--- title: "AI Visibility Glossary: Key AI Search Terms, Defined Once" slug: "glossary" canonical_path: "/glossary" meta_title: "AI Visibility Glossary: 40+ AI Search Terms Defined" meta_description: "Definitions for AI answer visibility, brand names, website citations, source links, the Prime AI Visibility Score, check cadence, Discovery, opportunity research, and Primey." h1: "AI Visibility Glossary: Key AI Search Terms, Defined Once" date: "2026-09-19" last_updated: "2026-09-25" --- # AI Visibility Glossary: Key AI Search Terms, Defined Once This AI visibility glossary gives plain-language definitions for the AI search terms that matter: AI answer visibility, brand names, website citations, source links, check cadence, GEO, RAG, llms.txt and the Prime AI Visibility Score. ## Answer engine A system that generates an answer to a question. Prime AI supports ChatGPT, Perplexity, Claude, Gemini, and an AI Overview-style preview using Google-grounded answers; the Google option is not a capture of the live Google Search AI Overview panel. ## Visible in AI answers Answers that name the brand, cite its website, or do both. Count each answer once. ## Brand named Answers that explicitly name the brand or an approved name variant. ## Website cited Answers containing at least one qualifying link to a tracked brand website. ## Source links Individual citation links observed in a saved answer. Multiple links in one answer do not mean multiple visible answers, and they do not prove the model's hidden reasoning. ## Explicit recommendation An answer clearly recommends the brand. Naming or citing alone does not establish this. ## Comparable change A comparison using a sufficiently aligned question set, tools, persona, market, and measurement method. ## Prime AI Visibility Score A 0–100 summary of observed visibility across the requested check scope. It considers saved brand and website evidence, competitor context, and completeness; missing requested answers cannot improve it. Prime AI does not publish internal weights. ## Check cadence How often saved questions are checked. Monthly is the default; weekly and temporary daily schedules are optional, and on-demand checks use credits. ## Engine breakdown Results separated for ChatGPT, Perplexity, Claude, Gemini, and the AI Overview-style preview using Google-grounded answers. Results reflect the chosen check scope and monthly-default cadence unless another schedule or on-demand check is selected. ## Discovery Research that finds candidate questions from an external citation index and website evidence, shows source context, and lets you select questions to save. A discovered question is not proof of current visibility or a guaranteed rank. ## Opportunity research Research that reviews domain strength, organic Keyword Difficulty, and competing pages to group topics into five work categories: Within reach, Strengthen your content, Build authority, Needs more research, and Review relevance. It is a prioritization guide, not a promised ranking or citation. ## Primey The AI Search Race Engineer in Prime AI. Primey explains authorized saved data for the selected client, points to evidence and possible next steps, and has a monthly reply allowance. ## Prompt set The saved questions selected for a check. Question sets should remain sufficiently aligned when results are compared over time. ## AI crawler An automated agent that fetches web pages for an AI company, for model training, search indexing or live retrieval. Examples include OpenAI's GPTBot and OAI-SearchBot, PerplexityBot and Anthropic's ClaudeBot. ## AI Mode A separate, conversational Google Search experience. Google says AI Mode and AI Overviews may use different models and techniques and show different answers and links. Prime AI does not currently check AI Mode. ## AI Overview A Google Search feature that shows an AI-generated snapshot with links to web pages for some searches. Prime AI's Google option is an AI Overview-style preview built from Google-grounded answers, not a capture of the live AI Overview panel. ## AI search Any search experience in which an AI system writes an answer, often with source links, instead of or alongside a ranked list of pages. ChatGPT search, Perplexity, Gemini and Google's AI features are examples. ## Answer engine optimization (AEO) A term often used interchangeably with GEO: improving content so answer engines can find, understand and quote it. Like GEO, it cannot guarantee a mention, citation or recommendation. ## Canonical URL The preferred address for a page when the same content is reachable at several URLs, declared with a canonical link element. It helps crawlers consolidate duplicates; search engines treat it as a strong hint, not a command. ## Direct answer A short paragraph near the top of a page that answers the main visitor question before supporting detail. ## E-E-A-T Experience, Expertise, Authoritativeness and Trustworthiness, a framework from Google's Search Quality Rater Guidelines for judging content quality. Google says E-E-A-T itself is not a specific ranking factor. ## Embeddings Numeric representations of text that place similar meanings close together. Retrieval systems use them to find passages related to a question before an answer is written. ## Entity resolution The normalization step that maps approved variants of a brand name to the same brand for Brand named reporting. ## FAQPage schema A schema.org JSON-LD type for question-and-answer pairs. Structured data must match visible page content and does not guarantee crawling, a rich result, or an AI citation. ## Generative Engine Optimization (GEO) The practice of improving content and brand evidence so it can be understood and considered by generative answer systems, without promising a ranking or citation. ## Google-Extended A robots.txt product token that lets site owners control whether content Google crawls may be used to train Gemini models and to ground some Gemini answers. It is not a separate crawler, and Google says it does not affect inclusion or ranking in Google Search. ## Grounding Connecting a model's answer to external sources, such as web search results, at answer time. Grounded answers often show source links; a link shows what the answer displayed, not why it was chosen. ## Hallucination A fluent statement in an AI answer that is false or not supported by its sources. Saving the full answer lets a team find and document these errors about its brand. ## JSON-LD A JSON-based format for structured data, placed in a script tag on the page. Google recommends it for most structured data, and this site uses it for its FAQPage and glossary markup. ## Knowledge cutoff The approximate date up to which a model's training data runs. The model may know little about later events, so answers written without web search can miss newer facts, products or prices. ## Large language model (LLM) A model trained on large amounts of text to predict and generate language. Answer engines combine an LLM with other parts, such as search or retrieval, to produce an answer. ## llms.txt A proposed machine-readable index that points AI systems toward useful site URLs. It does not force an AI system to crawl, cite, or recommend a page. ## Prompt The exact text sent to an AI tool in a check. Prime AI calls saved prompts questions. Small wording changes can change the answer, so keep the wording fixed when comparing checks. ## Query fan-out A technique in which an AI search system splits one question into several related searches and combines what they return. Google says AI Overviews and AI Mode may use it; other systems do not always say whether they do. ## References (article) The end-of-article list of publications, primary documents, and reports used to support an article's claims. ## Retrieval-augmented generation (RAG) An architecture in which a system retrieves material at answer time and uses it as context for generation. ## robots.txt A file at the root of a website that tells crawlers which paths they may fetch, standardized as the Robots Exclusion Protocol (RFC 9309). Its rules are requests that well-behaved crawlers follow, not access control. ## Sentiment Whether an answer describes a brand positively, neutrally or negatively. Some tools score it automatically; the saved answer shows the wording behind any score. ## Share of voice (AI) An analyst calculation of how often a brand appears in AI answers. Two formulas share the name: answer share (answers naming the brand divided by relevant answers) and brand-pool share (the brand's appearances divided by all brands' appearances), so always state which. It is not a labelled Prime AI metric, but it can be calculated from saved answers. ## Structured data Machine-readable information about a page, usually written with the schema.org vocabulary. It helps systems interpret a page but must match what the page visibly says, and it does not guarantee a rich result or an AI citation. ## Zero-click search A search that ends without a click to any website, often because the answer was shown on the results or answer page. It is one reason teams track what answers say, not only the traffic they send. ## Read more - [Metrics and methodology](https://primeaivisibility.com/metrics) - [Pricing and credits](https://primeaivisibility.com/pricing) - [How Prime AI works](https://primeaivisibility.com/how-it-works)