The GEO glossary

The language of AI search, in plain terms. What each word means, how it works, and why it decides whether buyers find you.

• Generative Engine Optimization (GEO)

  Generative Engine Optimization (GEO) is the practice of making a brand visible, cited and recommended inside AI-generated answers from engines such as ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews. Where SEO earns a ranking on a results page, GEO earns a place inside the answer itself.

• Answer Engine Optimization (AEO)

  Answer Engine Optimization (AEO) is the practice of structuring content so that answer engines, from AI assistants to Google's answer boxes, can extract it and present it as the direct answer to a question. It favours complete, self-contained, quotable answers over long narrative pages.

• AI search (answer engines)

  AI search is search that returns a synthesised answer instead of a list of links, delivered by answer engines such as ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews. The user asks in natural language and the engine composes a response, often naming a handful of brands and citing a handful of sources.

• AI visibility

  AI visibility is the degree to which a brand appears in AI-generated answers: whether it gets named, how it gets described, and whether its content gets cited when buyers ask engines relevant questions. It is the AI-search equivalent of rankings, measured by testing prompts across engines.

• Prompt share (share of voice in AI answers)

  Prompt share is the percentage of relevant buying prompts for which a brand appears in an AI engine's answer, measured per engine with repeated runs. It is the AI-search version of share of voice: of all the moments an engine recommends someone in your category, how many include you.

• AI citations

  AI citations are the sources an AI engine quotes or links when it composes an answer, the pages it used as evidence. Being cited means the engine treated your content as trustworthy enough to build its answer on, which is distinct from merely being mentioned in the answer.

• AI-sourced vs AI-influenced revenue

  AI-sourced revenue comes from deals that started in an AI conversation; AI-influenced revenue comes from deals the buyer touched an AI engine somewhere along the way. Keeping the two separate is what makes AI-search reporting honest, because blending them lets weak programmes borrow credit from strong pipelines.

• llms.txt

  llms.txt is a proposed web standard: a plain markdown file at the root of a website that gives AI systems a curated map of the site's most important content. It tells language models what the site is, what it offers, and which pages matter, in a format built for machine reading rather than human browsing.

• Structural prominence

  Structural prominence is how early and how clearly a page's key facts appear in its structure: headings, first sentences, lists and tables rather than buried prose. Content with high structural prominence is easy for retrieval systems to extract, which makes it more likely to be quoted in AI answers.

• Google AI Overviews and AI Mode

  AI Overviews are the AI-generated summaries Google shows above its classic results, and AI Mode is Google's fully conversational search experience. Together they move Google itself from a list of links to a synthesised answer, with a small set of cited sources doing the work ten results used to do.

• Grounding and retrieval (how AI engines pick sources)

  Grounding is how an AI engine anchors its answer in retrieved sources instead of relying on model memory alone: it searches, reads candidate pages, and builds the answer from what they say. Understanding grounding explains why some content gets quoted in AI answers and other content, saying the same thing, never does.

• The invisible shortlist

  The invisible shortlist is the set of brands an AI engine recommends when a buyer asks who to consider, formed inside a private conversation before any website is visited. Brands left off it lose deals they never knew existed, because no analytics tool records a recommendation that went to someone else.

• FAQPage JSON-LD (structured data for AI)

  FAQPage JSON-LD is schema.org structured data that marks up a page's questions and answers in machine-readable form, embedded as a JSON block in the page's HTML. It gives search and AI systems a clean, unambiguous version of each Q&A pair, independent of the page's visual layout.

• AI search attribution

  AI search attribution is the practice of tracing leads, meetings and revenue back to the AI conversations that produced them. It is hard because AI answers often produce no click, so it relies on captured conversation context, tracked cited paths and buyer self-reporting rather than classic referral analytics.