Glossary
The language of AI search.
Clear definitions of the terms behind getting cited and recommended by AI — written to be quoted.
- AEOAnswer Engine Optimization
- The practice of structuring content and reputation so that AI answer engines cite a brand as a source inside their generated answers. Where SEO competes for a ranked link, AEO competes to be the sentence the AI quotes.Compare: GEO, SEO.
- GEOGenerative Engine Optimization
- The practice of shaping how generative AI understands a brand so the model recommends it by name when a user asks for a suggestion. AEO makes you credible as a source; GEO makes you the pick.Compare: AEO.
- Answer enginee.g. ChatGPT, Perplexity, Claude, Gemini
- A system that responds to a query with a single synthesized answer rather than a list of links. It reads many sources, summarizes them, and usually credits a few. The answer — not a results page — is the destination.See also: AI Overviews.
- AI OverviewsGoogle's generative answer box
- The AI-generated summary Google places above traditional search results for many queries. It draws on indexed pages and cites a small set of sources, compressing the classic ten blue links into one answer with a few links.Reaches over a billion users per month.
- Mention rateThe core AEO/GEO metric
- How often an AI engine names a brand in its answer, measured across repeated samples of the same question. Because AI answers vary run to run, mention rate is expressed as a percentage over many samples, not a single yes/no.the ageo's primary reported number.
- AI citationA credited source
- A reference an AI answer attributes to a specific source — a linked domain, a named brand, or a quoted passage. Being cited is the AEO win condition: it places your name inside the answer and, often, drives the click.Driven more by brand mentions than backlinks.
- llms.txtA standard for AI crawlers
- A plain-text file at a site's root that gives AI systems a concise, structured summary of what the site offers and how to represent it — analogous to robots.txt, but for meaning rather than access. A missing llms.txt is a common, easily fixed visibility gap.Live at theageo.co/llms.txt.
- EntityA thing a model recognizes
- A distinct person, company, product, or concept that an AI model treats as a known node in its understanding of the world. Brands that exist as confirmed entities — corroborated across multiple independent sources — are far easier for AI to recommend with confidence.Reinforced by schema sameAs and off-site mentions.
- RAGRetrieval-Augmented Generation
- A technique where an AI model retrieves fresh documents from the web or a database at query time and generates its answer from them, rather than relying only on training data. RAG is why current, well-structured, crawlable content can be cited even by models trained before it existed.Powers most live answer engines.
- Structured dataSchema.org / JSON-LD
- Machine-readable markup that labels what content means — that this is an Organization, that a FAQPage, that a HowTo. It removes ambiguity for AI systems parsing a page, making facts safer to extract and cite.See also: entity.
- E-E-A-TExperience, Expertise, Authoritativeness, Trust
- Google's framework for content quality, increasingly mirrored by how AI systems weight sources. Named authors, credentials, original data, reviews, and legal transparency all raise the odds that a model treats a site as trustworthy enough to quote.Built in the Build step of our method.