The language of AI search 153 terms

The AI Search Glossary

A clearer understanding of AI search starts here. Explore the terms behind GEO, answer engines, citations and brand visibility.

153 definitions

A terms

38 terms

Agent Experience (AX)

Agent experience describes how easy your website and data are for an AI agent to read, understand, and act on. It is user experience built for software instead of people, and it is quietly becoming part of how brands win in AI search.

Agentic Checkout

Agentic checkout happens when an AI agent completes a purchase on a person's behalf, handling payment right inside the conversation instead of sending the shopper to a website. It marks the moment agentic commerce turns from research into revenue.

Agentic Commerce

Agentic commerce is shopping carried out by an AI agent on someone's behalf. The agent researches, compares, and sometimes completes the purchase based on the goal it was given, and the buyer may never visit a store directly.

Agentic Commerce Protocol (ACP)

The Agentic Commerce Protocol is an open standard built by OpenAI and Stripe that lets merchants sell directly inside AI conversations. It is the technology behind ChatGPT's Instant Checkout and one of the main rails turning answer engines into storefronts.

Agentic RAG

Agentic RAG is a more advanced form of retrieval augmented generation, where an AI agent helps plan, retrieve, evaluate, and use information across complex, multi step tasks rather than a single lookup. The term is still evolving as agent based workflows mature.

Agentic Search Optimization

Agentic search optimization is the work of preparing a brand's content, structured data, and site behavior so autonomous AI agents can find, evaluate, and act on it correctly. It goes beyond making content readable for a person, since an agent needs it to be usable by a program.

Agentic Storefront

An agentic storefront is a sales channel that automatically makes a brand's products available to AI shopping platforms, so they can be discovered and bought through AI agents without a custom integration for each one.

agents.md

agents.md is an emerging convention where a markdown file gives AI agents instructions and context for working with a project or a site. It is most common in software repositories, where it tells coding agents how the codebase is organized.

AI Answer Attribution

AI answer attribution is how a brand gets credited when an AI system uses its content in an answer. Levels range from being named with a link, which is the strongest, down to a bare footnote or no credit at all.

AI Answer Engine

An AI answer engine writes a direct answer instead of handing back a list of links. Perplexity, ChatGPT Search, and Google AI Overviews are the ones most people use today.

AI Answer Share

AI answer share is the percentage of relevant AI generated answers, across a set of tracked prompts, where a given brand shows up at all. It is one of the clearest ways to see whether a brand has a real presence in a category.

AI Attributed Conversion

An AI attributed conversion is a sale or signup that can be traced back to an AI engine, where the customer discovered or decided on a brand through an AI answer before converting. It is what turns AI visibility from a brand metric into a revenue line.

AI Brand Monitoring

AI brand monitoring is the ongoing practice of tracking how a brand is mentioned, described, and cited across multiple AI platforms over time, rather than checking it once and moving on. It reveals whether visibility is improving or slipping as models and content change.

AI Brand Sentiment

AI brand sentiment is the tone an AI engine uses when it describes a brand, positive, neutral, or negative. It is about how a brand gets talked about, not only whether it is mentioned.

AI Citation Tracking

AI citation tracking is the practice of measuring which specific domains and pages an AI system cites or links to when it generates answers involving a brand or topic. It shows exactly which sources are shaping how a brand is represented.

AI Crawler

An AI crawler is a bot that fetches web pages to feed an AI system, either for training or for answering live questions. Common ones include GPTBot, ClaudeBot, PerplexityBot, OAI SearchBot, and Google Extended.

AI Generated Answer

An AI generated answer is the written response a model gives to a question, built from its training plus whatever it retrieves live, instead of a list of links. It is the basic unit that GEO work is trying to influence.

AI Mode

AI Mode is Google's fully conversational search experience, where you ask a complete question and get a synthesized, multi part answer with citations. It goes a step further than AI Overviews, which sit on top of the regular results page.

AI Overview

AI Overview is Google's AI written summary shown at the top of some results pages, built by running sub queries against retrieved pages and citing the sources it drew from. It grew out of Google's earlier Search Generative Experience.

AI Search Impression

An AI search impression is a single instance of a brand, piece of content, or URL turning up inside an AI generated answer. It is essentially the AI era version of a traditional search impression.

AI Search Ranking

AI search ranking is the hidden order an engine uses when deciding which sources to lean on while writing an answer. There is no numbered list to see, but some sources clearly carry more weight than others.

AI Search Traffic

AI search traffic is the visits that arrive from AI products like Perplexity, ChatGPT Search, or Google AI Overviews. It behaves as its own channel, separate from regular organic search traffic.

AI Search Visibility

AI search visibility is how often, and how prominently, a brand appears across the AI systems and prompts that matter to its category. It is the umbrella metric that most GEO work is ultimately trying to move.

AI SEO (AISO)

AI SEO, sometimes shortened to AISO for AI search optimization, is the practice of improving how a brand shows up across AI powered search overall, from AI Overviews to ChatGPT to Perplexity. It largely overlaps with GEO, and the label people choose tends to matter less than the actual work.

AI Shopping Agent

An AI shopping agent is software that shops for a person. It takes a goal, such as finding trail runners under one hundred fifty dollars in a wide fit, then searches, compares, and increasingly buys, often without the shopper visiting a single store.

AI Systems

AI systems are AI powered platforms such as ChatGPT, Google AI Overviews, Perplexity, Claude, and Copilot that interpret a user's query and generate, synthesize, or retrieve an answer to it.

AI Visibility Score

An AI visibility score is a single number meant to summarize how often and how prominently a brand appears across AI answers. Every vendor calculates it differently and rarely publishes the formula, so scores from different tools are best treated as trend lines rather than numbers to compare across platforms.

Answer Engine

An answer engine is a system that responds to a question with a direct, synthesized answer instead of a list of links. ChatGPT, Perplexity, Gemini, and Google's AI Overviews are all answer engines.

Answer Engine Optimization (AEO)

Answer engine optimization is the work of shaping content so an AI engine or voice assistant picks it as the direct answer to a question. It leans heavily on clean question and answer formatting.

Answer First Content

Answer first content states the direct answer to a question right at the top, then expands with context afterward, instead of building up to the point. It matches how AI engines extract information, which makes it far more likely to get lifted into an answer.

Answer Presence

Answer presence measures whether a brand appears at all inside AI generated answers for the queries that matter to its category, service, or name. It is the most basic check before any deeper visibility question makes sense.

Answer Ready Content

Answer ready content is written so an AI system can lift it cleanly into a generated answer without needing to infer meaning from the surrounding page. A tight definition, a direct claim, or a clearly scoped comparison are all common examples.

Answer Units

Answer units are the individual pieces of content an AI system can extract, reuse, or cite clearly, such as a definition, a comparison, a step, a statistic, or a question and answer pair.

Attribution Quality

Attribution quality evaluates how a brand is credited when it is cited in an AI answer. Attribution levels range from named with a link, which is the best outcome, down to named only, a footnote, or no credit at all.

Author Authority

Author authority is the credibility a specific named person carries as a recognized expert, signaled through bylines, credentials, and a consistent publishing track record. AI systems use it as one signal of how much to trust a piece of content.

Authority Content

Authority content is deep, genuinely expert material that covers a topic thoroughly enough to become the reliable source on it. AI engines tend to reach for it because it answers hard questions completely rather than skimming the surface.

B terms

4 terms

Brand Hallucination

Brand hallucination is when an AI system states something about a brand that is not true, such as an invented price, a discontinued product, or a merger that never happened. It usually means the model found too little consistent information and filled the gap on its own.

Brand Mention Optimization

Brand mention optimization is the work of getting a brand mentioned accurately, positively, and often across the web, so models come to associate it with authority even when the mention is not a link.

Business to Agent (B2A)

Business to agent, or B2A, describes the emerging layer of business where the customer is not a person but an AI agent acting on their behalf. It sits alongside B2B and B2C as a third channel, and it changes what a brand needs to publish and expose.

C terms

18 terms

ChatGPT Ads

ChatGPT Ads are OpenAI's sponsored placements inside ChatGPT. They are clearly labeled, shown below or beside an answer, and matched to the context of the conversation rather than to keywords.

ChatGPT Shopping

ChatGPT Shopping is the product discovery and buying experience inside ChatGPT, where users get product recommendations in conversation and can increasingly purchase without leaving the chat. Appearing in the organic recommendations is free, which makes it a large and still open opportunity.

Citation Frequency

Citation frequency is how often AI models cite or reference a specific source when mentioning a brand. Higher citation frequency generally means the AI considers that source authoritative and trustworthy on the topic.

Citation Magnet

A citation magnet is a piece of content built specifically to get referenced again and again, usually original research, a unique dataset, or a definitive resource that others have to point back to.

Citation Optimization

Citation optimization is the work of structuring and formatting content so models are more likely to pick it, reference it, and link back to it. Clear definitions, tidy headings, short summaries, and structured data all help.

Citation Rate

Citation rate is how often a page, domain, or brand gets cited as a source in AI answers. It is shaping up to be a core metric for AI search, in much the same way click through rate has long mattered for traditional SEO.

Citation Share of Voice

Citation share of voice is a brand's proportion of all citations within a topic or query set, measured against competitors. It narrows the broader idea of share of voice specifically to citations rather than mentions.

Cited Source

A cited source is a page an AI engine references when it builds an answer, usually shown to the user as a link or an attribution line. Becoming one is a central goal of most GEO work.

Content Authority

Content authority is whether a specific page carries enough verifiable substance for a model to use it, as opposed to how well known the site it sits on happens to be. A well known brand's product page can lose out to a smaller independent blog on the same question if the blog compares options and the product page only sells.

Content Chunking

Content chunking is the practice of breaking longer content into discrete, self contained sections, each answering one question, so retrieval systems can pull out exactly the piece they need.

Content Freshness

Content freshness is how recent a piece of content is relative to right now. Engines that retrieve live pages tend to favor fresher content, especially on time sensitive questions.

Context Window

The context window is the amount of text a model can hold in mind at once, including the prompt, any retrieved documents, and its own answer. Anything beyond that limit simply falls out of view.

Conversational Commerce

Conversational commerce is shopping that happens through a chat interface, where a human buyer gets help discovering and comparing products in conversation but still makes and completes the purchase themselves.

Conversational Query

A conversational query is a search phrased the way you would ask a person, in full natural sentences instead of clipped keywords. AI engines are built to understand and answer them directly.

Core Web Vitals

Core Web Vitals are Google's user experience metrics for speed, responsiveness, and visual stability. They still matter for whether a page is eligible to appear in an AI Overview.

Crawlability

Crawlability is how easily bots and AI indexers can reach, move through, and read a site. If they cannot crawl a page, nothing else matters, since it can never be indexed or cited.

Cross Model Normalization

Cross model normalization is how a monitoring tool combines results from AI systems that do not answer the same way, and it is the reason a single blended visibility score can be misleading. Some engines cite many sources, others cite almost none, and averaging them together produces a number that no individual engine would recognize.

D terms

4 terms

Dark Traffic

Dark traffic is website visits whose source your analytics cannot identify, increasingly including AI referred visits that arrive without clear referrer data. It usually gets lumped under direct traffic, which means AI search is often doing more for a brand than its reports show.

Deep Research

Deep Research is an AI feature, offered across several assistants, that runs an extended multi step investigation of a question, browsing many sources and returning a long, cited report instead of a quick answer.

Digital PR

Digital PR is the work of earning mentions and coverage of a brand across reputable third party sites. It builds the kind of authority that feeds AI citations, whether or not the mention includes a link.

Domain Authority

Domain authority is a score, originally from Moz, that estimates how trustworthy and competitive a whole domain is, based mostly on its backlink profile and history. Higher domain authority tends to track with getting cited more often by AI engines.

E terms

10 terms

E E A T

E E A T stands for Experience, Expertise, Authoritativeness, and Trustworthiness, Google's framework for judging content quality. Its signals shape both traditional rankings and AI Overview inclusion.

Embedding

An embedding is a string of numbers that represents the meaning of a piece of text, so that things meaning similar things end up with similar numbers. It is the underlying unit that powers semantic and vector search.

Enhanced Entity Page

An enhanced entity page is a web page that is strategically structured to serve as the definitive, AI readable source for a specific entity. It combines structured data, author signals, answer units, and schema markup to maximize the chance of being retrieved and cited by generative AI engines.

Entity

An entity is a distinct thing AI systems can recognize and connect to other things, such as a company, founder, product, service, concept, or location. In the language of a knowledge graph, entities are the real world people, places, organizations, and things that everything else is described in relation to.

Entity Disambiguation

Entity disambiguation is the work of helping an AI system tell a brand apart from everything else that shares its name. It matters more often than it sounds, since plenty of brand names overlap with common words or with entirely unrelated companies.

Entity Footprint

An entity footprint is the full presence of a brand, its people, and its products across the web as recognized entities. It is the sum of how consistently and accurately a brand is described everywhere AI systems read, and a strong footprint makes AI describe it correctly and confidently.

Entity Health

Entity health measures whether AI systems represent a company, service, or piece of content accurately rather than distorting, oversimplifying, or inventing details about it.

Entity Optimization

Entity optimization is the work of making a brand, person, product, or concept a recognized, well defined entity in search knowledge graphs and model training data, so AI can identify and describe it correctly.

Entity SEO

Entity SEO is optimizing for how search and AI systems recognize and connect an entity, such as a brand, person, or product, rather than optimizing individual keywords or pages in isolation.

Extractability

Extractability is how easily an AI engine can lift a clean, self contained piece of content and use it in an answer. Two pages can hold the exact same facts, yet the more extractable one gets cited while the other gets skipped over.

F terms

5 terms

Factual Claim

A factual claim is a specific, verifiable statement within content, such as a statistic, a benchmark, or a named result, that AI systems can extract, evaluate for accuracy, and cite as evidence in a generated answer.

Factual Density

Factual density is how much specific, verifiable information a piece of content packs in, including concrete numbers, dates, named examples, and sourced claims. AI engines strongly prefer to cite it, and raising it is one of the highest leverage things a team can do for GEO.

Faithfulness

Faithfulness is a GEO metric that assesses whether an AI model accurately represents a brand's content when citing it. High faithfulness means the AI's summary or quote preserves the original meaning without distortion or hallucination.

FAQ Content Strategy

An FAQ content strategy is building pages around the specific questions people actually ask, each paired with a short, authoritative answer. The format mirrors how people query AI engines and gets lifted as a direct answer often.

G terms

8 terms

Gemini

Gemini is Google's family of large language models and the assistant built on them. It powers Google's conversational AI products and contributes to AI features across Search.

Generative Engine

A generative engine is any system that answers a question by generating new text from what it knows and retrieves, rather than returning a list of links. It is the engine in generative engine optimization.

Generative Engine Optimization (GEO)

Generative engine optimization is the practice of improving how a brand appears in AI generated answers. GEO focuses on visibility, prominence, context, and the sources used to support a brand's presence in tools like ChatGPT, Claude, Gemini, and Perplexity.

GEO Score

A GEO score is a single number that sums up how visible a brand is across AI search. It rolls up things like how often the brand gets mentioned, whether it gets cited, how it is described, and how it compares to rivals, so a team can track one figure instead of drowning in dashboards.

GEO vs SEO

SEO optimizes for ranked lists of links based on keywords and backlinks, while GEO optimizes for how information is selected, summarized, and cited by AI models. SEO targets search engine results pages, and GEO targets AI generated answers.

Google Extended

Google Extended is a control that lets a site opt out of having its content used to train and ground some of Google's AI models, set through robots.txt, without affecting normal Google Search crawling.

Google SGE (Search Generative Experience)

Google SGE was Google's earlier experimental AI search layer, generating synthesized summaries at the top of results. It has since evolved into AI Overviews and marked Google's first big move toward putting large language models inside core search.

Grounding

Grounding is the practice of anchoring an AI model's answer in real, retrieved sources rather than letting it rely only on memory. It is what allows an engine to cite where its information actually came from.

H terms

3 terms

Hallucination

A hallucination is when an AI model states something false or made up with full confidence, presenting it as fact. It happens because models are built to generate plausible sounding language, not to verify truth.

Hand off Success

Hand off success measures whether visibility inside AI answers leads a user to take a meaningful next step, such as visiting a site, requesting more information, booking a call, or becoming a lead.

I terms

3 terms

Indexability

Indexability is whether a page can actually be added to a search engine's index, which is the prerequisite for appearing in any result, classic or AI generated. Pages blocked by noindex tags, login walls, or errors cannot be cited.

Information Gain

Information gain is how much new, original value a page adds beyond what already exists on a topic. Content that mostly repeats existing material has low information gain, and as AI compresses the web into single answers, that kind of sameness gets squeezed out first.

Instant Checkout

Instant Checkout is the feature that lets people buy products directly inside ChatGPT, completing the purchase in the conversation instead of going to the merchant's site. It was built by OpenAI and Stripe on the Agentic Commerce Protocol.

J terms

1 term

JSON LD

JSON LD is a lightweight structured data format used to embed machine readable information into web pages. It helps search engines and AI models understand the meaning, relationships, and context of content, making it easier to retrieve and cite accurately.

K terms

4 terms

Knowledge Cutoff

A knowledge cutoff is the date after which a model has no built in knowledge, because its training data stops there. Anything more recent is unknown to the model unless it retrieves that information live.

Knowledge Graph

A knowledge graph is a structured map of entities and the relationships between them, used by search and AI systems to work out that a name refers to a specific real thing rather than just a word. A well connected brand tends to get described consistently, while an ambiguous one gets confused with its namesakes.

Knowledge Panel

A knowledge panel is an information box, shown by Google and increasingly by AI tools, that summarizes key facts about an entity such as a brand, person, or organization, drawn from structured data and authoritative sources.

Knowledge Update Cycle

The knowledge update cycle is how often an AI system refreshes what it knows, either by retraining or by re retrieving from the live web. It sets the floor on how quickly any change a brand makes can actually show up in answers, and it varies a great deal from one system to another.

L terms

9 terms

Large Language Model (LLM)

A large language model is an AI trained on very large amounts of text that predicts and generates language, and now often answers questions directly instead of returning links. ChatGPT, Claude, Gemini, Perplexity, and Copilot are all built on one.

LLM Advertising

LLM advertising is paid placement inside AI generated answers, where sponsored content or citations appear alongside the organic response. It is an emerging paid layer of AI search, distinct from earning citations naturally through content.

LLM Audit

An LLM audit checks how one or more models currently perceive, describe, and represent a brand, product, or topic. It surfaces the gaps, errors, and openings before a team decides what to act on.

LLM Optimization (LLMO)

LLM optimization is the work of improving how a brand or its content is understood, represented, and surfaced specifically by large language models. It is used roughly interchangeably with GEO and AEO, though it tends to put more focus on the model layer itself rather than any one particular search surface.

LLM Visibility

LLM visibility is how much, and how accurately and favorably, a brand or its content turns up in AI responses across the major engines and chatbots. It is the thing most GEO work exists to improve.

llms full.txt

llms full.txt is a companion to llms.txt. Where llms.txt is a short curated index of a site's key pages, llms full.txt contains the full text of that content in one file, so an AI can ingest it in a single fetch.

llms.txt

llms.txt is a proposed standard, a simple markdown file placed at a site's root, similar to robots.txt, that hands AI systems a clean, curated map of a site's most important content and key facts. Its real world impact is still fairly mixed today.

Log File Analysis

Log file analysis is examining a server's raw access logs to see exactly which bots visited, what they fetched, and when. It is the most direct way to confirm that AI crawlers are actually reading a site.

Long Tail AI Query

A long tail AI query is a highly specific, often multi word question aimed at a narrow intent. These queries usually face less competition and carry clearer purpose, which makes them strong targets for focused content.

M terms

6 terms

Machine Legible Content

Machine legible content is content structured and marked up so that software, including crawlers, retrieval systems, and AI models, can parse its meaning reliably, not just so a human reader can follow it.

Mentions

Mentions are the number of AI generated answers where a brand appears. Tracking mentions over time reveals whether a brand's presence is growing or declining across AI platforms.

Model Context Protocol (MCP)

The Model Context Protocol is an open standard for connecting AI assistants to external tools and data sources in a consistent way, so an agent can take real actions instead of just talking.

Multi Model Monitoring

Multi model monitoring is tracking a brand across several AI systems at once rather than checking just one and generalizing from it. It matters because the engines genuinely disagree, and because for many brands there is no ranking page to inspect, only the answers themselves.

Multi Turn Conversation Modeling

Multi turn conversation modeling tests a brand's AI visibility across a real back and forth exchange rather than a single question, because that is how people actually use AI assistants. They ask, refine, object, and follow up before deciding.

N terms

1 term

Named Entity Recognition (NER)

Named entity recognition is the AI technique of identifying and classifying the real world things named in text, such as people, companies, products, and places, so a system understands what a passage is actually about.

O terms

1 term

Original Research

Original research is data or analysis a brand produces itself, such as a survey, benchmark, study, or proprietary dataset, that does not exist anywhere else. It is one of the most reliable ways to earn AI citations.

P terms

10 terms

Passage Ranking

Passage ranking is a search engine's ability to index and rank individual passages within a long page, instead of judging the whole page as one block. A highly relevant paragraph can surface even when the page overall is not a strong match.

Perplexity AI

Perplexity is an AI native answer engine that retrieves live web results and turns them into cited, conversational responses. Its heavy reliance on citations makes it one of the most important platforms to optimize for.

Persona Based Querying

Persona based querying is testing a brand's AI visibility by asking questions the way specific customer types would actually ask them, rather than relying on one generic prompt. It is the difference between a flattering demo and a real read on where a brand stands.

Pillar Page

A pillar page is the central, comprehensive page on a broad topic that links out to and ties together the more specific pages around it. It anchors an entire topic cluster.

Product Feed Optimization

Product feed optimization is the work of structuring and enriching product data, including titles, descriptions, attributes, price, and stock, so AI shopping agents and answer engines can read it, trust it, and recommend the products. In AI commerce, the feed effectively becomes the storefront.

Prompt Engineering (for GEO)

Prompt engineering for GEO is choosing which questions to measure a brand's visibility against, written in the words a real customer would actually type. It is a different job from the prompt engineering used to get better output from a model, since here the prompts are the measuring instrument and need to stay fixed once a test begins.

Prompt Intelligence

Prompt intelligence is the analysis of the actual questions people ask AI systems within a category, including their phrasing, intent, and frequency. It is used to decide which prompts are worth tracking and targeting in the first place.

Prompt to Citation Funnel

The prompt to citation funnel is the path from a user's question, through the engine's answer, to the sources it cites, ending in a click or a brand impression. GEO work is largely about slotting a brand into that path as a cited source.

Prompt Tracking

Prompt tracking is monitoring how a brand appears across a defined set of prompts over time, watching whether it gets mentioned, how it is described, and how that changes. It turns a one off snapshot into a trend line a team can actually manage.

Prompts

Prompts are the natural, consumer style questions used to query AI models, similar to what people actually ask tools like ChatGPT. Prompts can be custom written or based on common high intent use cases within a category.

Q terms

1 term

Query Fan Out

Query fan out is when an AI engine takes a single question, quietly splits it into several related searches, runs them all at once, and blends the results into one answer. Google's AI Mode, ChatGPT, and Perplexity all do this, and it changes how brands need to think about visibility.

R terms

6 terms

Rank

Rank is where a brand appears when it is mentioned in an AI generated answer, such as first, second, or third. A higher rank generally means the AI model considers the brand more relevant to the query.

Reasoning Model

A reasoning model is a large language model built to work through problems in steps before answering, trading speed for more careful, accurate responses on harder questions.

Reranking

Reranking is a second pass in retrieval where an initial set of candidate results is reordered by a more careful model, putting the most relevant ones on top before they ever reach the final answer.

Research Intent Queries

Research intent queries are searches where the user's goal is to learn, compare, or evaluate rather than to navigate directly to a specific site. These queries are increasingly answered by AI systems, which makes them a key target for GEO strategy.

Retrieval Augmented Generation (RAG)

Retrieval augmented generation is an AI architecture where the model retrieves relevant documents before generating a response, rather than relying solely on training data. Tools like Perplexity use it to give more current and source backed answers, and most modern AI search runs on some version of it.

robots.txt

robots.txt is a long standing file at a site's root that tells crawlers which parts of the site they may and may not access. In the AI era it has also become how a site allows or blocks AI crawlers specifically.

S terms

13 terms

Schema Markup and Structured Data

Schema markup, or structured data, is machine readable code, usually written as JSON LD, that explicitly labels what a piece of content is, such as an FAQ, an article, a product, or an organization. It helps both traditional search engines and AI models interpret a page correctly, though its direct effect on AI citations is still not firmly established.

Search Engine Optimization (SEO)

Search engine optimization is the long running practice of improving a site so it ranks higher in traditional search results. In the AI era it is widening into a broader discipline that also covers visibility inside AI generated answers.

Search Engine Results Page (SERP)

A SERP is the page a search engine returns for a query. Modern ones increasingly mix AI answer boxes, featured snippets, and knowledge panels in with the traditional links, making the space more crowded and varied than it used to be.

Semantic Authority

Semantic authority is how much an AI system treats a source as reliable on a subject, expressed through whether it retrieves and cites that source. It is not a score any platform publishes, and no one outside the labs can measure it directly, so it is best understood as an effect visible in citation behavior rather than a number you can look up.

Sentiment

Sentiment is how positively or negatively a brand is described in AI generated answers. Sentiment analysis helps identify whether AI models associate a brand with favorable or unfavorable attributes.

Server Side Rendering (SSR)

Server side rendering means the server delivers a fully built HTML page, rather than sending bare code that the browser fills in with JavaScript afterward. It makes content reliably visible to crawlers.

Share of Model

Share of model is the share of AI systems that name a brand when asked a relevant question. Coverage is not uniform, since a company can be recommended by ChatGPT and be entirely absent from Claude, and the pattern often differs again by country.

Share of Voice

Share of voice is a brand's share of AI mentions compared to competitors. It shows how much of the conversation a brand owns within a specific topic or category in AI generated answers.

Source Credibility Signals

Source credibility signals are the mix of factors, including domain authority, E E A T indicators, backlink quality, author credentials, structured data, and citation history, that lead an engine to judge a source trustworthy enough to cite.

Sources

Sources are the domains and URLs that AI models cite when generating answers about a brand. Tracking them helps a team understand which content is actually shaping how AI represents that brand.

Synthetic Prompt Testing

Synthetic prompt testing is running a large, generated set of test questions that stand in for the real prompts customers would use, to build a statistical picture of how often and how well a brand appears across AI answers.

T terms

3 terms

Tokenization

Tokenization is how a model breaks text into small pieces called tokens, the units it actually reads and generates. A token is often a word fragment rather than a whole word.

Topic Cluster

A topic cluster is a group of interlinked pages that cover a subject comprehensively, usually a central pillar page plus many supporting pages that link to it and to each other.

Topical Authority

Topical authority is the depth and breadth of expertise a site shows on a subject, judged by the quantity, quality, and interconnection of its content. Strong topical authority makes a site far more likely to be cited as a primary source.

U terms

1 term

Universal Commerce Protocol (UCP)

The Universal Commerce Protocol is Google's open standard for agentic commerce, letting merchants sell through Google's AI surfaces such as AI Mode and the Gemini app. It is the main alternative to OpenAI's Agentic Commerce Protocol.

V terms

3 terms

Visibility

Visibility is how often a brand appears across tracked prompts in AI generated answers. It is the foundational metric in GEO, since if a brand does not appear at all, nothing else about how it is described really matters yet.

Visibility Score

A visibility score is a composite metric that rolls up a brand's overall presence across AI answers, factoring in things like citation frequency, share of voice, sentiment, and the authority of the responses it shows up in.

Z terms

1 term
Back to top