SEO & AI Search Glossary

Short, practical definitions of the terms used across search engine optimization and AI-search visibility work. Written to be precise enough to cite and plain enough to act on.

Last reviewed: July 2026

AI search & generative engines
Structured data
Content & on-page SEO
Technical SEO

AI search & generative engines

AEO (Answer Engine Optimization): The practice of structuring content so answer engines — Google AI Overviews, ChatGPT, Perplexity, Copilot — can extract and present it as a direct answer. AEO focuses on question-shaped headings, self-contained answer paragraphs, FAQ markup, and machine-readable facts rather than only classic ranking signals.
GEO (Generative Engine Optimization): Optimizing a site so generative AI systems mention and cite it when composing answers. Where classic SEO targets a ranked list of links, GEO targets inclusion in the generated text itself: being retrievable, quotable, and attributable — with clear entities, verifiable facts, and consistent descriptions across the web.
LLMO (Large Language Model Optimization): An umbrella term for making a brand legible to large language models: allowing AI crawlers, publishing an llms.txt file, keeping entity descriptions consistent, and earning mentions in sources models retrieve from. Often used interchangeably with GEO.
AI Overviews: Google's AI-generated summary block shown above classic results for many queries. It synthesizes an answer from multiple sources and links a subset of them, which makes being cited in the overview a distinct visibility channel from ranking #1 below it.
AI crawler: A bot that fetches web pages for AI systems — GPTBot and OAI-SearchBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot, Google-Extended. Each respects robots.txt rules under its own user-agent, so a site chooses per bot whether to be available for AI answers and training.
llms.txt: A proposed plain-text file at /llms.txt that gives AI assistants a curated index of a site: what the product is, key pages, pricing, and canonical facts. It complements robots.txt — robots.txt says what bots may fetch, llms.txt says what they should read first.
AI citation: A link or named reference to a source inside an AI-generated answer. Citations are earned passage by passage: systems favor self-contained paragraphs that answer one question with verifiable facts, tables with consistent columns, and pages whose claims match other sources.
Brand mentions in AI answers: How often AI assistants name a brand when answering category questions ("best X for Y"). Tracked by running a fixed set of prompts across ChatGPT, Perplexity, and Gemini over time — the AI-era equivalent of rank tracking.
Entity SEO: Optimizing for how search and AI systems model things (brands, people, products) rather than strings of keywords. In practice: consistent naming, an About page with Organization markup, sameAs links to official profiles, and a distinct description that separates the entity from same-named ones.
Knowledge graph: A database of entities and their relationships that search engines use to answer fact questions directly. Getting a brand represented correctly — right category, founder, pricing — depends on consistent structured data and corroborating third-party sources.
RAG (Retrieval-Augmented Generation): The architecture behind most AI search: the system first retrieves relevant documents, then generates an answer grounded in them. For site owners the implication is direct — content must win retrieval (be findable and topically clear) before it can be quoted in the generated answer.

Structured data

Structured data (schema markup): Machine-readable annotations, usually from the Schema.org vocabulary, that state what a page is about: an article, a product, an organization, a FAQ. Search engines use it for rich results; AI systems use it to extract facts with less guessing.
JSON-LD: The recommended format for structured data: a JSON block in a <script type="application/ld+json"> tag, separate from the visible HTML. Connected entities are best expressed as one @graph so parsers see organization, page, author, and FAQ as a single linked structure.
FAQPage schema: Markup for question-and-answer pairs on a page. Google now shows FAQ rich results only for a narrow set of authoritative sites, but the markup remains valuable for AI systems, which extract Q&A pairs as ready-made answer passages.
sameAs: A Schema.org property listing the official profiles of an entity — social accounts, directories, registries. It is a primary disambiguation signal: it tells machines which external pages describe this exact organization and not a same-named one.
Organization markup: The Schema.org node describing a company: name, legal name, founder, founding date, address, logo, contact, sameAs. The more unique verifiable attributes it carries, the harder it is for search and AI systems to confuse the company with another of the same name.

Content & on-page SEO

E-E-A-T: Experience, Expertise, Authoritativeness, Trustworthiness — the quality dimensions from Google's Search Quality Evaluator Guidelines. Signaled in practice by named authors with real profiles, first-hand detail, cited primary sources, and accurate, verifiable claims.
Topical authority: The credibility a site earns by covering a subject completely and coherently rather than in one-off posts. Built with content clusters: a hub page for the core topic and interlinked articles answering every adjacent question.
Content cluster: A group of articles covering one topic from multiple angles, linked to a shared hub page. Clusters concentrate relevance and internal links, helping both classic rankings and AI retrieval, which favors sites that fully cover a topic.
Keyword cannibalization: When several pages on one site target the same query and split its ranking potential — the engine cannot pick a clear winner, so all of them underperform. Fixed by consolidating duplicates with 301 redirects to one canonical page or re-targeting the extras to distinct queries.
Internal linking: Links between pages of the same site. They distribute authority, define topic structure, and give crawlers (classic and AI) the path to every page. Descriptive anchor text matters: it tells the machine what the target page answers.
Content refresh: Updating an existing article — facts, examples, internal links, structure — instead of publishing a new one. Counteracts content decay, the gradual traffic loss as information ages and competitors publish fresher coverage.
Programmatic SEO: Generating many pages from a data set and templates — locations, integrations, comparisons. It works when every page carries genuinely distinct, useful data, and collapses into scaled content abuse (a Google spam policy violation) when pages are thin permutations of the same text.

Technical SEO

Canonical URL: The single URL a site declares as the authoritative version of a page when duplicates or variants exist, via the rel="canonical" tag. Consolidates ranking signals and prevents variants from competing with each other.
Crawl budget: The number of URLs a crawler will fetch from a site in a period, driven by server capacity and perceived value. Wasted by duplicate, parameterized, and broken URLs; protected by clean internal links, accurate sitemaps, and fast responses.
XML sitemap: A machine-readable list of a site's canonical URLs with optional last-modified dates, submitted to search engines. It does not guarantee indexing — it tells crawlers what exists and what changed, which matters most for new and frequently updated sites.
IndexNow: An open protocol for pinging search engines the moment a URL is added, updated, or deleted, instead of waiting for a recrawl. Supported by Bing and Yandex among others; Google uses its own systems.
Core Web Vitals: Google's field metrics for user experience: LCP (loading), INP (interactivity), CLS (visual stability). Measured on real visits; passing thresholds are a lightweight ranking factor and a strong UX proxy.
SERP: Search Engine Results Page — everything shown for a query: classic links, AI Overviews, featured snippets, People Also Ask, local packs, shopping units. Modern visibility work targets the whole page, not only the ten blue links.
hreflang: Annotations that declare language and regional variants of a page so engines serve the right version to the right audience. Every variant must reference all others plus an x-default, or the signals are ignored.