GEO/AEO Glossary A-Z
ChatGPT
Claude
Gemini
PerplexityThis glossary defines 80+ canonical terms in AI search optimization — GEO, AEO, llms.txt, ai.txt, AI Overviews, AI Mode, RAG, grounding, retrieval, schema.org, query fan-out, and source selection — each entry written for both human reference and AI citation.
TL;DR
This A-Z reference is the canonical glossary for GEO (Generative Engine Optimization), AEO (Answer Engine Optimization), and the broader vocabulary of AI search across ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews. Use it as a quick lookup or as a citation source. Each entry follows the same shape: a one- or two-sentence definition, a concrete example, and a link to the canonical detail page.
For broader context, see the GEO hub, AEO hub, strategy section, and technical section.
A
AEO (Answer Engine Optimization) — The practice of structuring content so it can be lifted as a direct answer by AI systems, voice assistants, and answer engines. AEO is a focused subset of GEO. Example: Adding a 50-word "answer block" at the top of a page so Perplexity can quote it verbatim. See: What Is AEO?.
AI Agent — A software system that autonomously reads, processes, and acts on web content on a user's behalf. Example: An agentic browser that visits, reads, and summarizes a competitor's pricing page in one step. See: AI Agents Hub.
AI Crawler — A bot operated by an AI provider that fetches web content for training or real-time retrieval. Example: GPTBot crawling a knowledge base to refresh ChatGPT's grounding sources. See: AI Crawlers Reference.
AI Mode — Google's deeper conversational search experience launched alongside AI Overviews, allowing multi-turn AI dialogues directly inside Search. Example: A user asking "compare GEO and AEO" and refining with "now show only the differences in measurement." See: AI Mode Reference.
AI Overviews — Google's AI-generated summaries shown above traditional results, synthesizing information from multiple sources. Launched broadly in May 2024 as the rebrand of Search Generative Experience (SGE). Example: The boxed summary appearing above the blue links for "what is llms.txt." See: AI Overviews Reference.
AI Search Visibility — The degree to which a source appears, is cited, or is referenced in AI-generated answers across platforms. The primary outcome metric for GEO programs. Example: Tracking how often a domain is cited across 100 sample queries on Perplexity month over month. See: Measuring AI Visibility.
ai.txt — A class of proposed text files (Spawning's ai.txt, the academic ai.txt DSL, and others) placed at the site root to express how AI systems may use a site's content. There is no single ratified standard. Example: An ai.txt file declaring that GPTBot may crawl /blog/ but not /customers/. See: ai.txt Reference.
Answer Block — A short, self-contained passage (typically 40-80 words) written so an AI system can lift it as a complete answer. Example: The 60-word definition at the top of a "what is RAG?" page that Perplexity quotes word for word. See: Answer Blocks.
Answer Engine — Any system that returns a direct answer rather than a list of links. Example: ChatGPT, Perplexity, Google AI Overviews, Alexa, and Siri all qualify as answer engines. See: Answer Engines Hub.
Answer Extraction — The process by which an AI system pulls a specific span of text from a source and uses it as the answer. Example: A featured snippet pulling the first sentence under an
and bolding it as the displayed answer. See: Answer Extraction Guide.
Answer Grounding — Anchoring an AI-generated response to specific source material so the answer is verifiable and citable. Strong grounding correlates with both factual accuracy and being cited. Example: Perplexity attaching numbered citations to every sentence in its response. See: Answer Grounding.
Answer-First Formatting — A content structure principle where the direct answer appears in the first 2-3 sentences, before context, history, or qualification. Example: Opening a definition page with "GEO is the practice of …" rather than "Over the past two years …". See: Answer-First Writing.
Applebot — Apple's web crawler, used for Spotlight, Siri, and Apple Intelligence retrieval. Example: The user-agent string Applebot-Extended in server logs indicates Apple Intelligence access. See: AI Crawlers Reference.
B
Bing Chat / Copilot — Microsoft's AI search assistant, integrated into Bing and Microsoft 365. Example: Asking Copilot in Edge to compare two product pages and return a recommendation. See: Copilot Optimization.
Bot Detection — Server-side identification of automated user agents, used to govern access for AI crawlers separately from human traffic. Example: Cloudflare flagging requests from ClaudeBot and routing them through a separate cache and rate limit. See: Bot Management.
C
Canonical Concept — A single, authoritative page that serves as the definitive reference for a specific concept. Example: /geo/what-is-geo is the canonical concept page for "GEO" across geodocs.dev. See: Canonical Concept Model.
Canonical URL — The preferred URL for a piece of content, used to prevent duplicate-content issues. Example: The on a print version pointing back to the main article URL. See: Canonical URLs.
ChatGPT — OpenAI's conversational AI product, including ChatGPT Search. A primary GEO measurement surface. Example: A query like "best CRM for SaaS startups" returning a synthesized answer with cited sources inside ChatGPT. See: ChatGPT Optimization.
Citation Frequency (Citation Rate) — The number of times a source is cited or referenced in AI-generated responses over a measurement window. A primary KPI in GEO programs. Example: "Domain X was cited in 27% of sampled Perplexity answers about GEO in March 2026." See: Citation Rate.
Citation Readiness — The degree to which content is structured for AI citation. Tracked in geodocs.dev frontmatter as draft, reviewed, or verified. Example: A page graduating from draft to reviewed after passing the editorial citation checklist. See: Citation Readiness.
Citation Signal — Any structured element that increases the chance of being cited: clear definitions, factual claims, schema markup, and stable URLs. Example: A Definition schema block paired with a one-sentence answer at the top of the page. See: Citation Signals.
ClaimReview — A schema.org type used to mark up fact-checks of public claims. Helps AI systems and search engines surface verified claims. Example: A fact-check article using ClaimReview to mark up "Claim: X. Rating: False." See: ClaimReview Markup.
ClaudeBot — Anthropic's web crawler for Claude. Identified via the ClaudeBot user agent. Example: ClaudeBot/1.0 requests appearing in access logs after Claude added web browsing. See: AI Crawlers Reference.
Content Cluster — A group of related pages organized around a central pillar page. Example: A pillar page on "AEO" surrounded by spokes on FAQs, schema, answer blocks, and measurement. See: Content Clusters.
Content Structure — The hierarchical organization of content using semantic HTML, headings, lists, and tables that help AI systems parse meaning. Example: Using
for sections and for definition lists rather than styled s. See: Content Structure.
Crawler — Any bot that fetches and indexes web content. Includes traditional search crawlers and AI crawlers. Example: Googlebot, GPTBot, ClaudeBot, and PerplexityBot are all crawlers. See: Crawlers Reference.
D
Definition Block — A short, structured passage that defines a single term, optimized for AI extraction. Example: A bolded term followed by a 30-word definition wrapped in a Definition schema block. See: Definition Blocks.
E
E-E-A-T — Experience, Expertise, Authoritativeness, Trustworthiness. Originally a Google quality framework, now widely cited as a proxy for AI source selection. Example: A medical article authored by a credentialed clinician with linked credentials and citations to peer-reviewed studies. See: E-E-A-T for AI Search.
Embedding — A numerical vector representation of text used for semantic search and retrieval. Example: A blog post stored as a 1,536-dimensional vector so a RAG system can find it by meaning, not just keywords. See: Embeddings Explained.
Entity — A named concept, person, organization, or thing that AI systems can identify and track. Example: Declaring entities: ["GEO", "AEO"] in frontmatter so the page is indexed under both. See: Entities.
Entity Clarity — Defining entities explicitly through consistent naming, schema markup, and clear contextual relationships. Example: Always using "Generative Engine Optimization (GEO)" on first mention rather than alternating between "GEO" and "generative search optimization." See: Entity Clarity.
Entity Coverage — The breadth and depth with which a domain covers an entity and its related sub-entities. Example: A site covering GEO, AEO, llms.txt, ai.txt, and AI Overviews has stronger entity coverage on "AI search" than one covering only GEO. See: Entity Coverage.
F
FAQPage — A schema.org type for marking up FAQ content as structured Question/Answer pairs. Example: A FAQPage block listing five Q&A pairs about llms.txt at the bottom of an article. See: FAQPage Schema.
Featured Snippet — A highlighted answer box at the top of Google search results, often called "position zero." Example: The boxed paragraph answering "what is JSON-LD" with a source link below. See: Featured Snippets.
Freshness Signal — Any signal that conveys content recency: published date, updated date, version, last-reviewed timestamp, or recently added citations. Example: The updated_at field bumped to "2026-05-01" with a visible "Last updated" line in the UI. See: Freshness Signals.
G
GEO (Generative Engine Optimization) — The practice of structuring content so AI systems can understand, retrieve, synthesize, and cite it in generated answers. Example: Restructuring a 5,000-word essay into a TL;DR, definition, and FAQ so Perplexity cites it for "what is GEO." See: What Is GEO?.
Gemini — Google's family of multimodal AI models, powering AI Overviews, AI Mode, and the Gemini app. Example: A Gemini-powered AI Overview synthesizing five sources for a "best running shoes 2026" query. See: Gemini Optimization.
Generative Engine — Any AI system that generates text responses by synthesizing information from multiple sources. Example: ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews are all generative engines. See: Generative Engines.
Google-Extended — A standalone product token Google honors in robots.txt to control whether Google may use a site's content for Gemini training and grounding. Example: User-agent: Google-Extended followed by Disallow: / to opt out of Gemini training. See: Google-Extended Reference.
GPTBot — OpenAI's web crawler for ChatGPT training and retrieval. Identified via the GPTBot user agent. Example: The line User-agent: GPTBot followed by Disallow: /private/ in robots.txt to block GPTBot from a private section. See: GPTBot Reference.
Grounding — See Answer Grounding.
H
Hub-and-Spoke — A site architecture where a central hub page links out to many related spoke pages, each linking back to the hub. Example: An /aeo/ hub linking to 30 spokes covering FAQ schema, answer blocks, and citation patterns. See: Hub-and-Spoke Architecture.
HowTo — A schema.org type for marking up step-by-step instructional content. Example: A HowTo block describing the five steps to deploy llms.txt. See: HowTo Schema.
I
Inclusion Rate — The share of relevant queries on a given AI surface where your domain is included as a source. Example: "Our inclusion rate on Perplexity for GEO queries is 31% in March 2026." See: Inclusion Rate.
Internal Linking — Linking between pages on the same domain to convey structure and authority. Example: Every glossary entry linking to its canonical concept page reinforces the hub-and-spoke structure. See: Internal Linking for AI.
J
JSON-LD — JavaScript Object Notation for Linked Data. The recommended format for implementing schema.org structured data. Example: A
Crawler — Any bot that fetches and indexes web content. Includes traditional search crawlers and AI crawlers. Example: Googlebot, GPTBot, ClaudeBot, and PerplexityBot are all crawlers. See: Crawlers Reference.
D
Definition Block — A short, structured passage that defines a single term, optimized for AI extraction. Example: A bolded term followed by a 30-word definition wrapped in a Definition schema block. See: Definition Blocks.
E
E-E-A-T — Experience, Expertise, Authoritativeness, Trustworthiness. Originally a Google quality framework, now widely cited as a proxy for AI source selection. Example: A medical article authored by a credentialed clinician with linked credentials and citations to peer-reviewed studies. See: E-E-A-T for AI Search.
Embedding — A numerical vector representation of text used for semantic search and retrieval. Example: A blog post stored as a 1,536-dimensional vector so a RAG system can find it by meaning, not just keywords. See: Embeddings Explained.
Entity — A named concept, person, organization, or thing that AI systems can identify and track. Example: Declaring entities: ["GEO", "AEO"] in frontmatter so the page is indexed under both. See: Entities.
Entity Clarity — Defining entities explicitly through consistent naming, schema markup, and clear contextual relationships. Example: Always using "Generative Engine Optimization (GEO)" on first mention rather than alternating between "GEO" and "generative search optimization." See: Entity Clarity.
Entity Coverage — The breadth and depth with which a domain covers an entity and its related sub-entities. Example: A site covering GEO, AEO, llms.txt, ai.txt, and AI Overviews has stronger entity coverage on "AI search" than one covering only GEO. See: Entity Coverage.
F
FAQPage — A schema.org type for marking up FAQ content as structured Question/Answer pairs. Example: A FAQPage block listing five Q&A pairs about llms.txt at the bottom of an article. See: FAQPage Schema.
Featured Snippet — A highlighted answer box at the top of Google search results, often called "position zero." Example: The boxed paragraph answering "what is JSON-LD" with a source link below. See: Featured Snippets.
Freshness Signal — Any signal that conveys content recency: published date, updated date, version, last-reviewed timestamp, or recently added citations. Example: The updated_at field bumped to "2026-05-01" with a visible "Last updated" line in the UI. See: Freshness Signals.
G
GEO (Generative Engine Optimization) — The practice of structuring content so AI systems can understand, retrieve, synthesize, and cite it in generated answers. Example: Restructuring a 5,000-word essay into a TL;DR, definition, and FAQ so Perplexity cites it for "what is GEO." See: What Is GEO?.
Gemini — Google's family of multimodal AI models, powering AI Overviews, AI Mode, and the Gemini app. Example: A Gemini-powered AI Overview synthesizing five sources for a "best running shoes 2026" query. See: Gemini Optimization.
Generative Engine — Any AI system that generates text responses by synthesizing information from multiple sources. Example: ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews are all generative engines. See: Generative Engines.
Google-Extended — A standalone product token Google honors in robots.txt to control whether Google may use a site's content for Gemini training and grounding. Example: User-agent: Google-Extended followed by Disallow: / to opt out of Gemini training. See: Google-Extended Reference.
GPTBot — OpenAI's web crawler for ChatGPT training and retrieval. Identified via the GPTBot user agent. Example: The line User-agent: GPTBot followed by Disallow: /private/ in robots.txt to block GPTBot from a private section. See: GPTBot Reference.
Grounding — See Answer Grounding.
H
Hub-and-Spoke — A site architecture where a central hub page links out to many related spoke pages, each linking back to the hub. Example: An /aeo/ hub linking to 30 spokes covering FAQ schema, answer blocks, and citation patterns. See: Hub-and-Spoke Architecture.
HowTo — A schema.org type for marking up step-by-step instructional content. Example: A HowTo block describing the five steps to deploy llms.txt. See: HowTo Schema.
I
Inclusion Rate — The share of relevant queries on a given AI surface where your domain is included as a source. Example: "Our inclusion rate on Perplexity for GEO queries is 31% in March 2026." See: Inclusion Rate.
Internal Linking — Linking between pages on the same domain to convey structure and authority. Example: Every glossary entry linking to its canonical concept page reinforces the hub-and-spoke structure. See: Internal Linking for AI.
J
JSON-LD — JavaScript Object Notation for Linked Data. The recommended format for implementing schema.org structured data. Example: A