Glossary
The AI visibility glossary.
Plain-English definitions of the AI search vocabulary — and the metrics Buffy Intel reports on your brand.
Buffy= a metric or term specific to Buffy Intel — everything else is industry-standard AI-search vocabulary.
"Best of" List (Listicle)A ranked "best X" article — one of the most-cited page types in AI shopping answers, and usually published by third parties rather than brands.Action PlanBuffyA prioritised, regularly-refreshed set of recommendations to improve your AI visibility — grouped into Owned, Earned, and Optimize.Agentic CommerceShopping where an AI agent discovers, compares, and buys products on a person's behalf — moving the buying decision from human browsing to machine selection.Agentic Commerce Protocol (ACP / UCP)Open standards that let AI agents complete a purchase — exposing product, price, and checkout data in a machine-readable way. Examples: ACP (OpenAI + Stripe) and Google's UCP.AI Crawler (bot)An automated bot that fetches web pages for an AI company — to train a model, build a search index, or pull a page live to answer a question.AI HallucinationWhen an AI generates plausible-sounding but false information — including wrong facts about a brand, invented features, or fake citations.AI OverviewsGoogle's AI-generated answer summaries that appear above traditional search results, synthesising and citing sources for a query.AI Visibility ScoreBuffyA composite 0–100 score blending share of voice, top-ranked %, average rank, and citation coverage into one headline read of how visible your brand is across AI answers.Answer Engine Optimization (AEO)Optimizing your content to become the cited source when AI answer engines generate a direct response to a question.Average RankBuffyYour brand's average position in AI-recommended lists across prompts where it's mentioned. Lower is better — #1 means consistently first.Brand PerceptionBuffyA qualitative read of how AI platforms collectively describe and position your brand, based on branded queries.Citation (AI / LLM)When an AI answer references and links a source it used — being cited (not just mentioned) is how you earn trust and AI-referred traffic.Citation CoverageBuffyOf the AI answers that mention your brand, the share that also cite your brand's own website as a source.Citation Gap (Source Gap)The difference between the sources AI engines cite for your category and the ones that mention you — revealing where you're absent from the answer.Cited PagesBuffyThe unique URLs AI cites as sources when mentioning your brand.Content Decay (Citation Cliff)The decline in how often AI engines cite a page as it ages — live retrieval favors recently-updated content, so citations fall off after roughly a quarter.Conversational SearchSearching by asking full, natural-language questions — often in a back-and-forth — instead of typing keywords, the way people use ChatGPT, AI Mode, and Perplexity.CorroborationMultiple independent, credible sources agreeing on the same fact — a strong signal that makes AI engines more willing to state and cite it.Dark TrafficReal visits whose true source is invisible to analytics — they arrive with no referrer and get filed as 'Direct,' hiding much of your AI-driven traffic.E-E-A-TGoogle's quality framework — Experience, Expertise, Authoritativeness, Trustworthiness — signals that also shape whether AI engines treat your content as a credible source.EmbeddingsNumeric representations of meaning that let an AI judge how related two pieces of text are — the basis of semantic search and retrieval.Entity AuthorityHow strongly and credibly AI associates your brand with a topic — a strong predictor of AI citations, often more than backlinks or technical SEO.Featured SnippetA short, direct answer Google lifts to the top of search results — the pre-AI ancestor of today's AI-generated answers, and a useful model for extractable content.Generative Engine Optimization (GEO)The practice of getting your brand mentioned, cited, and recommended across generative AI engines like ChatGPT, Gemini, and Claude.Google AI ModeGoogle's conversational, AI-generated search experience — a full chat interface powered by Gemini, distinct from the shorter AI Overview summary.GroundingConnecting an AI's answer to verifiable external sources, so claims are backed by retrieved evidence rather than the model's memory.Hero ProductsBuffyYour flagship products — each gets a dedicated deep-dive prompt and is scored individually for how favourably AI describes it.Knowledge CutoffThe date after which a model's training data stops — so without live search, it doesn't 'know' anything more recent.Knowledge GraphA structured map of real-world entities — brands, products, people — and the relationships between them, used by search and AI engines to 'know' what your brand is.Large Language Model (LLM)An AI model trained on vast amounts of text that predicts and generates language — the engine behind ChatGPT, Gemini, Claude, and AI search.llms.txtA proposed standard — a Markdown file at /llms.txt — that gives AI crawlers a concise, curated guide to your site's most important content.Multimodal SearchSearching with images or voice, not just typed text — a fast-growing share of AI queries, where the engine interprets a photo or spoken question.PromptThe question or instruction a person gives an AI — the AI-search equivalent of a search query, but usually longer and more conversational.Prompt InjectionTrying to manipulate an AI's output by feeding it crafted text — including attempts to plant fake rankings or claims a model might repeat.Query Fan-OutWhen an AI decomposes one user question into several parallel sub-searches, then synthesises the results — so a single prompt pulls from many sources.Retrieval-Augmented Generation (RAG)An architecture that pairs a language model with live document retrieval, so answers are grounded in fetched sources rather than memory alone.robots.txtA file at your domain root that tells crawlers — including AI bots — which pages they may or may not fetch.Semantic SearchRetrieval based on meaning rather than exact keywords — matching the intent behind a query to content that answers it, using embeddings.Sentiment AnalysisMeasuring not just whether a brand is mentioned in AI answers, but how it's described — positive, negative, or neutral.Share of Voice (SoV)BuffyThe percentage of AI answers to category prompts that mention your brand at all.Structured Data (Schema markup)Machine-readable Schema.org markup (Product, FAQ, Organization, etc.) that labels what your content is, making it far easier for AI to parse and cite.Top-Ranked %BuffyOf the AI answers that mention your brand, the share that rank it #1 — the single top pick.Total MentionsBuffyThe total number of times your brand was mentioned across all tracked prompts in the latest analysis.Zero-Click SearchWhen a search or AI answer satisfies the query on the results surface itself, so the user never clicks through to a website.
say hi
Let's go
Show upwhere shoppers are looking.
Free tier · no card required · live in five minutes. Buffy will be wagging on the other side of the install.
(or just come say hi)