GEO & AI Visibility Glossary

The current period in which AI-generated content, AI-mediated search, and LLM-assisted decision-making are becoming an increasingly important layer of buyer discovery and vendor evaluation, alongside conventional search, social platforms, and publisher research.

A contextual measure of how frequently qualifying authoritative sources substantiate a brand, entity, or claim within a defined set of AI-generated answers. Reported separately from mention frequency, sentiment, and recommendation strength. Distinct from Citation Authority below, which is Brainpan.AI’s narrower, disclosed AIVI™ companion metric operationalizing this broader concept.

The share of a category’s AI answers in which a given brand is named, relative to all brands named for the same queries — the generative-era equivalent of traditional share of voice in media planning. Brainpan.AI operationalizes this as Share of Model, the largest component of the AIVI™ Brand Score.

Brainpan.AI’s term for the strength, consistency, and corroboration of an entity’s identity and attributes across first-party content, structured data, search systems, and authoritative external sources. Strong entity signals may improve accurate recognition and retrieval, but do not guarantee inclusion or citation.

A systematic diagnostic process that queries major LLMs (ChatGPT, Gemini, Perplexity, Copilot, and Claude) and answer engines with category-relevant prompts to determine whether and how a brand is mentioned, cited, and recommended. As performed by Brainpan.AI, the audit benchmarks Mention Rate, Citation Rate, Answer Accuracy, and competitive Share of Model.

An audit methodology that identifies the specific query types, topic areas, and content formats for which competitors are mentioned or cited by AI systems while the target brand is not. Citation gap analysis is the primary input to a content remediation roadmap.

The practice of reviewing which sources an AI system cites in its answers, and why, to identify which content types, publishers, and pages earn citation and which do not. Distinct from Citation Gap Analysis, which compares citation outcomes against named competitors rather than analyzing citation patterns in general.

How often an AI names a brand inside the generated answer itself, counted only when the model surfaces the brand on its own — paid or sponsored placements are excluded. It measures whether the AI retrieves the brand as part of its response. Example: in “For affordable car insurance, compare GEICO, Progressive, State Farm, and Allstate,” each named carrier earns one organic mention. Organic mentions answer the question “Did the AI talk about the brand?” and are the raw signal behind Mention Rate and Share of Model.

The percentage of eligible AI-generated responses in which a tracked brand or entity is mentioned at least once. Multiple mentions of the same entity within one response count as one response-level occurrence unless frequency-based counting is explicitly disclosed. Answers “do you show up at all?” — distinct from Share of Model, which measures your share of total mentions against competitors.

How often an AI answer backs a brand with a cited source — a link or reference that supports, associates, or grounds the brand in evidence, rather than merely naming it. Citations are classified by origin: owned (the brand’s own pages, 1st-party), affiliated quote/comparison marketplaces (2nd-party), and independent editorial sources such as NerdWallet (3rd-party). All cited mentions feed Citation Influence, the index’s citation component; Citation Authority (third-party) and Citation Independence (non-affiliated) are reported as companion metrics. Citation count answers the question “Did the AI back the brand with source evidence?”

How much the AI’s cited sources put a brand in front of the buyer — the citation component of the AIVI™ Brand Score (15%). It counts every cited mention regardless of source, because a sourced mention still reaches and reassures the reader whether the source is the brand’s own page or an independent one. Citation Influence is the buyer-impact lens — the AI-answer analogue of search Share of Voice, which counts all ranking presence. Authority and Independence refine it by source origin.

A brand's share of all eligible organic brand mentions captured within a defined query set, competitive universe, AI platform set, market, language, and measurement period. Paid placements are excluded and reported separately. It is the largest component of the AIVI™ Brand Score (35%), built from Organic Mentions, and the scored, operational form of AI share of voice (AI-SoV). Distinct from Mention Rate (do you show up at all) and Citation Rate (does a response back you with a source).

The percentage of eligible AI-generated responses in which a tracked brand is explicitly recommended, endorsed, shortlisted, or presented as a suitable option. A brand mention does not automatically constitute a recommendation — this is a stricter bar than Mention Rate, and the basis for Recommendation Strength, the AIVI™ Brand Score’s second-largest component (25%).

Isolates the extent to which independent sources substantiate a brand in observed AI answers. It measures earned source support within the study — not the general authority of the brand or domain. A disclosed companion metric reported alongside the AIVI™ Brand Score, not weighted into it; only third-party citations count, so a brand’s own pages can’t inflate it. Rooted in E-E-A-T: authority is conferred by others, not claimed. Distinct from LLM Citation Authority above, the broader concept this metric operationalizes.

The strictest citation lens: only genuinely independent third-party sources count — a brand’s own pages (1st-party) and affiliated quote/comparison marketplaces (2nd-party) are both removed. A disclosed companion metric to the AIVI™ Brand Score, not part of its weighting. It mirrors how search authority metrics discount sponsored and affiliate links (rel="sponsored"), isolating pure editorial endorsement.

The degree to which factual statements in an AI-generated answer agree with validated, current information about the entity, product, service, price, availability, or policy being discussed. Assessed separately from Narrative Alignment.

The degree to which an AI-generated answer describes a brand's category, audience, capabilities, and positioning consistently with the brand's current market identity. Narrative alignment does not require favorable coverage and should be assessed separately from sentiment.

Brainpan’s AI Visibility Index — a single 0–100 score for how visible a brand is across the major AI engines (ChatGPT, Gemini, Claude, Perplexity, and Copilot). As published in the Insurance Q2 2026 benchmark, it combines five weighted components: Share of Model (35%), Recommendation Strength (25%), Position weighting (20%), Citation Influence (15%), and Top-3 rate (5%), normalized so the category leader scores 100. Built on organic placements only; paid is quarantined and reported separately. See QVI™ below for the volume-neutral companion metric used in the Medicare Q3 2026 benchmark.

Brainpan.AI’s volume-neutral quality score, introduced in the Medicare AIVI™ Q3 2026 benchmark: 8 × mean Position Score + 7 × mean Recommendation Score + 25 × own-domain citation rate, scaled 0–100. Unlike the Insurance AIVI™ Brand Score, QVI™ is deliberately volume-neutral, so a brand’s rate of being answered well is measured separately from how often it is answered at all — which is how Medicare.org leads QVI™ at 48.05 while holding just 1.64% of commercial mention share. See the AIVI™ methodology for how it relates to the Brand Score.

A brand’s share of commercial-brand mentions: the mentions where competing brands appear against each other. In the Medicare AIVI™ Q3 2026 benchmark, the denominator is 59,055 commercial-brand mention rows. Used for competitive benchmarking (how much of the carrier conversation is yours), and the metric behind the published leaderboard (e.g. UnitedHealthcare 18.35%). See also Total SOV.

A brand’s share of every extracted mention in a study, including generic category terms (e.g. ‘Medicare Advantage’, ‘Part D’) and non-commercial organizations such as government, advocacy and media sources. In the Medicare Q3 2026 benchmark, the denominator is 229,372 rows. Used for category-opportunity sizing. Because the two denominators differ by roughly 4×, a brand’s rank can shift by more than 40 places between the two measures. For example, Kaiser Permanente has 2.47% Commercial SOV (9th among carriers) but 0.63% Total SOV. The two are reported separately, never blended. See also Commercial SOV.

A Google SERP placement in which a content excerpt is extracted and displayed above organic results in a dedicated box. Google now counts the featured snippet as a regular ranking position rather than a special “Position Zero” slot outside the normal count. Featured snippets and AI-generated search experiences may draw on overlapping content-quality, relevance, and retrieval signals, but performance in one does not guarantee visibility in the other.

A structured information card Google displays for recognized entities — brands, people, and organizations — alongside relevant search results. Placement is device-neutral: it can appear beside, above, or below organic results depending on device and query, not in one fixed spot. Knowledge Panel presence indicates that Google recognizes an entity sufficiently to display a structured entity result. It should be treated as an entity-recognition signal, not proof of visibility or citation across independent AI platforms.

Google’s AI-generated summary that answers a search query directly at the top of the results page, synthesizing and citing multiple sources above traditional organic listings. AI Overviews and AI Mode are Google’s two generative search surfaces today — AI Overviews injects a synthesized answer into standard results, while AI Mode is a separate, dedicated conversational tab. AI Overviews remains a primary AEO target and a major driver of zero-click search.

Google’s structured system for understanding real-world entities and their relationships. It supports entity-oriented Search experiences such as Knowledge Panels. Accurate entity representation can reduce ambiguity, but Knowledge Graph inclusion does not guarantee selection or citation in AI-generated answers.

Google Search’s conversational AI mode — a dedicated, chat-style tab distinct from traditional results and from AI Overviews. Clicks from AI Mode report in GA4 and Adobe Analytics as Organic Search, not as a distinct AI channel. See FAQ Q31.

Google’s own description of how AI Mode breaks a single question into multiple parallel sub-queries, retrieving results for each and synthesizing them into one answer — why a single AI Mode prompt can surface a far broader set of pages than the equivalent traditional search.

AI agents navigating and interacting with websites on a user’s behalf — filling forms, comparing options, completing tasks — rather than a human clicking through page by page. Google’s PageSpeed Insights began running dedicated “Agentic Browsing” checks in 2026, a sign that agent-driven traffic is now a real category sites need to account for, not a theoretical one.

OpenAI’s two separate crawlers, controlled independently in robots.txt: OAI-SearchBot handles search and retrieval inclusion for ChatGPT’s live answers, while GPTBot handles model-training data collection. Blocking one does not block the other. See FAQ Q13 and OpenAI, Crawlers.