Glossary
Working definitions as this site uses them — grouped by foundations, structured data and crawling, entity authority, content and measurement, and surfaces. These describe how the terms are used here, not exhaustive industry definitions.
Foundations
AI SEO — Umbrella discipline for earning visibility in AI search and AI-generated answers. This site treats AI SEO as the only top-level term; GEO, AEO, and LLMO are labels to explain below, not separate topics.
AEO — Answer Engine Optimization: structuring content to be surfaced as a direct answer.
GEO — Generative Engine Optimization: earning citations and synthesis in generative answers.
LLMO — LLM Optimization: being recognized and surfaced by large language models.
AI Visibility — How often and how favorably a brand appears across AI answers.
Answer Engine — A system that returns a synthesized answer instead of a list of links.
Generative Engine — An AI that composes responses from many sources rather than linking to them.
RAG — Retrieval-Augmented Generation: a model retrieves external documents to ground its answer.
Grounding — Anchoring an AI answer to verifiable source data to reduce hallucination.
Semantic Search — Matching on meaning and intent rather than exact keyword strings.
Structured Data and Crawling
Structured Data — Machine-readable markup that labels content so AI can parse and verify it.
JSON-LD — Google’s preferred structured-data format; JSON embedded and decoupled from HTML.
Schema.org — Shared vocabulary of types such as Product, Offer, FAQ, Review, SoftwareApplication, and VideoGame.
@graph / @id — JSON-LD constructs that connect entities and give each entity a stable identifier.
sameAs — Schema property linking an entity to authoritative profiles such as Wikidata, Wikipedia, LinkedIn, or Crunchbase.
llms.txt — Plain-text file proposed to guide AI systems to preferred content; currently not a proven enterprise ecommerce visibility lever. See Schema to Avoid for the skeptical verdict.
robots.txt — Advisory crawler instructions under RFC 9309; useful for compliant crawlers but not a security control.
Product Feed — Structured catalog file exposing product, price, availability, and inventory to platforms and AI shopping surfaces.
JSONL — JSON Lines; one JSON object per line, common in data feeds and large-scale exports.
Extractability — How easily an AI system can lift a clean, self-contained answer from a page section.
GPTBot — OpenAI’s training crawler.
OAI-SearchBot — OpenAI’s crawler for ChatGPT search visibility.
ChatGPT-User — OpenAI user agent for user-triggered fetches.
Google-Extended — Google’s control for Gemini training, independent of Search indexing.
PerplexityBot — Perplexity crawler for indexing and retrieval.
ClaudeBot — Anthropic crawler; separate from search-specific retrieval agents.
Entity Authority
Entity — A distinct thing a model recognizes: brand, product, person, organization, product generation, or market-specific offer.
Entity Home — The authoritative, machine-readable definition a brand establishes for itself.
Knowledge Graph — Structured network of entities and relationships models treat as facts.
Wikidata — Structured knowledge base that feeds knowledge graphs and multilingual entity resolution.
Digital PR — Earning third-party media mentions to build entity authority.
Earned Media — Third-party coverage the brand does not own.
Brand Mention — A reference to a brand, linked or unlinked.
E-E-A-T — Experience, Expertise, Authoritativeness, and Trustworthiness: Google’s quality signals.
Disambiguation — Resolving which entity a name refers to before an AI system can cite it accurately.
Brand Drift — When a model describes a brand or product using outdated training-cutoff information.
UGC — User-generated content such as reviews, forums, Reddit, YouTube comments, and community posts.
Content and Measurement
Answer-First — Leading a passage with the direct answer so AI can extract it cleanly.
Passage / Chunk — The self-contained unit of text an AI retrieves, extracts, or cites.
Citation — An AI answer’s attribution to a source.
Citation Decay — Turnover of cited sources over time.
Original Research — Proprietary data or benchmarks used as citable evidence.
Content Freshness — Recency signals that affect whether content remains eligible and attractive for citation.
Scaled Content Abuse — Mass-producing low-value pages; the risk is thin, manipulative velocity, not AI assistance itself.
Query Fan-Out — When an AI expands one query into many sub-queries to build an answer.
Share of Model — A brand’s share of AI answers for target topics versus competitors.
Share of Voice — How often a brand is named or recommended versus rivals.
Citation Rate — How often a brand or URL is cited when relevant to a query.
Mention Frequency — Baseline count of how often a brand appears across AI answers.
Sentiment Analysis — Whether AI describes a brand positively, neutrally, or negatively.
Hallucination — An AI stating something false, such as incorrect pricing, specs, availability, or feature support.
AI Referral Traffic — Site visits originating from AI platforms.
Prompt Set — The defined list of queries a visibility tool samples.
Surfaces, Commerce, and Governance
AI Overviews — Google’s AI-generated answer format in Search results.
AI Mode — Google’s conversational search mode.
Zero-Click — A search answered in-surface without sending a click to a source site.
ChatGPT — OpenAI assistant and AI search/product-discovery surface.
Perplexity — Citation-heavy answer engine.
Gemini — Google’s model powering AI features and assistant experiences.
Claude — Anthropic assistant.
Copilot — Microsoft assistant surfaced across Bing, Windows, and Microsoft experiences.
Alexa for Shopping — Amazon’s AI shopping assistant, unified with Rufus in 2026.
COSMO — Amazon commonsense knowledge model used to understand shopping intent.
Google Lens — Google’s visual-search tool.
Visual Search — Product discovery via images, screenshots, or camera input.
Multimodal Search — Queries combining image and text.
Shopping Graph — Google’s large product-listing dataset behind shopping and visual surfaces.
ACP — Agentic Commerce Protocol associated with OpenAI/Stripe commerce workflows.
EU AI Act Article 50 — EU transparency obligations for certain AI outputs, enforceable beginning August 2, 2026.