AI search glossary

Plain definitions of the terms used in AI search and AI visibility: how brands are measured in AI answers, how those answers are produced, and how to tell whether a change made a difference.

New to the vocabulary? Start with the three labels for the same work, AEO, GEO and LLMO, then the measures that tell you whether it worked: AI share of voice, citation share and the P25-P75 range.

• Agentic commerce

  Agentic commerce is shopping in which an AI agent acts for a person, researching products, comparing options and, in some cases, completing a purchase on their behalf.

• AI citation decay

  AI citation decay is the gradual loss of citations a page or brand receives in AI-generated answers over time, often while the page itself has not changed.

• AI crawler

  An AI crawler is an automated bot that fetches web pages for an AI company, either to collect training data or to retrieve pages in real time to answer a person's question.

• AI Mode

  AI Mode is a conversational search experience in Google Search that answers a question with an AI-generated response and supporting links, and lets people ask follow-up questions.

• AI Overviews

  AI Overviews are AI-generated summaries that Google Search can show at the top of some results pages, with links to web pages that support the summary.

• AI referral traffic

  AI referral traffic is the visits that arrive on your website from links in AI assistants and AI search tools.

• AI search audit

  An AI search audit is a structured check of how AI engines describe, mention and cite a brand for the questions its buyers ask, usually compared with its competitors.

• AI share of voice

  AI share of voice is the portion of brand mentions in AI-generated answers that go to your brand, compared with the competitors named in the same answers, across a defined set of prompts.

• Answer engine optimization (AEO)

  Answer engine optimization (AEO) is the practice of making a brand's content easy for AI-powered answer engines to find, understand and cite, so the brand is named in the answers people receive.

• Answer variability

  Answer variability is the way an AI engine can give different answers to the same prompt when it is asked more than once, including naming different brands or citing different sources.

• Attribution window

  An attribution window is the period after a change during which an observed effect is credited to that change.

• Citation half-life

  Citation half-life is the time it takes for the citations a piece of content earns in AI-generated answers to fall to half of an earlier level.

• Citation share

  Citation share is the percentage of the citations shown in AI answers for a query that point to your site, out of all citations shown for that query. Microsoft reports it in Bing Webmaster Tools, and the same ratio can be computed for any engine that shows its sources.

• Generative engine optimization (GEO)

  Generative engine optimization (GEO) is the practice of improving how a brand appears in responses produced by generative AI systems, such as AI chat assistants and AI features in search.

• Grounding query

  A grounding query is a search an AI system runs to find web sources that support and cite its answer, as distinct from the prompt the person typed.

• Holdout control

  A holdout control is a group of items deliberately left untouched by a change, so that their movement can be compared with the items that were changed.

• LLMO

  LLMO (large language model optimization) is the practice of improving how large language models, such as those behind ChatGPT and Gemini, describe and recommend a brand, by publishing accurate, consistent information both in the data models learn from and in the pages they retrieve when answering.

• llms.txt

  llms.txt is a proposed standard for a Markdown file, usually placed at a site's root, that gives large language models a concise guide to the site's most useful content.

• Mention rate

  Mention rate is the share of AI answers, across a set of prompts, that mention your brand at all.

• P25-P75 range

  The P25-P75 range is the span between the 25th and 75th percentiles of a set of readings: the middle half of the values, with a quarter of readings below it and a quarter above it.

• Query fan-out

  Query fan-out is a technique in which an AI search system breaks one question into several related searches, gathers results for each, and combines them into a single answer.

• Share of answer

  Share of answer is a measure of how much of the AI-generated answers in a category feature your brand, used by some teams as another name for AI share of voice.

• Unlinked brand mentions

  Unlinked brand mentions are references to your brand name on other websites that do not include a hyperlink to your site.

• Zero-click search

  A zero-click search is a search that ends without the person clicking through to any website, because the answer was shown on the results page or in an AI-generated response.