AI Search Glossary

A

AEO

Answer Engine Optimization.

AEO is the practice of making information easier for search and AI systems to understand and use when answering questions.

It often emphasizes:

• direct answers
• clear definitions
• question-led sections
• logical structure
• useful supporting context

AEO complements SEO rather than replacing it.

Related guide:

/learn/aeo/

AI Citation

A visible source reference or link associated with information in an AI-generated answer.

Citations may appear as:

• inline links
• numbered references
• source cards
• source panels
• publisher links

A citation does not automatically mean the platform endorses the entire source.

Related guide:

/learn/ai-search-citations/

AI Crawler

A broad term for an automated web crawler associated with an AI-related product.

Different crawlers may support different purposes including:

• search discovery
• indexing
• training-related use
• grounding
• other retrieval functions

Do not assume every AI crawler serves the same purpose.

Related guide:

/learn/ai-crawlers/

AI Mode

Google AI Mode is an AI-powered Google Search experience designed for deeper, conversational and multi-part questions.

Google describes AI Mode as capable of exploring multiple related subtopics through techniques such as query fan-out.

Related guide:

/learn/google-ai-mode/

AI Overviews

AI-generated summaries that can appear within Google Search.

They are designed to help users understand a topic while providing links that allow further exploration of web sources.

Related guide:

/learn/google-ai-overviews/

AI Search

A broad term for search or discovery experiences that use generative AI to help users find, synthesize or explore information.

Examples can include:

• ChatGPT Search
• Google AI Mode
• Google AI Overviews
• Perplexity

AI search does not refer to one universal technology or ranking system.

AI Search Optimization

The broader process of improving how a brand and its information perform across traditional and AI-powered search environments.

It can incorporate:

• technical SEO
• AI SEO
• GEO
• AEO
• entity clarity
• authority
• original content
• measurement

Related guide:

/learn/ai-search-optimization/

AI SEO

A broad optimization approach focused on improving how brands and websites are discovered, understood, trusted and surfaced across traditional search engines and AI-powered search experiences.

It builds on SEO rather than replacing it.

Related guide:

/learn/ai-seo/

AI Visibility

The observed presence of a brand, website, product, expert or source across AI-powered search and answer environments.

Visibility may include:

• mentions
• recommendations
• citations
• source links
• comparison appearances

Related guide:

/learn/ai-visibility/

AI Visibility Score

A calculated metric intended to summarize AI-search visibility.

There is no universal industry-standard AI visibility score.

Any score should explain:

• platforms
• queries
• sample
• weighting
• methodology
• measurement period

A proprietary score should not automatically be interpreted as total market visibility.

Answer Engine

A general term sometimes used for systems that respond to questions directly rather than only displaying a traditional list of webpages.

The term is broad and does not describe one specific technical architecture.

B

Brand Mention

An occurrence where an AI-generated response refers to a company, product, organization or person by name.

A mention is different from:

• a citation
• a recommendation

A brand can be mentioned without its website being linked.

Branded Query

A search or AI question containing the name of a company, person, product or organization.

Example:

“What does AI SEO Experts Canada research?”

Branded and non-branded visibility should often be measured separately.

C

Canonical URL

The preferred URL representing a page when multiple URLs contain the same or very similar content.

Canonicalization helps search engines understand which version should generally be treated as primary.

A canonical tag is not a method for making weak duplicate content valuable.

ChatGPT Search

OpenAI's web-search capability within ChatGPT.

It can retrieve current web information and provide source links or citations.

For publishers, OAI-SearchBot is relevant to search discovery.

Related guide:

/learn/chatgpt-search/

Citation Rate

The proportion of monitored queries in which a website or domain receives a visible citation.

Example methodology:

queries with domain citation

divided by

total monitored queries.

Always explain the query set and platform when reporting a citation rate.

Crawlability

The ability of a crawler to access a webpage.

Crawlability can be affected by:

• robots.txt
• authentication
• firewalls
• server errors
• CDN settings
• broken links

A crawlable page is not automatically indexable or citation-worthy.

Crawler

Software that automatically requests and discovers content on websites.

Search engines and AI-related products can operate crawlers for different purposes.

Related guide:

/learn/ai-crawlers/

D

Domain Citation

A source appearance where any URL from a particular domain is cited by an AI-search experience.

Domain-level tracking is useful, but page-level citation tracking is usually more informative because it shows which specific content is being selected.

Doorway Page

A page created primarily to target a slightly different keyword or location without providing sufficiently distinct value.

For example, producing hundreds of nearly identical city pages with only the city name changed can create poor-quality doorway-like content.

AI SEO Experts Canada should avoid this approach.

E

Entity

A distinct person, organization, place, product, service, event or concept that can be identified and related to other information.

Examples:

• Google
• Toronto
• AI SEO Experts Canada
• Generative Engine Optimization

Entity understanding goes beyond exact keyword matching.

Entity Clarity

How clearly a person, company, service or other entity is represented across a website and the wider web.

Useful entity clarity includes consistent information about:

• name
• role
• services
• location
• relationships
• expertise

Entity Optimization

The process of reducing ambiguity around entities and their relationships.

For a business this can involve:

• clear organization information
• consistent descriptions
• author identities
• relevant structured data
• internal linking
• credible external profiles

Related guide:

/learn/entity-optimization/

G

Gemini

Google's AI assistant.

Gemini is distinct from Google Search, although Google's products and underlying technologies can interact in different ways.

Depending on the Gemini experience, responses may provide related public web sources.

Related guide:

/learn/gemini/

Generative Engine

A broad, non-standard term referring to a system that generates synthesized responses rather than presenting only traditional search listings.

Different products may rely on very different retrieval, ranking and generation systems.

Generative Engine Optimization

See GEO.

Generative Search

A search experience that uses generative AI to synthesize information into an answer or interactive result.

Generative search often combines:

• retrieval
• search
• ranking
• generation
• source presentation

The exact architecture differs by platform.

GEO

Generative Engine Optimization.

GEO focuses on improving how useful and understandable information is for generative search and answer systems.

It commonly emphasizes:

• source usefulness
• clarity
• authority
• evidence
• originality
• entity relationships

Related guide:

/learn/geo/

Googlebot

Google's crawler used for Google Search.

Google's AI features within Search, such as AI Overviews and AI Mode, continue to rely on the normal Google Search crawling foundation.

Google-Extended

A robots.txt control provided by Google for specified uses involving Gemini model training and grounding.

Google states that Google-Extended does not affect whether a site is included or ranked in Google Search.

It should not be confused with Googlebot.

Grounding

The process of connecting a generated response to external or supplied information so the output can be informed by relevant evidence or context.

Grounding can use sources such as:

• web search
• databases
• documents
• enterprise data

Grounding does not automatically guarantee that every generated statement is correct.

I

Index

A structured collection of discovered information used by a search or retrieval system.

A search engine can crawl content and store information about it in an index for future retrieval.

Indexability

Whether a page is eligible to be included in a search system's index.

A page can be crawlable but intentionally non-indexable, for example through a supported noindex directive.

Information Gain

A useful editorial concept describing the amount of new or differentiated information a page contributes compared with what is already available.

It should not be treated as one publicly confirmed universal AI-ranking factor.

Original:

• data
• experience
• research
• analysis

can increase the practical information value of a page.

Internal Link

A hyperlink connecting one page of a website to another page on the same website.

Internal links help users navigate and can reinforce relationships among:

• topics
• services
• entities
• research
• authors

K

Knowledge Graph

A structured representation of entities and relationships between them.

Knowledge graphs can help systems connect concepts such as:

Company

→ founder

Company

→ location

Person

→ organization

Product

→ manufacturer

Not every AI-search system uses one knowledge graph in the same way.

L

Large Language Model

Often abbreviated LLM.

A machine-learning model trained to work with language and, depending on the model, other forms of information.

LLMs can generate, summarize, classify, reason over or transform text.

An LLM by itself should not be assumed to contain current web information unless it is connected to retrieval or other current sources.

LLM

See Large Language Model.

llms.txt

A proposed website convention intended to make information easier for language-model-related systems to discover or interpret.

It should not be treated as a universal AI-search ranking requirement.

Support varies by platform.

Website owners should prioritize officially documented crawler and search controls.

M

Mention Rate

The proportion of monitored AI-search queries in which a brand or entity is mentioned.

Example:

20 brand mentions

across

100 monitored queries

would equal a 20% observed mention rate for that study.

It does not mean the brand owns 20% of all AI-search visibility.

Model Training

The process used to develop or improve a machine-learning model from data.

Model-training crawler controls and search-indexing crawler controls can be different.

For example, OpenAI distinguishes GPTBot-related training controls from OAI-SearchBot search discovery.

N

noindex

A directive used to tell supported search systems not to include a webpage in their index.

A system may need to crawl the page to detect the directive.

Do not use noindex accidentally on pages intended for organic discovery.

nosnippet

A Google Search preview control that can restrict snippets from a page.

Because Google's AI features are part of Search, Google documents established preview controls as mechanisms website owners can use to limit information displayed from pages.

Use these controls intentionally.

O

OAI-SearchBot

OpenAI's crawler associated with ChatGPT Search discovery.

Publishers who want eligible public content to be discoverable in ChatGPT Search should review whether OAI-SearchBot is permitted.

Allowing it does not guarantee a citation or placement.

Related guide:

/learn/chatgpt-search/

Organic Search

Unpaid search visibility generated through search-engine results rather than advertising placement.

Traditional organic SEO remains an important foundation even as AI-powered search experiences expand.

P

Perplexity

An AI-powered search and answer platform that searches web sources and commonly presents citations within generated responses.

Related guide:

/learn/perplexity/

PerplexityBot

Perplexity's crawler associated with its search index.

Perplexity currently states that PerplexityBot respects robots.txt directives.

Primary Source

The original or authoritative source for a piece of information.

Examples can include:

• original research
• government publication
• official platform documentation
• first-party dataset
• academic paper reporting the original study

When possible, cite the primary source instead of a webpage that merely summarizes it.

Prompt

An instruction, question or input provided to an AI system.

For AI-search measurement, exact prompt wording should be recorded because wording can influence results.

Prompt Set

A defined group of prompts used for testing or measurement.

For search research, “query set” is often a clearer term when the prompts represent search or discovery questions.

Q

Query

A search request or question submitted by a user.

In AI search, queries can be long, conversational and multi-part.

Query Fan-Out

A technique Google describes for AI Mode in which a complex user question can be explored through multiple related searches or subtopics.

This allows an AI-search system to investigate different aspects of the original information need.

Related guide:

/learn/google-ai-mode/

Query Set

A predefined group of queries used for measurement or research.

A stable query set makes longitudinal AI visibility studies more reproducible.

Related guide:

/learn/ai-search-measurement/

R

Recommendation Visibility

A form of AI visibility where an AI-generated answer presents a company, professional, product or service as a potential choice for the user's need.

Recommendation visibility is different from an informational brand mention.

Retrieval

The process of finding relevant information from an external collection, index, database or search system.

Retrieval can provide information that is then used to generate an AI response.

Retrieval-Augmented Generation

Often abbreviated RAG.

A technical pattern where a generative model receives relevant information retrieved from an external source before generating a response.

The external source could include:

• documents
• databases
• search results
• private knowledge bases

Not every AI-search architecture should automatically be described as identical RAG.

RAG

See Retrieval-Augmented Generation.

Robots.txt

A website file containing crawling instructions for compliant automated user agents.

It is used for crawler management.

It is not a security mechanism.

Related guide:

/learn/ai-crawlers/

S

Schema Markup

A common term for structured data embedded in webpages using vocabularies such as Schema.org.

Structured data can help machines interpret information about:

• organizations
• people
• articles
• events
• breadcrumbs

Schema markup does not guarantee AI citations or rankings.

Search Intent

The underlying purpose behind a query.

Common broad categories can include:

• informational
• navigational
• commercial investigation
• transactional
• local

AI-search queries can combine several intents in one conversational question.

Search Visibility

The degree to which a website or brand appears across search experiences for relevant queries.

Traditional search visibility and AI visibility can overlap but are not identical.

Share of AI Visibility

A comparative metric representing an entity's proportion of observed appearances within a defined study.

It should always specify:

• platform
• query set
• sample
• time period
• methodology

It should not be presented as universal market share.

Source

A webpage, document, dataset or other information resource used or presented in connection with an answer.

Different AI systems handle sources differently.

Source Visibility

A form of AI visibility where a website or page appears as a cited or linked source even when the brand itself is not prominently discussed.

Source visibility is especially important for publishers and research organizations.

Structured Data

Machine-readable information embedded in or associated with webpages.

JSON-LD is a commonly used format on the web.

Structured data should accurately represent visible, legitimate information.

T

Topical Authority

A broad SEO concept describing the strength and depth of a website's expertise or relevance around a topic.

It can be supported by:

• useful content
• connected topic coverage
• internal linking
• original research
• expert authorship
• external recognition

It should not be reduced to publishing as many pages as possible.

Training Crawler

A crawler or control associated with collecting or governing data for potential model-training use.

Training-related controls should not automatically be assumed to govern AI-search discovery.

Crawler purposes must be evaluated individually.

U

User Agent

An identifier sent in an HTTP request that can indicate which browser, crawler or software is making the request.

robots.txt commonly uses user-agent names to apply crawler-specific rules.

User-agent strings can be spoofed, so identity verification may require additional methods.

V

Vector Search

A retrieval approach that represents information numerically so items can be found based on semantic similarity rather than exact keyword matching alone.

Vector search is widely used in modern information-retrieval systems.

However, a website owner generally does not need to “optimize vector embeddings” directly for public AI-search visibility.

Useful, clear and relevant information remains the practical priority.

Visibility

The degree to which an entity or source appears in a discovery environment.

In AI search, visibility can include:

• mentions
• citations
• recommendations
• source links

Always define what kind of visibility is being measured.

W

Web Search Grounding

A process where an AI system uses web-search information to help inform a generated response.

Grounding can improve access to current or externally verifiable information.

The implementation differs among platforms.

How to Use This Glossary

Terms in AI search evolve quickly.

When reading an article, report or software dashboard, ask:

• How is this term defined?
• Which platform does it refer to?
• Is the term an official product name or an industry label?
• What methodology is behind the metric?
• Is the claim based on official documentation or interpretation?

Terminology is useful only when it improves understanding.

Important Terms That Are Often Confused

AI SEO vs GEO

AI SEO is typically used as the broader optimization category.

GEO focuses more specifically on generative-answer visibility and source usefulness.

GEO vs AEO

GEO emphasizes generative source visibility.

AEO emphasizes clear answers to questions.

The practices overlap.

Mention vs Citation

A mention names an entity.

A citation visibly references a source.

They can occur independently.

Citation vs Recommendation

A citation identifies supporting information.

A recommendation presents an entity as a potential option.

A cited business is not automatically being recommended.

Crawler vs Model

A crawler fetches information.

A model processes and generates information.

They are not the same thing.

OAI-SearchBot vs GPTBot

OAI-SearchBot is associated with ChatGPT Search discovery.

GPTBot is controlled separately for potential model-training use.

Googlebot vs Google-Extended

Googlebot supports Google Search crawling.

Google-Extended is a separate control for specified Gemini-related training and grounding uses and does not determine Google Search inclusion.

Crawlability vs Indexability

Crawlability describes whether a crawler can access a page.

Indexability describes whether the page is eligible for inclusion in an index.

A page can be crawlable but non-indexable.

Frequently Asked Questions About AI Search Terminology

Is AI SEO an official Google term?

AI SEO is an industry term used to describe optimization for AI-powered discovery alongside traditional search. Individual platforms use their own official product terminology.

Are GEO and AEO the same thing?

No, although they overlap significantly. GEO focuses more on generative source visibility, while AEO focuses more directly on useful answers.

Is AI visibility the same as traffic?

No. AI visibility can occur without a website click.

Is an AI citation the same as a backlink?

Not necessarily. A visible AI source link and a conventional webpage hyperlink are different contexts, even though both may lead users to a website.

Does being crawled mean a website will appear in AI search?

No. Crawling only addresses access or discovery. Selection depends on the platform and query.

Is RAG the same thing as AI search?

No. RAG is one technical pattern. AI-search products can use more complex and varied retrieval and generation architectures.

Do all AI systems use the same ranking factors?

No. There is no publicly confirmed universal AI ranking formula shared across all platforms.

Continue Learning

AI SEO

Start with the broad optimization framework.

AI Search Measurement

Learn how to measure AI-search visibility systematically.

AI Crawlers

Understand how different crawlers access public web content.