HelloGeoLabsTM
← All articles

19 May 2026

GEO glossary: 38 AI visibility terms explained

From agentic search to zero-click: the key terms in GEO and AI visibility, in alphabetical order and explained in plain English.

GEO (Generative Engine Optimization) is the work of making sure AI assistants such as ChatGPT, Gemini, Perplexity and Claude mention your organisation. A lot of new jargon has appeared around GEO in a short time. In this glossary we explain the key terms briefly and in plain English, in alphabetical order. Where useful, we link to an article with more detail.

A to C

Agentic search: search in which an AI agent takes several steps by itself. It searches, compares options and can sometimes book or order something on the user's behalf. Read more about AI agents that search for customers.

AI agent: an AI program that carries out tasks independently, such as visiting websites, filling in forms or making an appointment.

AI crawler: a bot from an AI company that fetches web pages, for example GPTBot, PerplexityBot or ClaudeBot. See AI crawlers in your server logs.

AI Overviews: the AI summaries Google shows at the top of some search results, with links to sources.

AI SEO: another name for GEO. Optimising so that AI assistants find, understand and mention you.

AI visibility: how often and how well AI assistants mention your organisation for relevant questions.

Brand mention: every time your brand or business name is mentioned in an AI answer or on another website, with or without a link.

Citation: a reference to a source in an AI answer. Being mentioned and being cited are not the same: you can be mentioned without your website being the source. See citations in AI answers.

Citation share: the share of all citations in a set of answers that point to your website or pages.

Competitor tracking: keeping track of which competitors AI mentions, how often and for which questions.

Content gap: topics or questions where competitors are mentioned and you are not, often because you have no good content on them.

E to G

E-E-A-T: experience, expertise, authoritativeness and trust. A Google framework for judging quality, also relevant for AI. See E-E-A-T and AI.

Entity: a recognisable "thing" such as a person, place or organisation that AI understands as a single whole. See entities.

GEO (Generative Engine Optimization): optimising your online presence so that AI assistants mention you. Read what GEO is.

Grounding: backing up an AI answer with current sources the model fetches live, rather than relying only on knowledge from training.

H to O

Hallucination: an AI answer that sounds convincing but is wrong, such as an incorrect address or a service you do not offer.

Knowledge cutoff: the date up to which a language model has training data. It only knows about later events through live search.

Knowledge panel: the information box Google shows next to search results about a recognised entity.

LLM (Large Language Model): the technology behind AI assistants. See what is an LLM.

LLM SEO: yet another name for GEO, with the emphasis on language models.

llms.txt: a proposed file in the root of your website that gives AI an overview of your most important pages. See llms.txt explained.

NAP: Name, Address, Phone. Your basic details, which should be the same everywhere.

Onsite and offsite: onsite is about your own website. Offsite is everything outside it, such as reviews, directories, media and forums.

P to R

Persona: a description of a type of customer. By writing prompts per persona, you measure visibility per audience.

Position: where you appear in an AI answer, for example as the first or third option.

Prompt: the question or instruction someone gives an AI assistant. See which questions customers ask AI.

Prompt tracking: asking AI models a fixed set of prompts regularly and recording whether and how you are mentioned.

Prompt volume: how often a particular question is asked to AI. Reliable figures hardly exist yet; it is usually a matter of estimates.

RAG (Retrieval-Augmented Generation): a technique in which a model first retrieves relevant information and then uses it to write the answer.

robots.txt: a file that tells crawlers which parts of your website they may visit. See allowing AI crawlers.

S to Z

sameAs: a property in structured data that links your website to your official profiles elsewhere.

Sentiment: the tone in which AI talks about you: positive, neutral or negative, and which strengths and weaknesses it mentions.

Share of voice: your share of all mentions within a set of questions, compared with competitors. See share of voice in AI.

Source: a web page an AI assistant uses to support an answer. Often shown with a link.

Structured data: code on your website (usually schema.org) that states in a fixed format who you are and what you offer.

Training data: all the text a language model was trained on.

Visibility score: a percentage showing for how many of the questions asked you are mentioned. See the AI visibility score explained.

Zero-click: a search where someone gets the answer straight away and does not click through to a website. See zero-click search and AI.

From terms to numbers

Now that you know the language, you probably want to know where you stand. With the free AI check we ask five customer questions to ChatGPT, Gemini, Perplexity and Claude. Want to follow your visibility, position and share of voice every week? Take a look at GEO tracking.

Curious what AI says about your business?

Request the free check and receive your report within 1 day.

Start the free AI check