Back to Home
AI Buyer Perception Glossary
40 key terms and concepts in AI buyer perception monitoring, Generative Engine Optimization (GEO), and Answer Engine Optimization (AEO) — explained for B2B SaaS teams. Each term links to a full definition with practical context.
- AI Brand Monitoring
- AI brand monitoring is the continuous tracking of what AI systems say about a brand across engines, prompts, and time — covering mentions, sentiment, claim accuracy, recommendations, and cited sources. It extends social listening into a channel analytics cannot see, where answers are generated privately for each user rather than published once. Full definition
- AI Buyer Perception
- AI buyer perception is how AI assistants such as ChatGPT, Perplexity, Gemini, and Claude describe, position, and recommend a product when prospective buyers ask evaluation questions — comparisons, pricing, alternatives, and fit. It functions as a parallel reputation layer: buyers increasingly form vendor shortlists from AI answers before ever visiting a website. Full definition
- AI Citations
- AI citations are the source links an AI engine attaches to its answer, showing which pages grounded each statement. Engines with live retrieval — Perplexity, ChatGPT search, Gemini, and Google's AI Overviews — cite routinely. Citations are the clearest lever in GEO: whichever pages get cited largely control what the answer says. Full definition
- AI Crawler
- An AI crawler is a bot that fetches web content for AI systems — either to build training corpora (GPTBot, ClaudeBot) or to retrieve live pages for grounded answers (PerplexityBot, OAI-SearchBot). Whether you allow or block each crawler in robots.txt determines whether your content can shape AI training data and real-time answers. Full definition
- AI Overviews
- AI Overviews are Google's AI-generated summaries that appear above traditional results for many queries, synthesizing an answer from multiple sources with citation links. For brands they are a high-stakes surface: an Overview can answer a buyer's question — accurately or not — before any organic listing gets the chance to earn a click. Full definition
- AI Recommendation Rate
- AI recommendation rate is the share of buyer-intent prompts for which an AI engine explicitly recommends your product — names it first, calls it the best fit, or picks it over alternatives — rather than merely mentioning it. It is stricter and more revenue-correlated than mention rate, because recommendations steer shortlists directly. Full definition
- AI Search Monitoring
- AI search monitoring is the systematic tracking of how AI-powered search surfaces — ChatGPT search, Perplexity, Gemini, Copilot, and Google's AI Overviews — answer queries relevant to a brand. It typically covers mentions, recommendations, sentiment, claim accuracy, and citations, sampled on a schedule so changes are caught rather than discovered by a prospect. Full definition
- Answer Diff
- An answer diff is a side-by-side comparison of an AI engine's responses to the same prompt at two points in time, highlighting claims that were added, removed, or changed. It is the changelog of your AI presence: diffs reveal displacement, new hallucinations, or recovered recommendations that a single snapshot would miss. Full definition
- Answer Engine
- An answer engine is a system that responds to a query with a synthesized answer rather than a list of links — ChatGPT, Perplexity, Copilot, Gemini, and Google's AI Overviews all qualify. The shift from search engine to answer engine collapses the buyer journey: vendor selection and framing happen inside the answer, not on your website. Full definition
- Answer Engine Optimization (AEO)
- Answer Engine Optimization (AEO) is the discipline of formatting content so answer engines can extract it as a direct response: question-shaped headings, concise standalone definitions, FAQ and how-to schema, and clear entity references. AEO overlaps with GEO but focuses specifically on being quoted as the answer — in a featured snippet, voice reply, or AI-generated response. Full definition
- Answer Synthesis
- Answer synthesis is the process by which an AI engine merges information from multiple retrieved sources and its training data into a single fluent response. Synthesis is where brand claims get blended, compressed, and sometimes distorted: an answer may combine your pricing page with an outdated review and present both with equal confidence. Full definition
- Brand Hallucination
- A brand hallucination is a confident but false statement an AI model makes about a company or product — invented pricing, features that don't exist, wrong integrations, or fabricated policies. Unlike a bad review, it carries the authority of a neutral assistant and repeats to every buyer who asks a similar question. Full definition
- Brand Mention Rate
- Brand mention rate is the percentage of AI answers, across a defined prompt set, in which your brand appears at all. It is the top of the AI visibility funnel: before you can be recommended or accurately described, you must be mentioned. A low rate usually signals weak entity presence and citations rather than negative sentiment. Full definition
- Buyer-Intent Prompt
- A buyer-intent prompt is a question a prospective customer asks an AI assistant while actively evaluating a purchase — 'best X for Y', 'alternatives to Z', 'does X integrate with Y', 'is X worth the price'. These prompts matter more than informational queries because the answer often decides which vendors make the buyer's shortlist. Full definition
- Citation Gap
- A citation gap is the set of sources an AI engine relies on for answers in your category that you neither own nor appear in. Closing the gap — earning coverage or accurate mentions on those third-party domains — is often the fastest way to change what AI says, because those pages already feed the answers. Full definition
- Citation Monitoring
- Citation monitoring is the ongoing tracking of which URLs and domains an AI engine cites when answering questions about your brand or category. It shows whether AI grounds its answers in your documentation and pricing pages or in third-party sources — review sites, Reddit threads, competitor comparisons — that you don't control. Full definition
- Claim Extraction
- Claim extraction is the automated parsing of AI-generated answers into discrete, checkable statements about a brand — pricing figures, feature assertions, integration support, compliance status, and competitive comparisons. Each claim can then be classified as accurate, outdated, false, or missing, turning a wall of AI prose into an auditable list of facts to verify. Full definition
- Competitor Displacement
- Competitor displacement is when an AI engine that previously named or recommended your product on a buyer-intent prompt shifts to recommending a competitor instead. Because AI answers change with model updates and fresh sources, displacement can happen silently between two scans — and it directly reroutes buyers who trust the answer they get. Full definition
- Conversational Search
- Conversational search is information seeking through multi-turn dialogue with an AI assistant instead of one-shot keyword queries. Context carries across turns, so a buyer can narrow from 'best CRM' to team size, budget, and integrations within one session — meaning brands are evaluated against increasingly specific criteria that keyword tools never see. Full definition
- Dark Funnel
- The dark funnel is the portion of the buyer journey invisible to a vendor's analytics — peer recommendations, private communities, podcasts, and now AI chat sessions. AI assistants are the fastest-growing dark-funnel channel: prospects arrive at 'request demo' already shaped by answers you never saw, from conversations no tracking pixel can observe. Full definition
- Entity SEO
- Entity SEO is optimization focused on making a brand an unambiguous, well-connected entity in the knowledge systems machines use — consistent naming, organization schema, Wikipedia/Wikidata presence, and corroborating third-party references. It matters for AI answers because models reason over entities and relationships, not keywords: an ambiguous entity gets conflated, ignored, or hallucinated. Full definition
- Generative Engine Optimization (GEO)
- Generative Engine Optimization (GEO) is the practice of structuring and publishing content so generative AI engines — ChatGPT, Perplexity, Gemini, Claude, and Google's AI Overviews — cite your pages and represent your brand accurately in their answers. Where SEO targets rankings on a results page, GEO targets inclusion and framing inside the answer itself. Full definition
- GPTBot
- GPTBot is OpenAI's web crawler that collects publicly available content for training its models. Site owners control it via robots.txt: blocking GPTBot keeps pages out of future training data, but does not remove existing knowledge and does not affect ChatGPT's separate live-search fetching, which uses other user agents such as OAI-SearchBot. Full definition
- Grounding (LLM)
- Grounding is the technique of anchoring a language model's answer in retrieved, verifiable sources — search results, documents, or a knowledge base — rather than in parametric memory alone. Grounded answers cite their sources and track current facts; ungrounded answers rely on training data, so brand information degrades toward the model's knowledge cutoff. Full definition
- Knowledge Cutoff
- A knowledge cutoff is the date after which a language model has no training data; events, launches, and pricing changes past that date are unknown to the base model. Cutoffs explain many outdated brand claims — and why engines with live retrieval can be accurate while the same underlying model without browsing stays stale. Full definition
- Knowledge Graph
- A knowledge graph is a structured database of entities — companies, products, people — and the relationships between them, used by search engines and AI systems to ground facts. Presence in graphs like Google's Knowledge Graph and Wikidata gives models a canonical record of what your company is, reducing conflation and hallucination. Full definition
- LLM SEO
- LLM SEO is an umbrella term for tactics that improve a brand's presence in large language model outputs — entity clarity, extractable definitions, citation-worthy pages, crawler access, and third-party corroboration. Used near-interchangeably with GEO and AEO, it emphasizes continuity: existing search skills extended to answer engines, rather than a discipline built from scratch. Full definition
- LLM Visibility
- LLM visibility is the degree to which a brand appears — and appears accurately — in the outputs of large language models when users ask relevant questions. It spans mention frequency, recommendation position, sentiment, and citation of the brand's own pages. Unlike search visibility, it cannot be read from a rankings page; it must be measured by querying the models directly. Full definition
- llms.txt
- llms.txt is a proposed standard file placed at a site's root that gives large language models a curated, markdown-formatted guide to the site's most important content. Modeled loosely on robots.txt but permissive rather than restrictive, it helps AI systems find canonical pricing, documentation, and product facts instead of guessing from scattered pages. Full definition
- Model Refresh
- A model refresh is the release of an updated AI model version — new training data, new tuning, or a new architecture — that can materially change what the model says about a brand overnight. Refreshes are a major cause of sudden answer shifts, which is why AI monitoring compares answers across time and model versions. Full definition
- Perception Score
- A perception score is a composite metric summarizing how favorably and accurately AI engines represent a brand, typically combining mention rate, recommendation position, claim accuracy, sentiment, and citation ownership across a fixed prompt set. Its value lies in trend and comparison — tracking movement scan over scan — rather than in the absolute number. Full definition
- Prompt Pack
- A prompt pack is a curated, reusable set of buyer-intent prompts organized around one evaluation theme — pricing, integrations, security, head-to-head comparisons, or use-case fit. Running the same pack on a schedule makes AI answers measurable over time: identical inputs isolate what changed in the model's output rather than in your questions. Full definition
- Prompt Simulation
- Prompt simulation is the practice of systematically running the questions your buyers would ask — across multiple AI engines, on a repeatable schedule — and capturing the responses for analysis. It is the only reliable way to observe AI answers at scale, since outputs vary by engine, phrasing, and time and cannot be looked up anywhere. Full definition
- Retrieval-Augmented Generation (RAG)
- Retrieval-Augmented Generation (RAG) is an architecture in which a system first retrieves documents relevant to a query, then feeds them to a language model to generate the answer. RAG powers most AI search engines, which is why the pages that get retrieved — yours or a third party's — largely determine what AI says about you. Full definition
- Structured Data for AI
- Structured data for AI is the use of machine-readable markup — schema.org JSON-LD for organizations, products, FAQs, and offers — to state facts about your business unambiguously. While LLMs primarily read prose, the retrieval and parsing layers of answer engines use structured data to disambiguate entities and extract facts, making markup a low-effort GEO foundation. Full definition
- Verbatim Answer Capture
- Verbatim answer capture is the practice of storing the full, word-for-word text an AI engine produced for a prompt, rather than only derived scores or mention counts. The raw answer is the evidence layer: it lets teams see exactly what buyers were told, quote it internally, and prove that a change actually occurred. Full definition
- Zero-Click Research
- Zero-click research is buyer research completed entirely inside a search or AI interface, with no visit to any vendor's website. AI answers accelerate the pattern: a buyer can compare vendors, check pricing claims, and form a shortlist from synthesized answers alone — which makes what those answers say your de facto landing page. Full definition
Learn More
Monitor your AI buyer perception
See how AI engines describe your brand to potential buyers — start monitoring free, no credit card required.
Start Free Trial