Every AI search term, explained in plain English.
This glossary defines the 90 terms that B2B SaaS marketers meet when they work on Answer Engine Optimization. Each term has its own page with a direct definition, a worked example, a way to measure it and the sources behind it.
A
Core concepts
Agentic Search
Agentic search is an information-seeking process in which an AI system chooses searches or tools, evaluates intermediate results, and adjusts its next actions. It describes an adaptive workflow rather than a single fixed retrieval request.
Core concepts
AI Agent
In language-model applications, an AI agent is a system in which a language model directs actions and tool use to pursue a goal. Unlike a fixed workflow, an agent can choose its next step based on intermediate results.
Measurement
AI Answer Sentiment
AI answer sentiment is the evaluative tone an AI response expresses toward a specified brand, product, or claim. It is usually classified as positive, neutral, negative, or mixed, with the target and scoring rules defined before analysis.
Measurement
AI Citation
An AI citation is an explicit source reference attached to an AI-generated answer, usually presented as a link or citation marker. It identifies material used or presented as support, but it does not automatically prove that the associated claim is correct.
Technical
AI Crawler
An AI crawler is an automated client that collects web content for an AI-related purpose, such as model development or search indexing. Its purpose depends on the specific provider and user agent, not the general label “AI bot.”
Core concepts
AI Hallucination
An AI hallucination is generated content that is false, fabricated, or unsupported by the relevant evidence while presented as an answer. It can involve invented facts, nonexistent sources, incorrect product capabilities, or claims that exceed what retrieved material establishes.
Measurement
AI Mode Citation
An AI Mode citation is a supporting source link displayed with a generated response in Google AI Mode. It identifies a linked page, but it does not by itself prove that every surrounding claim is fully supported.
AI search engines
AI Overviews
AI Overviews are Google Search features that provide AI-generated summaries for some searches, often with links to supporting sources. They appear within the search results experience and do not replace every conventional result or appear for every query.
Measurement
AI Referral Traffic
AI referral traffic is website traffic attributed to links clicked from AI assistants or AI search experiences. It measures observed visits rather than answer visibility, and attribution can be incomplete when referrers, campaign parameters, or reporting categories are unavailable.
Core concepts
AI Search Optimization
AI search optimization is the practice of improving a website and its supporting evidence for discovery and use by AI-powered search experiences. It combines search fundamentals with content clarity, entity consistency, and answer-level measurement.
Core concepts
AI Visibility
AI visibility is the extent to which a brand, product, or source appears in AI-generated responses for a defined set of questions. It can include mentions, linked citations, recommendation context, and the accuracy of the information presented.
Measurement
AI Visibility Score
An AI visibility score is a composite metric that summarizes selected signals of brand presence in AI answers. Its interpretation depends on the provider’s inputs, weights, normalization, prompt sample, and engine coverage; no universal official score exists.
Core concepts
Answer Engine
An answer engine is a system that responds directly to a question instead of only listing documents to visit. It may use retrieved sources, stored knowledge, structured data, or a combination of these inputs.
Core concepts
Answer Engine Optimization
Answer engine optimization is the practice of making a brand and its information easier for answer engines to discover, understand, and use in responses. It combines accessible content, clear evidence, and consistent entity information.
Content
Answer-First Writing
Answer-first writing states a direct answer before supporting explanation, evidence, and qualifications. It helps readers identify the main point quickly while preserving the context needed to interpret that point accurately.
Technical
Applebot-Extended
Applebot-Extended is a robots.txt control that lets publishers opt out of having Applebot-crawled content used to train Apple’s generative foundation models. It is distinct from allowing Applebot to crawl for discovery.
Technical
Article Schema
Article schema is structured data that describes an editorial work using Schema.org’s Article type or a more specific subtype. It can identify the article’s subject, author, and publishing details.
B
Technical
Bingbot
Bingbot is Microsoft Bing’s primary web crawler for discovering and fetching web content. Its crawling role is different from a conversational assistant that generates a response for a user.
Technical
BM25
BM25 is a lexical relevance scoring method that uses term frequency, inverse document frequency, and document-length normalization. It ranks documents by statistical relationships between query terms and document text rather than by generated answers.
Authority
Brand Mentions
Brand mentions are references to a company or product by name in content or conversations. In AI visibility analysis, a mention means the answer names the brand, whether or not it includes a clickable citation or recommends the product.
Technical
Breadcrumb Schema
Breadcrumb schema describes a page’s place within a linked site hierarchy using Schema.org’s BreadcrumbList type. It represents navigation relationships rather than the full set of links on a page.
C
Technical
Canonical URL
A canonical URL is the representative URL selected for a group of duplicate or substantially similar pages. Publishers can signal a preferred version, but a search engine can choose a different canonical.
Technical
CCBot
CCBot is Common Crawl’s web crawler, used to collect data for its open web crawl repository. Allowing CCBot access is not equivalent to submitting a page to a specific AI answer engine.
AI search engines
ChatGPT Search
ChatGPT Search is ChatGPT’s web-search capability for finding online information and producing answers with source links. It differs from a response based only on model knowledge because retrieved web material can inform the answer.
Measurement
Citation Rate
Citation rate is the percentage of eligible AI answers that cite a defined website, page, or source set. It measures source inclusion within a specified sample, so meaningful comparisons require the same denominator, citation rules, and collection method.
Technical
ClaudeBot
ClaudeBot is Anthropic’s web crawler for collecting content that could contribute to training its generative AI models. Anthropic distinguishes it from Claude-SearchBot for search and Claude-User for user-directed retrieval.
Authority
Co-Citation
Co-citation occurs when two sources or entities are cited together by another source. In SEO and AEO discussions, the term is sometimes extended to shared brand references, so analysts should distinguish actual shared citations from simple textual co-occurrence.
Content
Comparison Page
A comparison page evaluates products, approaches, or vendors against explicit criteria. A useful comparison explains fit, differences, and limitations rather than declaring an unsupported universal winner.
Content
Content Chunking
Content chunking divides information into coherent sections or passages that can be understood and retrieved separately. Editorial chunking organizes published content, while retrieval pipelines may split documents again according to their own processing rules.
Content
Content Freshness
Content freshness is the extent to which information reflects the current state of its subject. It concerns substantive accuracy and timely updates, not simply a recent publication date displayed on unchanged content.
Technical
Context Engineering
Context engineering is the practice of selecting, organizing, and maintaining the information available to a language model during generation. It includes instructions, retrieved evidence, conversation history, and tool outputs.
Technical
Context Window
A context window is the amount of information a model can consider within a generation request, measured in tokens. Depending on the model, that budget includes instructions, conversation history, retrieved material, and generated output.
Content
Conversational Query
A conversational query expresses an information need in natural language, often with context, constraints, or follow-up references. It can ask for an explanation or comparison rather than only naming a topic.
Technical
Crawl Budget
For Google, crawl budget is the set of URLs Googlebot can and wants to crawl, determined by crawl capacity and crawl demand. It concerns fetching resources, not a guaranteed allowance of indexed pages.
D
E
Authority
E-E-A-T
E-E-A-T means experience, expertise, authoritativeness, and trustworthiness. Google uses these concepts in its guidance for evaluating content quality, with trust at the center; E-E-A-T is not a single ranking factor or a public numerical score.
Content
Entity
An entity is a distinct thing that can be identified and described, such as a company, product, person, or concept. Search and knowledge systems use entity representations to distinguish identities and connect facts about them.
Authority
Entity Home
An entity home is the primary page a publisher maintains to identify and describe a specific organization, person, or product. It acts as a consistent reference point for identity, but it is a practical publishing concept rather than an official universal search feature.
Authority
Entity SEO
Entity SEO is the practice of making identifiable people, organizations, products, and their relationships clear to search systems. It combines accurate content, consistent identity information, and relevant structured data rather than relying only on keyword matching.
F
Technical
FAQ Schema
FAQ schema is Schema.org markup for a page containing questions and their publisher-provided answers. Google stopped showing FAQ rich results on 7 May 2026, so FAQ markup should not be presented as a current Google rich-result tactic.
Content
Featured Snippet
A featured snippet is a Google search result that displays an extracted answer before the usual page link. It highlights information from a web page rather than generating a new synthesis across sources.
G
Core concepts
Generative Engine Optimization
Generative engine optimization is the practice of improving content and evidence so generative search systems can discover, interpret, and reference them. Its focus is visibility within generated responses rather than placement in a conventional list of links.
AI search engines
Generative Search
Generative search is a search approach that uses a generative model to produce a synthesized response to a query. It often combines retrieved information with generated text and may include citations, follow-up questions, or conventional search results.
AI search engines
Google AI Mode
Google AI Mode is a Google Search experience that uses AI to answer questions and support conversational follow-ups. It can explore related searches, synthesize information, and provide links that let users investigate sources and continue their research.
Technical
Google-Extended
Google-Extended is a robots.txt product token for controlling certain uses of crawled content in Gemini model training and grounding in specified Google AI products. It is not a separate crawler and does not control Google Search inclusion or ranking.
Technical
GPTBot
GPTBot is OpenAI’s crawler for collecting web content that may be used to improve its generative AI foundation models. Its robots.txt control is separate from OAI-SearchBot, which supports search visibility in ChatGPT.
Core concepts
Grounding
Grounding is the process of connecting an AI response to identifiable evidence or an external reference. In search applications, it usually means basing claims on retrieved sources rather than relying only on learned model associations.
H
Technical
HowTo Schema
HowTo schema is Schema.org markup for instructions that achieve a result through a sequence of steps. The vocabulary still exists, but Google Search has deprecated its How-to rich results.
Technical
Hybrid Search
Hybrid search combines lexical matching with semantic retrieval to produce a unified result list. It can preserve exact-term matches while also retrieving documents that express related meaning with different wording.
I
Technical
IndexNow
IndexNow is a protocol for notifying participating search engines when website URLs are added, updated, or deleted. It helps communicate changes without waiting solely for routine discovery, but it does not guarantee crawling, indexing, ranking, or AI citations.
Content
Information Gain
Information gain, in content planning, is the useful new understanding a page adds beyond existing material. This editorial usage is distinct from the formal mathematical concept used in statistics and machine learning.
J
K
Core concepts
Knowledge Cutoff
A knowledge cutoff is a stated time boundary associated with the information included in a model’s training. It indicates a freshness limitation of learned knowledge, not a guarantee that the model knows every earlier fact or cannot access newer information through tools.
Authority
Knowledge Graph
A knowledge graph is a structured representation of entities and the relationships between them. It connects identifiable things, such as companies, products, and people, so systems can distinguish identities and retrieve related facts.
L
Core concepts
Large Language Model
A large language model is a machine learning model trained on extensive language data to learn patterns and generate or interpret text. Its output reflects learned statistical relationships and supplied context rather than an inherently verified database of facts.
Technical
Lexical Search
Lexical search retrieves information using the words or terms present in a query and indexed content. It emphasizes term matching, while semantic search focuses on relationships in meaning.
Core concepts
LLM Optimization
LLM optimization is the practice of improving how large language model applications discover, interpret, and represent a brand or topic. In marketing, it usually concerns accessible evidence and answer visibility, not changing the model’s internal parameters.
Technical
llms.txt
llms.txt is a proposed Markdown file that helps language-model applications find useful website resources. It usually provides a concise site overview and curated links, but it is not an access-control standard or a guaranteed AI visibility signal.
M
Measurement
Mention Rate
Mention rate is the percentage of eligible AI answers that name a defined brand or product. It measures answer-level presence rather than citation links, recommendation strength, or the number of times the name appears inside each answer.
AI search engines
Microsoft Copilot
Microsoft Copilot is Microsoft’s family of AI assistant experiences. In search-related contexts, Copilot can use web information to generate responses with supporting links, including through Copilot Search in Bing.
Technical
Model Context Protocol (MCP)
Model Context Protocol is an open standard for connecting AI applications with external tools and data sources. It provides a shared integration interface rather than a ranking signal or a public search index.
O
Technical
OAI-SearchBot
OAI-SearchBot is OpenAI’s search crawler, used to surface websites in ChatGPT search features. Its access control is independent of GPTBot, so publishers can allow search crawling while disallowing the crawler associated with foundation-model training.
Technical
Organization Schema
Organization schema is Schema.org structured data that describes a business or other organization. It can express the organization’s name, official website, logo, identifiers, and relationships to other entities.
Authority
Original Research
Original research produces new evidence through a documented investigation rather than merely summarizing existing publications. Its usefulness depends on the method, transparency, and limits of the findings.
P
Technical
Passage Ranking
Passage ranking is a Google system that identifies individual sections of a web page to better understand the page’s relevance to a search. It does not turn each passage into a separately published page.
Content
People Also Ask
People Also Ask is a Google Search feature that groups related questions and expandable answers. It helps searchers explore connected information needs without starting every question as a separate search.
AI search engines
Perplexity
Perplexity is an AI-powered answer and search service that responds to questions using information sources and typically displays citations. It supports research-style interactions in which users can inspect references and ask follow-up questions.
Technical
PerplexityBot
PerplexityBot is Perplexity’s crawler for discovering content that can surface and be linked in its search results. Perplexity states that this crawler is not used to collect content for training AI foundation models.
Technical
Product Schema
Product schema is structured data that describes an offered product or service using Schema.org’s Product type. It can express product identity and related details without guaranteeing a rich result or AI recommendation.
Measurement
Prompt Tracking
Prompt tracking is the repeated collection and analysis of AI answers to a defined set of questions. It monitors observable changes in brand mentions, citations, positioning, and accuracy while preserving the conditions needed to interpret those changes.
Q
R
Technical
Reciprocal Rank Fusion (RRF)
Reciprocal rank fusion is a method for merging ranked result lists using document positions rather than directly comparing their raw relevance scores. It is commonly used to combine results from complementary retrieval methods.
Technical
Reranking
Reranking is a later retrieval stage that reorders an initial set of search results using another relevance assessment. It can improve which candidate documents reach an answer generator without changing the original document collection.
Core concepts
Retrieval-Augmented Generation
Retrieval-augmented generation is an approach that retrieves relevant information and supplies it to a generative model as context for an answer. It lets an application use external documents without relying solely on facts encoded in model parameters.
Authority
Review Sites
Review sites are platforms that collect and publish evaluations of products or services. For B2B SaaS, they may combine customer feedback, vendor profiles, comparison pages, and commercial placements, so readers should inspect verification practices and disclosure rules.
Technical
robots.txt
robots.txt is a public text file that gives participating crawlers instructions about which website paths they may request. It controls crawling behavior, not user authentication, and it does not reliably keep a URL out of search indexes.
S
Authority
sameAs
sameAs is a Schema.org property that points to a reference page unambiguously identifying the same entity. It helps describe identity across pages or platforms and should not be used for partners, competitors, categories, or merely related organizations.
Technical
Schema Markup
Schema markup is code that describes webpage entities and relationships using the Schema.org vocabulary. It helps compatible systems interpret facts such as an organization’s identity, a product’s attributes, or an article’s author.
Content
Semantic Search
Semantic search retrieves information by interpreting meaning and relationships rather than relying only on exact keyword matches. Implementations may use embeddings, entity understanding, language models, and lexical signals to match a query with relevant content.
Technical
Server-Side Rendering
Server-side rendering generates a webpage’s HTML on the server before sending it to the client. It can expose meaningful content in the initial response instead of requiring client-side JavaScript to construct that content.
Measurement
Share of Model
Share of model is a nonstandard metric for a brand’s relative presence in outputs from a specified AI model or assistant. Its meaning depends on the provider’s counting rules, prompt sample, comparison set, and model configuration.
Measurement
Share of Voice
Share of voice is a brand’s proportion of a defined set of measured category visibility. For AI answers, it commonly compares qualifying brand appearances across selected competitors, prompts, and engines, using a methodology that must be stated explicitly.
Technical
Structured Data
Structured data is information organized into an explicit format with defined fields and relationships. On websites, it can describe entities and page content in machine-readable forms such as JSON-LD using Schema.org.
T
Authority
Third-Party Validation
Third-party validation is evidence about a company, product, or claim provided by a separate organization or person. Its credibility depends on the source’s expertise, independence, methods, and relevance, not simply on appearing outside the company’s website.
Technical
Token
A token is a unit of content that a language model processes. For text, tokens can represent words, word fragments, or other character sequences, and the mapping depends on the model’s tokenizer.
Authority
Topical Authority
Topical authority is the perceived depth and reliability of a publisher’s coverage of a subject. In SEO and AEO, it is a working concept for assessing expertise and usefulness, not a universal score published by Google or AI engines.
Core concepts
Training Data
Training data is the collection of examples used to adjust a machine learning model’s parameters during training. For language models, it can include text and other material used to learn language patterns, knowledge associations, and task behavior.
V
X
Z
Ask AI about your category. Then ask us why you're missing.
We will run your buyers' real questions through every major AI engine and show you exactly who gets named instead of you.
Get my free AI search grade →