GEO Glossary

Plain-English definitions for every term in AI search, GEO, and brand visibility measurement.

AI Crawler: A bot that fetches web content for AI systems — either to build training corpora (GPTBot, ClaudeBot) or to ground live answers (PerplexityBot, ChatGPT-User). Blocking them trades visibility for control.
AI Search: Search experiences where the primary result is a generated answer rather than a ranked list of links — ChatGPT, Perplexity, Google AI Overviews and AI Mode, Gemini, and Claude among them.
AI Visibility: The degree to which AI engines mention, recommend, and accurately describe a brand when answering relevant questions — the AI-era equivalent of search visibility.
AI Visibility Score: A 0–100 composite metric summarizing how often a brand appears in AI engine answers across a panel of category-relevant prompts, typically weighted by prominence and recency.
Answer Engine: A system that responds to a query with one synthesized, conversational answer — often with citations — rather than a ranked list of web pages.
Answer Engine Optimization (AEO): Optimizing content to be selected as the direct answer to a question — whether by a featured snippet, a voice assistant, or a generative AI engine. Largely synonymous with GEO in AI-search contexts.
Brand Mention: Any instance of a brand being named in an AI answer — as a recommendation, a comparison point, or a passing reference. The atomic unit of AI visibility measurement.
Brand Sentiment (AI Answers): The favorability of how AI engines frame a brand when they mention it — enthusiastic recommendation, neutral listing, or mention-with-caveats. Two brands with equal mention rates can have very different sentiment.
Citation (AI Search): A source link an AI engine attaches to part of its answer, crediting the page the claim was drawn from. Being cited puts a brand's own material inside the answer.
Citation Rate: How often a domain appears as a cited source in AI answers for a tracked prompt set — distinct from mention rate, which counts the brand being named in the answer text.
Competitor Gap (AI Visibility): The difference between a brand's AI visibility and a competitor's, measured on the same prompt panel — the clearest signal of who is winning a category's AI answers.
Context Window: The maximum amount of text a model can consider at once — prompt, retrieved sources, and its own output combined. Retrieval must fit within it, so engines summarize and select aggressively.
Conversational Search: Search conducted as a multi-turn dialogue, where context accumulates and follow-ups refine the answer — so a brand can be introduced, compared, and eliminated across several turns of one conversation.
Digital PR: Earning coverage, reviews, and citations on authoritative third-party sites. Disproportionately powerful for GEO because models weight independent corroboration far above self-description.
E-E-A-T: Google's content-quality framework — Experience, Expertise, Authoritativeness, Trustworthiness. AI engines trained and grounded on the same web reward the same signals when choosing what to cite.
Embedding: A numeric vector that represents a piece of text's meaning, letting machines compare content by semantic similarity rather than exact words — the machinery beneath retrieval and semantic search.
Entity SEO: Optimizing for entities rather than keywords: establishing a brand as a distinct, consistently described 'thing' — with the same name, facts, and relationships everywhere machines look.
Fine-Tuning: Additional training that adapts a pretrained model to narrower data or behavior — distinct from pretraining (broad knowledge) and retrieval (live facts at answer time).
Flare Score: Brandflare's headline 0–100 score, combining five weighted measurements — presence, prominence, accuracy, reach, and sentiment — of how AI engines present a brand.
Generative Engine Optimization (GEO): The practice of improving how often, how prominently, and how accurately AI engines such as ChatGPT, Gemini, Perplexity, and Claude mention a brand in their generated answers.
Google AI Overviews: Google's AI-generated answer summaries shown above the traditional results, synthesized from multiple sources with links — formerly the Search Generative Experience (SGE).
Grounding: Constraining a model's answer to information retrieved from real, current sources — reducing hallucinations and making the choice of cited sources decisive for brand visibility.
Hallucination (AI): A confident, fluent statement by an AI model that is factually wrong — such as inventing a product, misstating a price, or attributing the wrong founder to a company.
Inference: The act of a trained model generating output for a given input — everything that happens between a user's question and the finished answer, including any retrieval and reasoning steps.
Knowledge Cutoff: The date at which a model's training data ends. Without live retrieval, everything the model says about a brand reflects the world as of that date — rebrands, launches, and price changes since are invisible.
Knowledge Graph: A database of entities and the relationships between them — like Google's Knowledge Graph or Wikidata. Presence in major knowledge graphs anchors a brand's identity for search and AI systems alike.
Large Language Model (LLM): A neural network trained on vast amounts of text to predict and generate language — the technology behind ChatGPT, Gemini, Claude, and every AI answer that mentions (or omits) your brand.
LLM-as-Judge: Using a strong language model to evaluate other AI outputs at scale — in GEO, judging whether an answer mentions a brand, how favorably, and whether its claims match verified facts.
llms.txt: A proposed standard: a markdown file at /llms.txt that gives LLM crawlers a curated summary of a site's most important pages, analogous to robots.txt or a sitemap but written for AI consumption.
Mention Rate: The fraction of AI answers, across a tracked question set, in which a brand is named at least once — counted only on questions that do not name the brand, because an engine naming you when asked about you measures nothing.
Prompt: The input a user gives an AI engine. In GEO, tracked prompts play the role keywords play in SEO: the finite, repeatable queries against which visibility is measured.
Prompt Tracking: Repeatedly running a fixed panel of category-relevant prompts against AI engines on a schedule, recording which brands are mentioned — the AI equivalent of keyword rank tracking.
Query Fan-Out: An AI engine's expansion of a single user question into multiple background sub-queries whose results are synthesized into one answer — meaning a brand can enter an answer via searches the user never typed.
Retrieval-Augmented Generation (RAG): An architecture where the model first retrieves relevant documents (often via live web search) and then generates its answer from them — the reason current, crawlable content can enter AI answers within hours.
Schema Markup: The practical application of Schema.org vocabulary to a site's pages — Organization, Product, FAQPage, Article — so search and AI systems can extract entities and facts without inference.
Share of Voice (SOV) in AI Answers: The percentage of all brand mentions in a category's AI answers that belong to a given brand. Brandflare counts this only on questions that do not name the brand, so being asked about directly cannot inflate it.
Structured Data: Machine-readable annotations — usually Schema.org vocabulary in JSON-LD — that state a page's facts explicitly (organization, product, price, FAQ), removing ambiguity for crawlers and AI systems.
Sweep (AI Visibility Tracking): One complete measurement pass in AI visibility tracking: every tracked prompt is run against every monitored engine, and the answers are scored and stored as a dated snapshot for trend analysis.
Training Data: The corpus of text a language model learns from — web pages, reference works, forums, code, licensed archives. A brand's footprint in this corpus determines what the model 'knows' about it by heart.
Vector Search: Retrieval that ranks content by the semantic similarity of embeddings instead of keyword overlap — so a page can be found for questions that share none of its exact words.
Zero-Click Search: A search that ends without a click to any website because the user's need is met on the results surface itself — the default outcome of AI answers, which makes being in the answer the new imperative.

GEO Glossary

Plain-English definitions for every term in AI search, GEO, and brand visibility measurement.

A bot that fetches web content for AI systems — either to build training corpora (GPTBot, ClaudeBot) or to ground live answers (PerplexityBot, ChatGPT-User). Blocking them trades visibility for control.

AI Search

Search experiences where the primary result is a generated answer rather than a ranked list of links — ChatGPT, Perplexity, Google AI Overviews and AI Mode, Gemini, and Claude among them.

AI Visibility

The degree to which AI engines mention, recommend, and accurately describe a brand when answering relevant questions — the AI-era equivalent of search visibility.

AI Visibility Score

A 0–100 composite metric summarizing how often a brand appears in AI engine answers across a panel of category-relevant prompts, typically weighted by prominence and recency.

Answer Engine

A system that responds to a query with one synthesized, conversational answer — often with citations — rather than a ranked list of web pages.

Answer Engine Optimization (AEO)

Optimizing content to be selected as the direct answer to a question — whether by a featured snippet, a voice assistant, or a generative AI engine. Largely synonymous with GEO in AI-search contexts.

Brand Mention

Any instance of a brand being named in an AI answer — as a recommendation, a comparison point, or a passing reference. The atomic unit of AI visibility measurement.

Brand Sentiment (AI Answers)

The favorability of how AI engines frame a brand when they mention it — enthusiastic recommendation, neutral listing, or mention-with-caveats. Two brands with equal mention rates can have very different sentiment.

Citation (AI Search)

A source link an AI engine attaches to part of its answer, crediting the page the claim was drawn from. Being cited puts a brand's own material inside the answer.

Citation Rate

How often a domain appears as a cited source in AI answers for a tracked prompt set — distinct from mention rate, which counts the brand being named in the answer text.

Competitor Gap (AI Visibility)

The difference between a brand's AI visibility and a competitor's, measured on the same prompt panel — the clearest signal of who is winning a category's AI answers.

Context Window

The maximum amount of text a model can consider at once — prompt, retrieved sources, and its own output combined. Retrieval must fit within it, so engines summarize and select aggressively.

Conversational Search

Search conducted as a multi-turn dialogue, where context accumulates and follow-ups refine the answer — so a brand can be introduced, compared, and eliminated across several turns of one conversation.

Digital PR

Earning coverage, reviews, and citations on authoritative third-party sites. Disproportionately powerful for GEO because models weight independent corroboration far above self-description.

E-E-A-T

Google's content-quality framework — Experience, Expertise, Authoritativeness, Trustworthiness. AI engines trained and grounded on the same web reward the same signals when choosing what to cite.

Embedding

A numeric vector that represents a piece of text's meaning, letting machines compare content by semantic similarity rather than exact words — the machinery beneath retrieval and semantic search.

Entity SEO

Optimizing for entities rather than keywords: establishing a brand as a distinct, consistently described 'thing' — with the same name, facts, and relationships everywhere machines look.

Fine-Tuning

Additional training that adapts a pretrained model to narrower data or behavior — distinct from pretraining (broad knowledge) and retrieval (live facts at answer time).

Flare Score

Brandflare's headline 0–100 score, combining five weighted measurements — presence, prominence, accuracy, reach, and sentiment — of how AI engines present a brand.

Generative Engine Optimization (GEO)

The practice of improving how often, how prominently, and how accurately AI engines such as ChatGPT, Gemini, Perplexity, and Claude mention a brand in their generated answers.

Google AI Overviews

Google's AI-generated answer summaries shown above the traditional results, synthesized from multiple sources with links — formerly the Search Generative Experience (SGE).

Grounding

Constraining a model's answer to information retrieved from real, current sources — reducing hallucinations and making the choice of cited sources decisive for brand visibility.

Hallucination (AI)

A confident, fluent statement by an AI model that is factually wrong — such as inventing a product, misstating a price, or attributing the wrong founder to a company.

Inference

The act of a trained model generating output for a given input — everything that happens between a user's question and the finished answer, including any retrieval and reasoning steps.

Knowledge Cutoff

The date at which a model's training data ends. Without live retrieval, everything the model says about a brand reflects the world as of that date — rebrands, launches, and price changes since are invisible.

Knowledge Graph

A database of entities and the relationships between them — like Google's Knowledge Graph or Wikidata. Presence in major knowledge graphs anchors a brand's identity for search and AI systems alike.

Large Language Model (LLM)

A neural network trained on vast amounts of text to predict and generate language — the technology behind ChatGPT, Gemini, Claude, and every AI answer that mentions (or omits) your brand.

LLM-as-Judge

Using a strong language model to evaluate other AI outputs at scale — in GEO, judging whether an answer mentions a brand, how favorably, and whether its claims match verified facts.

llms.txt

A proposed standard: a markdown file at /llms.txt that gives LLM crawlers a curated summary of a site's most important pages, analogous to robots.txt or a sitemap but written for AI consumption.

Mention Rate

The fraction of AI answers, across a tracked question set, in which a brand is named at least once — counted only on questions that do not name the brand, because an engine naming you when asked about you measures nothing.

Prompt

The input a user gives an AI engine. In GEO, tracked prompts play the role keywords play in SEO: the finite, repeatable queries against which visibility is measured.

Prompt Tracking

Repeatedly running a fixed panel of category-relevant prompts against AI engines on a schedule, recording which brands are mentioned — the AI equivalent of keyword rank tracking.

Query Fan-Out

An AI engine's expansion of a single user question into multiple background sub-queries whose results are synthesized into one answer — meaning a brand can enter an answer via searches the user never typed.

Retrieval-Augmented Generation (RAG)

An architecture where the model first retrieves relevant documents (often via live web search) and then generates its answer from them — the reason current, crawlable content can enter AI answers within hours.

Schema Markup

The practical application of Schema.org vocabulary to a site's pages — Organization, Product, FAQPage, Article — so search and AI systems can extract entities and facts without inference.

Share of Voice (SOV) in AI Answers

The percentage of all brand mentions in a category's AI answers that belong to a given brand. Brandflare counts this only on questions that do not name the brand, so being asked about directly cannot inflate it.

Structured Data

Machine-readable annotations — usually Schema.org vocabulary in JSON-LD — that state a page's facts explicitly (organization, product, price, FAQ), removing ambiguity for crawlers and AI systems.

Sweep (AI Visibility Tracking)

One complete measurement pass in AI visibility tracking: every tracked prompt is run against every monitored engine, and the answers are scored and stored as a dated snapshot for trend analysis.

Training Data

The corpus of text a language model learns from — web pages, reference works, forums, code, licensed archives. A brand's footprint in this corpus determines what the model 'knows' about it by heart.

Vector Search

Retrieval that ranks content by the semantic similarity of embeddings instead of keyword overlap — so a page can be found for questions that share none of its exact words.

Zero-Click Search

A search that ends without a click to any website because the user's need is met on the results surface itself — the default outcome of AI answers, which makes being in the answer the new imperative.