where to start

Three checks cost nothing and take an hour.

  • Ask two or three assistants about your own business and your main service. Note anything wrong, out of date or missing.
  • Compare the facts on your website with your profiles elsewhere. Names, services and contact details should match.
  • Write down five questions customers ask you most. For each one, find the page on your site that answers it. Where there is none, that is the first page to write.

If the list of gaps is longer than you expected, tell us what’s on your mind and we’ll call you back.

Have a question about your own website?

Request a call back