AI search glossary: AEO, GEO and the rest

A program run by an AI company that fetches web pages automatically, either to collect training data or to read pages live while an assistant answers a question. Examples include GPTBot, OAI-SearchBot, ClaudeBot and PerplexityBot.

Google's conversational AI search, in which Google answers a question in writing and lets you ask follow-ups, drawing on Google Search.

The AI-written summary Google shows above its normal results for some searches. For many local searches Google shows a map of businesses instead of an AI Overview.

How often a business is named or cited when people ask AI assistants the questions that matter to it. It is best measured as a rate across many askings, because individual answers vary.

Making a business easier for AI assistants and AI search features to find, understand and cite in their answers. It overlaps with SEO but focuses on being named in a written answer rather than ranked in a list of links.

A link an AI assistant shows as a source for its answer. Being cited means the assistant used a page as evidence; it is a different fact from being named in the text, and the two are best measured separately.

The share of answers in which an assistant linked to a page on the business's own website as a source.

Short for query fan-out.

Another name for answer engine optimisation: making a business easier for generative AI systems to find, understand and cite.

When an AI assistant searches the web or another source while answering, and bases its answer on what it finds rather than on its trained memory alone. Grounded answers usually show citations.

An AI answer that states something false with confidence, such as a wrong address or a service a business does not offer. Correcting the sources an assistant reads is the lasting fix.

The date after which a model learned nothing in training. Anything that changed since then can reach its answers only through web search.

The kind of AI model behind ChatGPT, Claude, Gemini and similar assistants: a system trained on large amounts of text to predict and generate language.

A proposed plain-text file at the root of a website that summarises the site and links to its key pages for AI systems. It is not an official standard, and assistants are not obliged to read it.

The share of answers in which a business's name appears in the text, across a fixed set of questions.

The way an AI assistant breaks one question into several smaller searches behind the scenes before writing an answer. A page can be cited because it answers one of those smaller questions well.

A text file at the root of a website that tells crawlers which pages they may read. A rule written for one purpose can accidentally block AI crawlers.

Code on a web page, invisible to visitors, that states facts about the business in the standard schema.org format: name, type, address, phone, opening hours.