HTML&HTML / AI SEARCH REFERENCE
AI Search & Visibility Reference Glossary
A source-aware reference for GEO, AEO, LLMO, AAO, RAG, E-E-A-T, llms.txt, sitemaps and AI search visibility—separating official platform guidance from industry terminology.
AI Search Visibility
INDUSTRY_TERMHow discoverable, understandable, source-eligible and clickable a website is across generative AI search experiences.
Definition
How discoverable, understandable, source-eligible and clickable a website is across generative AI search experiences.
What does Google say?
Google now publishes official guidance for succeeding in generative AI Search features and provides dedicated Search Console visibility reporting for AI Overviews, AI Mode and related experiences.
What it means for AI systems
AI search systems retrieve candidate pages, evaluate relevant passages and present links or citations when useful.
Why it matters to a website owner
The commercial opportunity begins before a website visit: a brand must first be eligible to appear when a customer asks AI what to choose.
How HTML&HTML measures it
HTML&HTML combines crawler access, technical eligibility, answer extractability, entity integrity, source-worthiness and conversion-path evidence.
What it does not guarantee
No tool can guarantee an AI recommendation, citation, ranking, click, customer or revenue outcome.
Official / primary sources
GEO — Generative Engine Optimization
INDUSTRY_TERMA market term for work intended to improve visibility and source eligibility in generative AI search experiences.
Definition
A market term for work intended to improve visibility and source eligibility in generative AI search experiences.
What does Google say?
Google explicitly recognizes the term GEO, but says that from Google Search's perspective optimizing for generative AI search is still SEO and there are no special technical requirements for AI Overviews or AI Mode.
What it means for AI systems
Useful GEO work focuses on eligibility, retrieval, original information, trust, clear entities, usable links and measurable referrals rather than unsupported AI ranking hacks.
Why it matters to a website owner
A business wants to be considered when a customer asks an AI system for brands, products, providers or sources.
How HTML&HTML measures it
HTML&HTML maps GEO readiness across AI/search crawlability, answer extractability, information gain signals, entity graph integrity, trust and citation/referral surfaces.
What it does not guarantee
GEO is not a Google-approved shortcut and it cannot guarantee inclusion in AI-generated answers.
Official / primary sources
AEO — Answer Engine Optimization
INDUSTRY_TERMA market term for structuring content so answer-oriented systems can more easily identify clear, useful responses.
Definition
A market term for structuring content so answer-oriented systems can more easily identify clear, useful responses.
What does Google say?
Google recognizes AEO as a term used in the market, while stating that success in generative AI Search still relies on Search fundamentals, helpful content and technical eligibility.
What it means for AI systems
Answer extractability improves when headings, paragraphs, definitions, evidence and page intent make the answer clear without losing context.
Why it matters to a website owner
If a customer asks a direct question, the site should make its useful answer easy to identify and verify.
How HTML&HTML measures it
HTML&HTML evaluates answer-first structure, heading-to-answer continuity, self-contained sections, evidence and structured content signals.
What it does not guarantee
There is no special AEO markup required by Google Search and no fixed word-count formula guarantees extraction.
Official / primary sources
LLMO — Large Language Model Optimization
INDUSTRY_TERMA market term for making web information easier for large language model and retrieval systems to resolve, understand and use correctly.
Definition
A market term for making web information easier for large language model and retrieval systems to resolve, understand and use correctly.
What does Google say?
Google does not define LLMO as a separate Google Search discipline. Its official AI Search guidance emphasizes accessible text, crawlability, unique useful content, internal links and matching structured data.
What it means for AI systems
Clear entities, stable canonical URLs, semantic HTML, internal links and coherent sections reduce ambiguity for retrieval and language-model workflows.
Why it matters to a website owner
The practical goal is not to teach an LLM your site; it is to remove ambiguity that makes your offer harder to retrieve or interpret.
How HTML&HTML measures it
HTML&HTML uses entity integrity, semantic structure, internal-link alignment, canonical consistency, language consistency and knowledge-surface checks.
What it does not guarantee
LLMO is not an official Google ranking factor or a guarantee that a model will mention a brand.
Official / primary sources
AAO — AI Agent Optimization
EMERGING_TERMAn emerging term for making websites easier for AI agents to understand and interact with safely.
Definition
An emerging term for making websites easier for AI agents to understand and interact with safely.
What does Google say?
Google's 2026 generative AI optimization guidance includes initial guidance related to AI agents and describes the area as quickly emerging and evolving.
What it means for AI systems
Agents benefit from semantic controls, accessible names, explicit states, clear forms, stable actions and machine-readable interfaces where appropriate.
Why it matters to a website owner
As delegated AI actions grow, businesses need critical user journeys to remain understandable, safe and testable for both people and agents.
How HTML&HTML measures it
HTML&HTML checks accessibility semantics, forms, actionable controls, OpenAPI/agent discovery signals and public interaction surfaces.
What it does not guarantee
AAO is an emerging market term, not a stable universal standard and not a promise that an external agent can complete every action.
Official / primary sources
RAG — Retrieval-Augmented Generation
TECHNIQUEA technique that retrieves relevant external information and uses it to ground a generated response.
Definition
A technique that retrieves relevant external information and uses it to ground a generated response.
What does Google say?
Google's 2026 generative AI Search guide explicitly describes retrieval-augmented generation (RAG), also called grounding, as a technique used with Search ranking systems to retrieve relevant up-to-date pages before generating responses.
What it means for AI systems
Retrieval works better when important information is textual, coherent, uniquely useful, linked and accessible at stable URLs.
Why it matters to a website owner
The site owner wants useful facts, products and expertise to be retrievable in the right context when an AI system answers a relevant question.
How HTML&HTML measures it
HTML&HTML evaluates oversized content blocks, context loss, duplication, self-contained sections, discovery paths and semantic integrity.
What it does not guarantee
RAG is a retrieval technique, not a standalone Google ranking factor and not a guarantee that a particular page will be cited.
Official / primary sources
E-E-A-T
GOOGLE_QUALITY_CONCEPTExperience, Expertise, Authoritativeness and Trustworthiness: a quality-evaluation concept used in Google's guidance and quality rater framework.
Definition
Experience, Expertise, Authoritativeness and Trustworthiness: a quality-evaluation concept used in Google's guidance and quality rater framework.
What does Google say?
Google says E-E-A-T can help creators self-assess content quality. Search quality raters use it to evaluate results, but rater data is not used directly as a ranking algorithm score.
What it means for AI systems
Clear authorship, first-hand experience, sourcing, organization identity, editorial accountability and factual accuracy make content easier to trust and verify.
Why it matters to a website owner
For commercial and high-stakes topics, visitors and systems both need evidence of who is responsible for the information and why it should be trusted.
How HTML&HTML measures it
HTML&HTML observes authorship, About/Contact/Privacy presence, organization/person schema, citations, freshness and trust signals; it does not claim to measure Google's internal E-E-A-T score.
What it does not guarantee
E-E-A-T is not a single public numeric Google ranking factor that third-party tools can read.
Official / primary sources
llms.txt
PROPOSALA community proposal for publishing a concise machine-readable directory that points language-model tools and agents to important site resources.
Definition
A community proposal for publishing a concise machine-readable directory that points language-model tools and agents to important site resources.
What does Google say?
Google Search explicitly says it does not use llms.txt or other special AI text files to determine inclusion in Google Search or its generative AI features.
What it means for AI systems
The llms.txt v2 proposal adds standard discovery links such as rel=describedby and Markdown alternates; some documentation and coding-agent ecosystems use the proposal.
Why it matters to a website owner
It can be a useful optional knowledge-directory surface, especially for documentation-heavy sites, but it should complement—not replace—crawlable HTML, sitemaps and internal links.
How HTML&HTML measures it
HTML&HTML validates H1, summary structure, link reachability, v2 discovery relationships and Markdown alternates while labeling llms.txt as a proposal.
What it does not guarantee
llms.txt is not an IETF/W3C standard, Google ranking requirement, AI recommendation guarantee or replacement for robots.txt.
Official / primary sources
XML Sitemap
OFFICIAL_PROTOCOLA machine-readable list of canonical URLs that helps search engines discover the pages a site wants crawled and considered.
Definition
A machine-readable list of canonical URLs that helps search engines discover the pages a site wants crawled and considered.
What does Google say?
Google recommends standard sitemap formats, root-level placement where practical, absolute URLs, canonical URLs and the protocol limits of 50MB or 50,000 URLs per sitemap.
What it means for AI systems
For Google AI features, pages still rely on the Search index; sitemap hygiene therefore remains part of the discovery foundation rather than a special AI-only file.
Why it matters to a website owner
A correct sitemap reduces discovery ambiguity and helps site owners monitor important URL sets in Search Console.
How HTML&HTML measures it
HTML&HTML checks sitemap discovery, URL validity, canonical consistency, status codes, hreflang relationships and orphan/discovery-path signals.
What it does not guarantee
A sitemap does not guarantee crawling, indexing, ranking, AI citation or traffic.
Official / primary sources
Query fan-out
GOOGLE_DESCRIBED_TECHNIQUEA technique where a model issues multiple related searches to gather enough context for a complex question.
Definition
A technique where a model issues multiple related searches to gather enough context for a complex question.
What does Google say?
Google's generative AI Search guide explicitly describes query fan-out as concurrent related queries used to fetch additional relevant results.
What it means for AI systems
It means a page may be discovered through a narrower sub-question rather than only through the user's original wording.
Why it matters to a website owner
Deep, specific and uniquely useful pages can create more entry points into AI-assisted research journeys.
How HTML&HTML measures it
HTML&HTML uses intent, internal-link, information-gain and answer-extractability signals to identify whether important subtopics have clear, retrievable surfaces.
What it does not guarantee
Query fan-out does not mean creating thin pages for every keyword variation.
Official / primary sources
Source eligibility
HTMLANDHTML_DECISION_TERMHTML&HTML's decision term for whether a page clears observable website-side barriers to being retrieved, evaluated and linked as a source.
Definition
HTML&HTML's decision term for whether a page clears observable website-side barriers to being retrieved, evaluated and linked as a source.
What does Google say?
Google says pages shown as supporting links in AI Overviews or AI Mode must be indexed and eligible to appear in Search with a snippet; meeting requirements still does not guarantee serving.
What it means for AI systems
Eligibility is a prerequisite, not an outcome: access, indexability, usable text, relevance and trust all matter before an external system can choose a source.
Why it matters to a website owner
The site owner needs to know whether preventable website-side failures are removing the brand from consideration before the recommendation decision happens.
How HTML&HTML measures it
HTML&HTML combines deterministic access/indexability evidence with non-scoring intelligence context without claiming access to external ranking systems.
What it does not guarantee
Eligibility is not a probability score for being recommended.
Official / primary sources