Skip to content
BrandWater AI

The AI visibility glossary.

55 plain language definitions of the terms used in AI search, measurement and optimization. Each one starts with a direct definition, then explains why it matters.

A

AI crawlerSources
An AI crawler is an automated program, run by an AI company, that fetches web pages to train models, build a search index or answer a user's request.
AI ModePlatforms
AI Mode is a conversational search experience in Google Search that answers longer questions and follow ups with AI.
AI OverviewsPlatforms
AI Overviews are AI generated summaries that Google shows at the top of some search results, with links to supporting pages.
AI referral trafficMeasurement
AI referral traffic is website visits that arrive from AI assistants or AI search products, usually identified by the referring domain.
AI visibilityConcepts
AI visibility is how often and how favourably a brand appears in AI generated answers, recommendations, comparisons and cited sources for the questions its customers ask.
AI visibility toolConcepts
An AI visibility tool is software that runs a set of questions against AI platforms such as ChatGPT, Gemini and Perplexity on a schedule, then reports how often, how favourably and alongside which sources a brand is named in the answers.
Answer engine optimization (AEO)Optimization
Answer engine optimization is the practice of structuring content so that search engines and AI assistants can extract a direct, accurate answer from it.
Answer-first contentOptimization
Answer-first content states the conclusion at the start of a page or section and then supports it with detail, so a passage can be quoted without losing its meaning.

B

Brand IntelligenceProduct
Brand Intelligence is a structured, evidence backed record of a brand, in which every fact carries its source, a confidence level and a review status.
Brand mentionMeasurement
A brand mention is any occurrence of a brand's name in an AI generated answer.

C

Canonical tagSources
An HTML tag that tells search engines and crawlers which version of a page, among near-duplicates, is the one to treat as authoritative.
Citation coverageSources
Citation coverage is the share of AI answers that carry a visible source link back to a brand's website.
Citation rateSources
Citation rate is the share of eligible AI answers that show a link to a brand's own pages as a source.
Co-mentionMeasurement
A co-mention is when two or more brands are named in the same AI answer.
Content freshnessOptimization
How recently a page was meaningfully updated, as distinct from a cosmetic timestamp change.
Context windowConcepts
The amount of text a language model can consider at once when generating a response, measured in tokens.

D

Digital PROptimization
Earning coverage, links and mentions from independent publications and sites, as distinct from publishing content on channels a brand controls directly.

E

E-E-A-TConcepts
Google's framework for judging content quality: Experience, Expertise, Authoritativeness and Trustworthiness. It is not a ranking factor you can switch on directly, but a lens reviewers and some automated systems use to judge whether content, and the person or brand behind it, is credible.
Eligible observationMeasurement
An eligible observation is an answer that is relevant to the brand and topic being measured, and therefore counts in the denominator of a rate.
EntityConcepts
An entity is a distinct thing that search and AI systems recognise, such as a company, product or person, and can link facts to.
Entity resolutionConcepts
The process by which a system determines that different names, mentions or records refer to the same real-world entity.

G

Generative engine optimization (GEO)Optimization
Generative engine optimization is the practice of improving how often and how accurately a brand or page is used in answers written by generative AI systems.
GroundingConcepts
Grounding is connecting an AI model's answer to external sources, such as search results, so that claims can be supported and cited.

H

Hallucination (in brand answers)Concepts
A hallucination is an AI statement that sounds plausible but is false or unsupported.

I

IntentConcepts
Intent is what a person is trying to do when they ask a question, such as discovering options, comparing them or deciding.
Internal linkingOptimization
The practice of linking between a site's own pages in a deliberate structure, so crawlers and readers can find related content.

K

Knowledge graphConcepts
A structured database of entities, people, places, organizations, products, and the relationships between them, used by search engines and some AI systems to resolve who or what something refers to.
Knowledge panelConcepts
The information box, usually on the right side of a search results page, that summarizes facts about an entity drawn from a knowledge graph.

L

Large language model (LLM)Concepts
A large language model is an AI system trained on very large amounts of text to predict and generate language.
llms.txtOptimization
llms.txt is a proposed convention for a plain text file at a site's root that summarises the site and links to its key pages for language models.

M

Mention rateMeasurement
Mention rate is the share of eligible AI answers that name a brand.
Multimodal searchConcepts
Search or AI interaction that combines more than one type of input, such as text, images or voice, in a single query.

O

ObservationMeasurement
An observation is one stored answer to one prompt on one AI platform at one point in time.
OmissionMeasurement
An omission is a confirmed attribute of a brand that AI answers consistently leave out.

P

Position (in AI answers)Measurement
Position in AI answers is where a brand appears when an answer names several options, and how it is described there.
Product feedSources
A structured file, often XML or CSV, listing a retailer's products with attributes like price, availability and specifications, originally built for shopping ads and comparison engines.
PromptConcepts
A prompt is the question or instruction a person gives an AI system.
Prompt setMeasurement
A prompt set is a defined, stable list of prompts, grouped by topic and intent, that is rerun over time to measure AI visibility.

Q

Query fan-outPlatforms
Query fan-out is when an AI search system breaks one question into several related sub-queries, runs them and combines the results into one answer.

R

Recommendation rateMeasurement
Recommendation rate is the share of eligible AI answers that actively recommend a brand, rather than only naming it.
Retrieval augmented generation (RAG)Concepts
Retrieval augmented generation is a technique in which a system retrieves relevant documents and gives them to a language model to use when writing an answer.
robots.txtSources
A plain text file at the root of a website that tells crawlers, including AI crawlers, which parts of the site they may or may not access.

S

sameAs (schema.org)Optimization
sameAs is a schema.org property that lists other official web pages describing the same entity, such as verified profiles, so that machines can connect them.
Sample sizeMeasurement
Sample size is the number of observations that sit behind a reported rate.
Sentiment (interpreted)Measurement
Sentiment is a label, such as neutral or positive, that describes the tone with which an AI answer treats a brand.
Share of voice (in AI answers)Measurement
Share of voice in AI answers is a brand's mentions divided by all tracked brand mentions across the same set of answers.
Source gapSources
A source gap is a domain or page that AI answers cite repeatedly for a topic but that does not mention or link to a brand.
Structured dataOptimization
Structured data is markup, commonly schema.org in JSON-LD, that describes the meaning of a page's content in a form machines can read.

T

TopicConcepts
A topic is a subject a brand wants to be known for, such as invoicing for freelancers.
Topical authorityOptimization
Topical authority is the depth and consistency with which a site and its wider web presence cover a subject, so that systems treat it as a reliable source on that topic.
Training cutoffConcepts
The point in time after which a language model has no direct knowledge from its training data, because that data was collected before the cutoff date.

V

Verification (of an action)Product
Verification is rerunning the same observation set after a change and comparing the result with the earlier baseline.
Voice searchConcepts
A spoken query to a voice assistant, which typically returns a single answer read aloud rather than a list of results.

X

XML sitemapSources
A file listing a website's pages in a machine-readable format, submitted to help search engines and crawlers discover and prioritize content.

Z

Zero-click searchConcepts
A zero-click search is one where the person gets what they need from the results page or an AI summary and does not click through to a website.

Measure these terms on your own brand.

Start with a first audit. We show where you appear, who is named instead, and what to fix first.