GEO GLOSSARY
AI visibility & GEO glossary.
Plain-language definitions for the terms that come up when measuring and improving how a brand appears in AI-generated answers, tied to how MagRank AI actually measures each one where relevant.
AI visibility
How often, how favorably and how prominently a brand appears in answers generated by AI systems such as ChatGPT, Gemini and Perplexity.
Generative Engine Optimization (GEO)
Adjusting a website’s content, structure and public information so AI systems are more likely to cite and recommend it when answering user questions.
Answer engine
A system — ChatGPT, Gemini, Perplexity and similar tools — that generates a direct written answer to a question instead of returning a list of links to click through.
AI Overview
Google’s AI-generated summary that appears above traditional search results for many queries, synthesizing several sources into one answer.
Brand mention rate
The percentage of sampled AI answers to buyer-relevant questions that name a brand at all, whether positively, neutrally or in passing.
Recommendation rate
The percentage of sampled AI answers that actively suggest a brand as a good option, not merely mention it in passing.
AI share of voice
A brand’s mentions divided by all distinct commercial brand mentions across the sampled answers, showing its visibility relative to competitors rather than in isolation.
Citation rate
The percentage of sampled AI answers that cite a brand’s own website or a subdomain of it as a source.
Mention order (position bias)
Where a brand appears among the commercial brands listed in an AI answer — since names near the top of a list tend to get more attention and trust than names near the bottom.
Cross-model presence
How consistently a brand appears across different AI providers — for example OpenAI, Gemini and Perplexity — rather than being visible to only one.
LLM grounding
The process by which a language model anchors its answer in retrieved external sources — search results, crawled pages, a knowledge base — rather than relying only on what it memorized during training.
Retrieval-augmented generation (RAG)
A technique where a model retrieves relevant documents at answer time and includes them in its working context before generating a response, instead of answering purely from what it learned during training.
llms.txt
A proposed plain-text file, modeled on robots.txt, that gives AI crawlers and assistants a curated summary of a site’s most important pages.
Zero-click search
A search result where the user gets their answer directly on the results page or in an AI-generated summary, without clicking through to any website.
Brand entity recognition
The process of identifying and disambiguating a specific business by name within free-form generated text, whether done by the AI model itself or by a tool analyzing its output.
AI hallucination
A confident but factually incorrect statement produced by an AI model — for example inventing a feature, price or fact about a brand that the model was never actually given evidence for.