← Back to e21designs
Glossary · the language of AI search

GEO Glossary.

The category language used inside this firm. Definitions for the terms a CMO or CFO new to AI search needs to understand · GEO, AEO, citation density, llms.txt, N-E-E-A-T, RAG, schema.org, Speakable, SSR. Each definition is short, plain, and links to the deeper resource. Updated quarterly.

Terms · 17 Last updated · 14 May 2026 Canonical reference · schema.org

Answer Engine Optimization (AEO)

The discipline of engineering a brand into the single answer that appears when a buyer asks a direct question of an AI engine. AEO sits inside the broader GEO discipline; it focuses on the answer surface specifically. See: AEO vs SEO vs GEO · where revenue lives in 2026.

Citation density

The frequency with which a brand or entity appears in the corpora AI engines trust · news, journals, books, Wikipedia, podcast transcripts, GitHub, peer-reviewed papers. Citation density is the single highest-correlated predictor of whether an AI engine returns a brand name in a buyer answer.

Citation share

The percentage of prompts in a defined category where a brand is returned as a named entity by an AI engine. Measured per-engine (ChatGPT, Claude, Perplexity, Gemini) and aggregated. The GEO equivalent of Share of Voice in traditional marketing.

Compounding (citation)

The flywheel by which existing citations attract new citations. AI engines weight prior recognition · once a brand is in the corpus, subsequent crawls find it referenced and reinforce the entity. The first nine months of GEO work disproportionately seeds the next nine years.

DefinedTerm schema

A schema.org type that explicitly marks up a concept or term being defined on a page. Critical for category-defining content (definition pages for GEO, AEO, llms.txt, etc.) · DefinedTerm helps AI engines lift the definition verbatim into answers.

Entity graph

The web of connections AI engines build between entities (people, brands, products, places). A brand's entity graph is the set of all entities it is associated with, and the strength of those associations. Strong entity graphs cite cleanly; weak ones get confused with similarly-named entities or missed entirely.

E-E-A-T

Google's quality framework: Experience, Expertise, Authoritativeness, Trust. Originally for traditional search rankings; now a foundational input to AI search citation as well. Sites with strong E-E-A-T get cited; sites with thin authorial transparency do not.

FAQPage schema

A schema.org type that marks up question-and-answer pairs on a page. AI engines lift FAQ-schema-marked content disproportionately because the structure exactly matches how buyer prompts are phrased.

Generative Engine Optimization (GEO)

The discipline of engineering brand visibility, citation share and answer position inside generative AI engines (ChatGPT, Claude, Perplexity, Gemini, Copilot, Google AI Overviews). GEO sits alongside SEO (search-result rankings) and AEO (single-answer optimisation) as the third distinct discipline of the AI-mediated economy. See: canonical definition.

HowTo schema

A schema.org type that marks up step-by-step procedures. High-leverage for service firms because AI engines often respond to "how do I X" prompts by lifting HowTo-marked content verbatim.

llms.txt

A plain-text Markdown file at the root of a website (like robots.txt for crawlers, but for AI engines) that tells generative models what the firm is, what it does, and which canonical pages are authoritative. Co-authored format · ships at e21designs.com/llms.txt. See: llms.txt explained.

LLM (Large Language Model)

A type of generative AI model trained on vast text corpora · GPT-4 (OpenAI), Claude (Anthropic), Gemini (Google), Llama (Meta), Mistral, DeepSeek. LLMs power the AI engines that have made GEO necessary.

N-E-E-A-T

e21designs's extension of Google's E-E-A-T framework for the AI-mediated era: Namedness, Experience, Expertise, Authoritativeness, Trust. Namedness is added because AI engines weight name-recognition (entity-clarity) more heavily than search engines ever did.

RAG (Retrieval-Augmented Generation)

The architecture by which modern AI engines combine pre-trained knowledge with real-time retrieval from the open web. Most current AI search systems are RAG-based. GEO is the discipline of being retrievable inside RAG pipelines.

Schema.org

Schema.org · the open vocabulary for structured data on the web. Maintained by Google, Microsoft, Yahoo and Yandex. The substrate of all GEO-readable evidence · Article, Organization, Person, Service, Product, FAQ, HowTo, Speakable, Review, BreadcrumbList, and more.

Speakable schema

A schema.org type that marks up sections of a page suitable for spoken-word output. Originally for voice assistants; now a strong citation hint to AI engines that the marked text is the canonical answer to a likely question.

SSR (Server-Side Rendering)

The technique of generating HTML on the server before sending it to the browser, as opposed to building it in JavaScript in the browser (CSR). Critical for GEO: AI crawlers (GPTBot, ClaudeBot, PerplexityBot) do not reliably execute JavaScript. Content and schema that exist only after JS execution are invisible to most AI engines.

External references we trust

Once the language is clear, the field report makes more sense

Apply · ten partnerships a year, firm-wide

If the language now makes sense, see how AI sees your firm.

A glossary makes the category legible. The next step is seeing where your firm sits in that category · the free AI Visibility Read lands within five working days, reviewed personally by Dr. Avinaash.

Apply for the free read →