GEO & SEO Glossary

Definitions of key terms used in Generative Engine Optimisation and search optimisation.

GEO (Generative Engine Optimisation): The practice of optimising a brand's content, structured data, and authority signals so that AI answer engines — including ChatGPT, Claude, Perplexity, and Gemini — recommend the brand in their generated responses. GEO extends traditional SEO by targeting AI retrieval and generation rather than search result ranking.
Share of Model: A metric measuring the percentage of relevant AI-generated responses in which a specific brand is mentioned or recommended, across a defined set of tracked prompts and AI engines. Share of Model is the GEO equivalent of search market share: it quantifies how much of the AI answer landscape a brand occupies for its target queries. Calculated as: (responses mentioning the brand) / (total responses sampled) × 100.
Knowledge Graph: A structured representation of entities (people, organisations, products, places, concepts) and the relationships between them, stored as a graph of nodes and edges. Search engines and AI models maintain internal knowledge graphs to understand what entities exist and how they relate. For GEO, building a clear, consistent, well-cited entity graph for a brand increases the confidence AI engines have when attributing facts and making recommendations.
Topical Authority: The degree to which a website or brand is recognised as a comprehensive, trustworthy source on a specific topic. Topical authority is established through consistent, in-depth coverage of a subject area, backed by accurate citations and genuine expertise. High topical authority increases the likelihood that AI engines will retrieve and cite a brand's content when answering questions in that domain.
E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness): A framework originating in Google's Search Quality Rater Guidelines that describes the qualities of high-quality content and content creators. Experience refers to first-hand knowledge of a topic. Expertise is demonstrated skill or knowledge. Authoritativeness is recognition by others as a leading source. Trustworthiness is the accuracy, honesty, and safety of the content. AI models are trained on corpora that reward these signals, making E-E-A-T relevant to GEO as well as traditional SEO.
llms.txt: A plain-text file placed at the root of a website (at /llms.txt) that provides a curated, machine-readable summary of the site's content for Large Language Model crawlers. It follows a loose convention — a title, an optional summary, and a list of key pages with descriptions — designed to help LLMs quickly understand what a site covers and which pages are most authoritative. It is analogous to sitemap.xml but optimised for LLM consumption rather than search engine indexing.
Structured Data (JSON-LD): Machine-readable markup embedded in a webpage that describes its content using a standardised vocabulary, most commonly schema.org. JSON-LD (JavaScript Object Notation for Linked Data) is the preferred format recommended by Google and used by GEO.actor. Structured data helps both search engines and AI crawlers understand the entities, relationships, and facts present on a page without having to infer them from natural language.
Citation Frequency: The rate at which a specific URL, domain, or brand is explicitly cited as a source in AI-generated answers. High citation frequency indicates that AI engines trust and retrieve from a source reliably. Citation frequency is a component of Share of Model tracking and can be broken down by engine, prompt type, and time period.