AI Search Glossary

Agentic commerce

Agentic commerce is shopping in which an AI agent acts for a person: it finds and compares products and can complete the purchase, sometimes without the shopper opening the merchant's website. Open protocols such as the Agentic Commerce Protocol from OpenAI and Stripe and Google's Universal Commerce Protocol give agents and merchants a shared way to handle the order.

Agentic search

Agentic search is when an AI system runs its own research loop instead of a single search. It plans queries, reads the results, decides what's still missing, and searches again until it can answer. Deep research modes in ChatGPT, Gemini, Perplexity, and Claude work this way, and they read far more pages than one search returns.

Also called: agentic RAG, deep research

AI brand mentions

AI brand mentions are the times an AI engine like ChatGPT or Google AI Overviews names your brand in an answer. A mention counts whether or not the answer links to your website. That's the difference from a citation, where the engine credits a specific page as its source.

Also called: LLM brand mentions, mention rate

AI brand monitoring

Keeping watch on what AI engines say about your brand and which sources they draw on when they say it.

Also called: AI search monitoring

AI brand sentiment

The tone an AI engine takes when it describes your brand: positive, neutral, or negative, and the specific claims behind that tone.

AI citations

AI citations are the links an AI engine attaches to its answer to show which web pages the information came from. Each citation credits a specific URL. A brand mention is different: the engine names a company in the text, even when it doesn't link to that company's site.

Also called: citation frequency, citation rate, LLM citations

AI crawlers

AI crawlers are bots run by AI companies that fetch web pages. Some collect training data for future models, some build the search index an AI engine retrieves from, and some fetch a page on demand when a user asks a question. Each type has its own user agent, so robots.txt can allow one and block another.

Also called: AI bots, LLM crawlers, ClaudeBot, PerplexityBot

AI hallucination

An AI hallucination is a confident but false statement from an AI model, such as a price your product never had, a feature it doesn't offer, or a source that doesn't exist. Hallucinations happen because language models generate the most likely-sounding answer, and they don't check whether it's true.

Also called: hallucination

AI referral traffic

AI referral traffic is website visits that arrive when someone clicks a link inside an AI-generated answer, such as a cited source in ChatGPT, Perplexity, Gemini, Claude, or Microsoft Copilot. Analytics tools identify it by the referring domain or by a UTM tag, like the utm_source=chatgpt.com parameter ChatGPT adds to its links.

Also called: LLM referral traffic, AI traffic

AI search optimization

The umbrella term for improving how a brand appears across AI search surfaces, from ChatGPT and Perplexity to Google AI Overviews and AI Mode.

Also called: AI SEO

AI share of voice

AI share of voice is the percentage of all brand mentions in AI-generated answers that belong to your brand, measured across a fixed set of prompts. Each brand counts once per answer it appears in. If brands are named 200 times across your tracked answers and 50 of those mentions are yours, your share of voice is 25%.

Also called: AI SOV, share of voice in AI search

AI visibility

How often AI engines bring up your brand when people ask about your category, and how accurately they describe it.

Also called: AI search visibility, AI brand visibility

AI visibility score

An AI visibility score is the percentage of tracked AI answers that mention your brand. You divide the answers that name you by every answer in your tracked prompt set, including answers that name no brand at all. If your brand appears in 30 of 100 tracked answers, your AI visibility score is 30%.

AI visibility tracking

Measuring AI visibility over time by running the same prompts across engines and comparing mentions, citations, and positions from run to run.

Answer engine

A search tool that replies with a written answer instead of a list of links. ChatGPT, Perplexity, Gemini, and Google AI Overviews are all answer engines.

Also called: AI answer engine

Answer engine optimization (AEO)

The practice of making your brand and content easy for AI answer engines to find and cite when they answer a question.

Also called: AEO

Branded vs unbranded prompts

Branded prompts are AI questions that name your company, such as "Is Acme good for agencies?" Unbranded prompts describe a need without naming any brand, such as "best project software for agencies." Tracking unbranded prompts shows whether AI engines recommend you to buyers who don't know you yet.

Also called: branded prompts, unbranded prompts, non-branded prompts

ChatGPT

OpenAI's AI assistant. It answers from its training data and, when search is used, from live web results it retrieves and cites.

ChatGPT Shopping

Product results inside ChatGPT answers, shown as cards with images, prices, and links to merchants.

Claude

Anthropic's AI assistant. It can search the web and cite sources when a question needs current information.

Content chunking

Content chunking is the way AI search systems split a web page into smaller passages so each one can be retrieved and quoted on its own. The term also describes writing pages in clear, self-contained sections, so any passage an engine pulls out still makes sense without the rest of the page.

Also called: chunking, passage ranking, passage slicing

Content freshness

Signals that tell search engines and AI systems a page is current, such as a visible update date and recent data.

Context window

A context window is the maximum amount of text an AI model can work with at once, measured in tokens. It holds everything in play for a single response: the instructions, the conversation so far, any web pages or documents the system retrieved, and the model's own reply. Anything that doesn't fit is left out.

Also called: context length

Conversational search

Searching by asking questions in natural language and refining them through follow-ups, with the engine keeping the context of the conversation.

Domain authority

A score created by Moz that estimates how likely a domain is to rank, based mostly on its backlinks. It's a third-party metric that Google doesn't use.

E-E-A-T

Experience, expertise, authoritativeness, and trustworthiness: the qualities Google's quality rater guidelines use to judge whether content deserves to be trusted.

Also called: EEAT, experience, expertise, authoritativeness, trustworthiness

Earned media

Earned media is coverage of your brand that other people publish without being paid, such as press articles, reviews, forum discussions, social posts, and mentions in industry roundups. It sits alongside owned media (channels you control, like your website) and paid media (advertising).

Embeddings

Embeddings are lists of numbers that represent the meaning of a piece of content, such as a paragraph of text or an image. Content with similar meaning gets similar numbers, so a system can find a relevant passage even when it shares no words with the question. Embeddings are what make semantic search and AI retrieval work.

Also called: vector embeddings

Entity SEO

Making sure search engines and AI models recognize your brand as a distinct entity, with consistent facts about what it is and what it's connected to.

Also called: entity optimization

Featured snippet

A highlighted excerpt from one web page that Google shows near the top of the results to answer a query directly.

Gemini

Google's family of AI models and the name of its standalone AI assistant. Gemini models also power AI Overviews and AI Mode.

Generative engine optimization (GEO)

Generative engine optimization (GEO) is the practice of shaping your content and your presence on other sites so AI engines such as ChatGPT and Google AI Overviews include your brand in their answers and cite your pages as sources. The name comes from a 2023 research paper, and in practice it means much the same as answer engine optimization (AEO).

Also called: GEO

Google AI Mode

A conversational search mode inside Google Search that answers with an AI-written response and supports follow-up questions.

Also called: AI Mode

Google AI Overviews

AI-generated summaries that appear above the organic results in Google Search, with links to the pages they draw on.

Also called: AI Overviews, AIO

GPTBot

OpenAI's crawler for collecting model training data. ChatGPT search uses a separate crawler, OAI-SearchBot, so blocking GPTBot in robots.txt doesn't remove you from ChatGPT search results.

Also called: OAI-SearchBot, ChatGPT-User

Grounding

Grounding in AI means tying a language model's answer to specific source material, such as live search results or a company's own documents, so the model isn't relying only on what it learned in training. A grounded answer can point to its sources, which is why AI search engines show citations.

Also called: grounding in AI

Hybrid search

Retrieval that combines keyword matching with vector search, so results match both the exact terms and the meaning of a query.

Information gain

Information gain is how much new information a page adds beyond what a reader, or a search engine, has already seen on the same topic. The term comes from a Google patent that describes scoring documents this way. In practice, it means original data, firsthand experience, or specific details that other pages leave out.

Also called: information gain score

Knowledge cutoff

A knowledge cutoff is the date after which an AI model has no training data. The model can't know about anything newer, such as a product launch or a price change, unless it searches the web while it answers. Each model has its own cutoff, and newer model versions move it later.

Also called: training cutoff

Knowledge graph

A database of entities and the relationships between them. Google's Knowledge Graph powers knowledge panels and helps AI systems resolve which brand a question is about.

Large language model (LLM)

An AI model trained on large amounts of text to predict and generate language. LLMs are the engines inside ChatGPT, Claude, Gemini, and most AI search tools.

Also called: LLM

LLM SEO

Search optimization aimed at large language models, covering both your own pages and the third-party sources models draw on when they mention you.

Also called: LLM optimization, LLMO

llms.txt

llms.txt is a proposed standard for a markdown file at a website's root (example.com/llms.txt) that summarizes the site and links to its most useful pages for language models. Jeremy Howard of Answer.AI proposed it in September 2024. Google says its search ignores the file, and no other major AI search engine has said it uses one.

Microsoft Copilot

Microsoft's AI assistant, built into Windows, Edge, Bing, and Microsoft 365. It grounds web answers in Bing search results.

Also called: Copilot

Model Context Protocol (MCP)

The Model Context Protocol (MCP) is an open standard for connecting AI applications to outside tools and data. An MCP server describes what it can do in a format any MCP-compatible assistant understands, so one integration can work across AI apps like Claude and Cursor without custom code for each.

Also called: MCP

Perplexity

An AI answer engine that searches the web for every query and shows numbered citations beside its answer.

Prompt tracking

Prompt tracking is running a fixed set of buyer questions through AI engines like ChatGPT and Perplexity on a schedule, then recording which brands and sources appear in each answer and where they rank. Comparing the same prompts over time shows whether your AI visibility is rising or falling.

Also called: tracked prompts, prompt monitoring

Query fan-out

When an AI engine splits one question into several related searches, retrieves results for each, and combines them into a single answer.

Also called: query fanout, fan-out queries

Retrieval-augmented generation (RAG)

Retrieval-augmented generation (RAG) is a technique where an AI model first searches a set of documents for passages that match a question, then writes its answer from those passages. Because it looks things up at answer time, the model can use up-to-date information and point to where it came from.

Also called: RAG

Search Generative Experience (SGE)

Google's 2023 Search Labs experiment with AI-generated answers. It became AI Overviews when Google rolled the feature out in May 2024.

Also called: SGE

Semantic search

Semantic search is a way of finding information by matching the meaning of a query instead of its exact words. It lets a search engine or AI assistant return a page about "affordable CRM for startups" when someone asks for a "cheap customer database for a new company," even though the two phrases share almost no words.

Server-side rendering (SSR)

Server-side rendering (SSR) means the server builds a page's full HTML before sending it, so the content is there in the first response. With client-side rendering (CSR), the browser gets a near-empty page and JavaScript builds the content afterward. Crawlers that don't run JavaScript, including most AI crawlers, see only what's in that first response.

Also called: SSR, client-side rendering, CSR

Sources vs citations

In AI search, sources are the pages an AI engine used to write its answer, and citations are the links that actually appear in the answer for the reader to see. The two overlap, but they aren't the same: an engine often reads more pages than it shows, so a page can shape an answer without being credited in it.

Also called: AI sources, source rate

Structured data

Code, usually schema.org JSON-LD, that labels what a page contains, such as a product or an FAQ, in a format machines can parse.

Also called: schema markup, schema.org

Tokenization

Breaking text into tokens, the small units of words and characters that language models read and generate.

Also called: tokens

Topical authority

How thoroughly a site covers a subject, judged across all its pages rather than any single one.

Universal Commerce Protocol (UCP)

An open standard backed by Google that lets AI agents discover products and complete purchases with merchants through a common interface.

Also called: UCP

Unlinked brand mentions

An unlinked brand mention is any place your brand is named without a link to your website: in an article, review, forum post, or AI-generated answer. In SEO, unlinked mentions are a link-building opportunity. In AI search, they count on their own, because AI engines read brand names in text whether or not they're linked.

Also called: unlinked mentions, linked vs unlinked mentions

Vector search

Vector search finds content by meaning instead of exact words. It turns text into embeddings, lists of numbers that represent what the text means, then returns the stored items whose embeddings sit closest to the query's. That's how a search for "cheap flights" can find a page about "budget airfare."

Also called: vector database

WebMCP

A proposed web standard that lets a website expose actions, such as search or checkout, as tools an in-browser AI agent can call directly.

XML sitemap

A file that lists the URLs on a site you want search engines to crawl, often with the date each was last changed.

Zero-click search

A search that ends without a click to any website, because the answer is shown directly on the results page or in an AI response.