> ## Documentation Index > Fetch the complete documentation index at: https://areyoufoundbyai.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # AI Visibility Glossary: Essential GEO Terms Defined > Every key term in AI visibility and generative engine optimisation (GEO, AEO, E-E-A-T, llms.txt, DNS-AID, MCP, and more) defined in plain language. The world of AI search has produced a fast-growing collection of acronyms and concepts, some borrowed from traditional SEO, some entirely new. This glossary defines each term the way an engine can use it: one clear meaning, no filler. Where a term has a dedicated page on this site, you'll find a link to go deeper. **Generative Engine Optimisation (GEO)** is the practice of structuring your business's online presence so that AI engines (ChatGPT, Claude, Perplexity, Gemini, Grok, DeepSeek, and Google AI Overviews) find, trust, recommend, and cite you when people ask them for help. Where traditional SEO competes for a rank in a list of blue links, GEO competes to *be the answer* the AI gives. The term was introduced by researchers at Princeton, Georgia Tech, the Allen Institute for AI, and IIT Delhi in a 2023 paper formalised at ACM SIGKDD 2024. GEO rewards different signals than SEO: entity recognition, third-party citations, quotable content, and off-page brand authority carry far more weight than keyword density or backlink volume alone. See [AI Readiness and Footprint](/docs/concepts/geo-score) for the site signals Are you found by AI? checks. **Answer Engine Optimisation (AEO)** is the practice of structuring content so AI systems can lift it directly into an answer: a clear definition in the first sentence, headings that mirror how people ask questions ("what is…", "how do I…"), and facts an engine can quote without rewriting. Most practitioners use AEO and GEO interchangeably. Where a distinction is drawn: **AEO is about the page**: is it shaped so an engine can extract a direct answer from it? **GEO is about the entity**: does the wider web give the engine a reason to trust and name you at all? You need both: a perfectly structured page from a business AI has never heard of still won't get cited, and a well-known brand with vague copy still won't get lifted. Are you found by AI? measures both: whether the answers name you, and how well your site is set up to be understood. **AI Visibility** is whether AI engines actually name your business when a customer asks a buying question. Are you found by AI? measures it with the **named share** (see below). Reports and the API also carry an **AI Visibility score** out of 100. It is an older weighted diagnostic score, not a percentage of AI recommendations, and it cannot be converted into a named share. Being named in the answer prose counts in full toward that score; a positive result recorded only as a list entry, citation or source title counts half. Its bands are 60 or more "Visible to AI", 25 or more "Partially visible" and below 25 "Rarely visible". See [AI Visibility vs AI Readiness](/docs/concepts/visibility-vs-readiness) for how the results relate. **Named share** is the headline measure: the number of countable saved answers that name the business, divided by all countable saved answers. Named in 7 of 28 countable answers is 25%. Every saved repetition counts, with no majority vote. A link to the website on its own does not count as the business being named. Missing or incomplete evidence is left out, never counted as a no. Read it with the questions, engines, date and answer count beside it. **AI Readiness** (out of 100) is how well your website itself is set up to be understood by AI, measured from the site alone. It is composed of five site signals: Citability, E-E-A-T, Technical, Schema and Platform. A signal that could not be verified is left out rather than scored zero, and each report states the share each signal carries. You can be ready but not yet named: a well-built site the engines do not yet know or trust. See [AI Readiness and Footprint](/docs/concepts/geo-score) for what each signal checks. **Footprint** (out of 100) is the separate off-site score: Wikidata and Knowledge Graph entity, Google Business Profile, brand mentions, search demand and referring domains. It is scored beside AI Readiness, not inside it. The free scan does not measure it, so it shows as not measured rather than zero. An **answer engine** is an AI system that responds to questions directly (ChatGPT, Claude, Perplexity, Google AI Overviews, Gemini, Grok, DeepSeek) rather than returning a list of links for you to evaluate yourself. Answer engines name two or three businesses in their response. You are one of them or you are invisible. There is no page two. This is the core reason AI visibility requires a different measurement and optimisation approach than traditional SEO. **Google AI Overviews** is Google's AI-generated answer panel that appears above traditional organic results on roughly one in four searches. It synthesises an answer from multiple sources and typically names a small number of businesses or resources directly. Clicks that previously reached organic results are absorbed by AI Overviews, independent research (Seer Interactive, 2025) found organic click-through drops roughly 61% on queries where AI Overviews appear. Being cited inside AI Overviews is effectively the new position one. Google's AI Overviews uses a "query fan-out" technique: it issues multiple related searches across subtopics before composing the final answer. Where an engine returns its search queries, Are you found by AI? keeps them with the answer. **llms.txt** is a machine-readable plain-text file published at `/llms.txt` on your domain. It is designed to give AI agents and tools a concise summary of who you are, what you do, and where to find key information, the equivalent of a site map written for a language model rather than a search crawler. An `llms.txt` is a low-effort completeness signal. Some agent tools read it, and it costs nothing to publish. A starter `llms.txt`, written from your site, comes with every [free scan](https://areyoufoundbyai.com/). The full report generates the complete file. **AI crawlers** are the automated bots that AI engines use to read the web and update their retrieval indexes. The main ones to know: | Crawler | Engine | | - | - | | GPTBot | OpenAI / ChatGPT | | Google-Extended | Google (Gemini, AI Overviews) | | PerplexityBot | Perplexity | | ClaudeBot | Anthropic / Claude | If your `robots.txt` or a web application firewall blocks these crawlers, those engines cannot read or retrieve your content. Search access and training preferences are separate controls, so set each one deliberately. Check the published rules for search, training and user-request crawlers with the [free AI crawler checker](https://areyoufoundbyai.com/tools/ai-crawler-checker). An **entity** is the machine-readable identity of your business, a stable, resolvable record that AI engines can find and match across sources. Entities are anchored by signals such as a Wikidata entry, Google's Knowledge Graph, a government business registration (ABN, Companies House), and consistent schema markup with `sameAs` links to your verified profiles. Engines recommend entities they can resolve and trust. If your business name collides with a dictionary word, another company, or a defunct registration (or if your name, address, and trading details are inconsistent across directories) the engine hedges and names someone else. Wikidata and Knowledge Graph entity signals are part of your off-site Footprint score. **Brand mentions** are instances where independent websites name your business, in directory listings, reviews, press articles, community forums, supplier pages, association profiles, and similar third-party sources. Brand mentions are part of your off-site Footprint score, and new pages that mention you are reported on Pro. Review the sources saved with your answers to find relevant places where buyers could find you. **Citations** are the sources an AI answer references, the domains and pages the engine drew on when composing its response. Are you found by AI? keeps the sources returned with each answer, so you can see which sites the engines read in your market. Being cited is not the same as being named. For the **named share**, a link to your website on its own does not count as your business being named. For the older **AI Visibility score**, being named in the answer prose counts in full, and a positive result recorded only as a list entry, citation or source title counts half. **E-E-A-T** stands for Experience, Expertise, Authoritativeness, and Trustworthiness, a framework Google formalised for evaluating content quality and entity credibility, now widely applied to AI engine behaviour. In practice, E-E-A-T signals include: named authors with stated credentials, a real About page, visible contact details and physical address, a privacy policy, customer reviews, industry accreditations, and provenance for factual claims. Recommending a business is a trust decision, so sites that show who is behind them give an engine more to work with. E-E-A-T is one of the five site signals in AI Readiness. **Schema markup**, typically implemented as **JSON-LD** (JavaScript Object Notation for Linked Data), is structured data you embed in your pages to state your business facts (name, location, services, hours, contact details) in a format AI engines can parse without guessing. The most important types for businesses are `Organization`, `LocalBusiness`, `Product`, and `FAQPage`. Adding a `sameAs` array pointing to your verified profiles (Google Business, Wikidata, LinkedIn, etc.) strengthens your entity anchor. **Critical:** the major AI crawlers (GPTBot, OAI-SearchBot, ClaudeBot and PerplexityBot) fetch raw HTML and do not execute JavaScript. If your schema is injected by a tag manager or client-side script, those crawlers never see it. Deliver JSON-LD in the initial HTML response. Valid structured data helps AI *read* you; on its own it does not guarantee that AI will *recommend* you. **DNS-AID** is an experimental IETF draft that lets AI agents discover a site's machine entry point from DNS, through an SVCB index record at `_index._agents.yourdomain.com`. Are you found by AI? records DNS-AID as an unscored DNS observation. It does not count toward any score. Use the [free DNS-AID checker](/docs/tools/dns-aid-checker) to check your current status. **Model Context Protocol (MCP)** is an open standard (developed by Anthropic) for connecting AI agents to external tools and data sources over HTTP. An MCP server exposes a set of callable tools; an AI agent or assistant connects to it and can invoke those tools directly within a conversation. Are you found by AI? exposes each monitored site's measurement through an MCP server. The recommended connection is the static endpoint `https://areyoufoundbyai.com/mcp` with your site token in an `Authorization: Bearer` or `x-api-key` header; the path form `https://areyoufoundbyai.com/mcp/YOUR-TOKEN` keeps working. Connect it to the Claude app, Claude Code, Cursor or another MCP client and ask questions such as "which engines named me last week" or "what did Perplexity say about my competitors". MCP access is included on Free as well as Pro. The demo server (`https://areyoufoundbyai.com/mcp/demo`) is public and needs no key or signup. See the [MCP overview](/docs/mcp/overview) for the tool list and connection instructions. **Agent-ready** describes a site that AI agents can discover, read and act on, beyond basic crawlability. The scan API returns an `agentReadiness` level from 0 to 5: | Level | Name | What it needs | | - | - | - | | 0 | Not reachable | The site could not be reached | | 1 | Basic Web Presence | Online and crawlable | | 2 | Bot-Aware | `robots.txt` lets AI crawlers in | | 3 | Agent-Ready | Structured data and an `llms.txt` | | 4 | Agent-Integrated | Machine-readable Product or Offer data an agent can act on | | 5 | Agent-Native | A published agent interface (`.well-known` manifest or MCP) | At level 3 an agent can understand and recommend you; level 4 and above lets agents act for a customer. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.