Generative engine optimisation (GEO) is the work of making a business and its content more likely to be found, cited and recommended by AI tools that write answers, such as ChatGPT, Perplexity, Gemini, Claude and Google’s AI Overviews. Where traditional SEO aims for a place in a list of links, GEO aims for a mention inside the answer itself.
How generative engine optimisation works
When someone asks an AI assistant “who are good wedding photographers in Bristol” or “what does a conveyancing solicitor charge”, the tool builds an answer from two sources. One is what the model learned during training. The other, increasingly, is a live search: the tool retrieves current web pages, reads them and grounds its answer in what they say, often linking to them as sources. This retrieval step, known as retrieval-augmented generation, is where most GEO work has an effect.
To be used in that answer, a page generally has to be:
- Reachable. AI search crawlers must be allowed in, and key content must not depend on scripts they cannot run.
- Found by the retrieval step. Many assistants draw on conventional search indexes, so pages that rank well in search are more likely to be retrieved.
- Easy to quote. Clear, self-contained statements of fact, prices, steps and definitions are simpler for a model to lift than vague marketing copy.
- Corroborated. A business described consistently on its own site, review platforms, directories and press coverage is easier for a model to trust as an entity.
The term was popularised by an academic paper published in 2023 and is used loosely alongside answer engine optimisation and “AI SEO”. In practice they overlap heavily.
Why it matters
People increasingly start research in an AI tool rather than a search box, particularly for questions that need comparing and explaining. If an assistant recommends three accountants in Manchester and you are not one of them, you have lost that customer before they ever see a results page.
The good news for UK businesses is that most of GEO rests on work that already pays off in search: a well-built site, genuinely useful content, accurate listings and real reviews. The different part is attention to how you are described across the wider web, since AI tools combine many sources into one view of your business.
Measurement is the hard part. At the time of writing (October 2026), there is no equivalent of Search Console for most AI assistants, answers vary between sessions, and tracking tools work by sampling prompts. Treat any precise “AI visibility score” with caution.
Common mistakes
- Treating GEO as a replacement for SEO. Weak search visibility usually means weak retrieval too.
- Blocking AI crawlers by accident. CDN settings and security plugins sometimes do this without the owner knowing.
- Publishing masses of AI-written pages to “feed” the models. That risks Google’s scaled content abuse policy and adds nothing a model cannot already say.
- Inconsistent business facts. Different prices, services or addresses across sites make an AI answer more likely to be wrong about you.
- Buying promises of guaranteed AI mentions. No one controls what these tools say.
How to act on it
Ask the main assistants the questions your customers ask, and note who gets mentioned and which sources are cited. Check your robots.txt and CDN settings allow AI search crawlers. Then make your key pages answer the questions plainly, with facts, prices in pounds and named locations, and make sure your business details match everywhere they appear.
My AI search optimisation guide explains how AI answers choose their sources, and my generative engine optimisation service covers the audit and the changes for businesses that want help doing it.