Glossary

Plain-English definitions of the key terms in AI Authority Infrastructure, Generative Engine Optimisation, Answer Engine Optimisation, agentic commerce, and how AI agents find, evaluate and transact with businesses. Written for marketing leaders and decision-makers, not technologists.

Last updated: 10 May 2026
ABCDEFGHIJKLMOPRSTUW
A
A2A (Agent-to-Agent)
An open protocol for communication between AI agents from different vendors and systems. One of the foundational standards being stewarded by the Linux Foundation's Agentic AI Foundation, alongside MCP, AP2, ACP and UCP. Lets a buyer-side agent talk to a supplier-side agent without proprietary integration.
ACP (Agentic Commerce Protocol)
An open standard for AI agents to conduct commercial transactions on behalf of buyers, including identity verification, authorisation and payment. Sits in the same family as MCP, A2A, AP2 and UCP. Part of the technical infrastructure that makes transactable AI Authority possible.
Agent-to-agent commerce
A model of commercial transaction in which AI agents acting on behalf of buyers communicate directly with the systems of suppliers, including requesting quotes, comparing options and placing orders. Forecast by Gartner to account for $15 trillion of B2B spend by 2028, with 90% of B2B buying AI-agent intermediated.
Agentic AI
AI systems that go beyond answering questions to take action on a user's behalf: browsing, filling forms, transacting, comparing suppliers and managing tasks across systems. The shift from generative AI (which produces content) to agentic AI (which produces outcomes) is what makes the third dimension of AI Authority, transactability, commercially urgent.
Agentic commerce
The category of commerce in which AI agents make and execute purchasing decisions on behalf of human buyers, working from policies, budgets and preferences rather than direct instruction at point of purchase. Bain forecasts 15 to 25 percent of retail spend will move through agentic channels in the medium term. The shift demands that businesses are not only discoverable and trusted, but transactable.
AEO
See Answer Engine Optimisation.
AI agent
A software system that uses generative AI to perform tasks autonomously on behalf of a user. AI agents can browse, summarise, evaluate options, draft responses and increasingly transact. The buyer-side AI agent is the audience whose decisions AI Authority work is designed to influence.
AI Authority
The degree to which AI platforms recognise and trust an organisation as a credible source on the topics relevant to its business. An organisation with strong AI Authority is cited, referenced and recommended in AI-generated answers. Without it, the organisation is absent regardless of how well-established it is in the real world.
AI Authority Audit
A structured, documented investigation into how an organisation currently appears across all major AI platforms. Covers citation frequency, sentiment, entity definition, technical accessibility and competitive positioning. Delivered as a written report with prioritised roadmap. From £1,500, 1 to 2 weeks delivery.
AI Authority Infrastructure (AIAI)
KnownEntity.ai's category descriptor and our acronym ambition. AIAI sits alongside SEO, AEO and GEO as a recognised practice category. Where AEO and GEO describe the discipline of optimising for AI answers, AI Authority Infrastructure describes the underlying systems and assets, hosted entity records, citation graphs, agent endpoints, that make a business discoverable, trusted and transactable by AI agents. The full positioning line is "AI Authority Infrastructure - Built for Business."
AI Authority Ladder™
KnownEntity.ai's proprietary four-stage methodology for moving an organisation from invisible to appropriately and strategically cited in AI answers. The four stages are Diagnostic Audit, Entity Build (EntityCore™), Citation Strategy, and Tracking and Measurement. Each stage builds on the one before. The Ladder maps onto the three dimensions of AI Authority: discoverable, trusted and transactable.
AI Authority Snapshot
KnownEntity.ai's complimentary, top-level analysis of how an organisation currently appears across the major AI platforms. Turned around within 48 hours of request. Designed to give a sense of the opportunity before any paid engagement.
AI crawler
A software agent operated by an AI platform to read and index public web content for training data, retrieval or live grounding. The major crawlers include GPTBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot (Perplexity), Google-Extended and Bingbot. A robots.txt file that blocks these crawlers prevents the business from being indexed by the AI platforms its buyers use.
AI Overview
Google's name for the AI-generated summary that appears at the top of search results for many queries. Draws on a curated set of sources rather than the full ranked list, making AI Authority within those sources commercially significant.
Answer Engine Optimisation
The practice of structuring an organisation's expertise, content and online presence so that AI platforms recognise it as a credible, citable source and feature it in synthesised answers. Sometimes abbreviated AEO. Used interchangeably with Generative Engine Optimisation (GEO).
AP2 (Agent Payment Protocol)
An emerging open standard for AI agents to authorise and execute payments on behalf of users, with attribution and verification baked in. One of the agent commerce protocols being developed alongside MCP, A2A, ACP and UCP. Critical infrastructure for the transactable layer of AI Authority.
B
Buyer journey
The sequence of stages a B2B buyer moves through from first awareness of a problem to final supplier selection. Research from 6sense and Gartner shows that the majority of this journey now happens before a supplier is contacted, with 73% of B2B buyers using AI tools in their purchase research and 95% of winning vendors already on the buyer's Day One shortlist.
Buyer-side agent
An AI agent operating on behalf of a buyer, conducting research, comparing options, evaluating credentials and increasingly initiating transactions. The audience that AI Authority work is designed to reach. Distinct from supplier-side agents, which operate on behalf of the business being recommended.
C
ChatGPT
OpenAI's consumer AI assistant, the most widely used generative AI platform. One of the major AI platforms KnownEntity.ai monitors and optimises for. Draws a substantial share of citations from Wikipedia, established trade press and authoritative sources, making the citation graph particularly important for ChatGPT visibility.
Citation
In the context of AI search, a reference made by an AI platform to a specific business, product or source within a generated answer. Can be implicit (mentioning the business by name) or explicit (linking to the source URL). Citations are the primary outcome AI Authority work is designed to influence.
Citation frequency
A measurement of how often an organisation is cited or mentioned across a defined set of AI prompts. Tracked over time as a key indicator of AI Authority growth or decline.
Citation graph
The network of third-party sources that mention or cite an organisation, used by AI systems to assess credibility and decide whether to feature the organisation in answers. Includes trade press, expert commentary, review platforms, industry bodies, LinkedIn authority content and Wikipedia.
Citation Strategy
The third stage of the AI Authority Ladder™. The systematic development of an organisation's external citation graph through trade press coverage, expert bylines, review platforms, LinkedIn authority and Wikipedia where evidence supports it. Every placement is logged and tracked as a potential AI citation source.
Claude
A family of large language models built by Anthropic. One of the major AI platforms KnownEntity.ai monitors and optimises for, alongside ChatGPT, Perplexity, Gemini and Copilot.
ClaudeBot
The web crawler operated by Anthropic to gather content for Claude. Businesses that block ClaudeBot in their robots.txt file will not appear in Claude's answers, regardless of how strong their content or citation graph is. See AI crawler.
Copilot
Microsoft's AI assistant, integrated across Microsoft 365 and Bing. One of the major AI platforms KnownEntity.ai monitors and optimises for.
D
Diagnostic Audit
The first stage of the AI Authority Ladder™. A systematic prompt audit across all major AI platforms (ChatGPT, Perplexity, Claude, Gemini, Copilot) to establish citation baseline, share of voice, sentiment and entity definition. Produces the documented starting point against which all subsequent work is measured.
Discoverable
The first of the three dimensions of AI Authority. A business is discoverable when AI agents find it during the buyer research stage, in the synthesised answers buyers see when they ask category-relevant questions. Discoverability depends on technical access (AI crawlers can reach the content), content quality (the content is extractable and citable), and citation graph (third-party sources point to the business). The floor of AI Authority work and the entry condition for everything that follows.
E
Earned media
Coverage, commentary and citation that an organisation receives from third-party sources, including trade publications, industry bodies, journalists, review platforms, expert directories and podcasts. Called earned because it is not paid for and not self-published. Fundamental to AI Authority because AI systems trust content that third parties also trust.
E-E-A-T
Experience, Expertise, Authoritativeness and Trustworthiness. The framework Google publishes for evaluating content quality, increasingly used as shorthand for the signals AI platforms also consider when deciding which sources to cite.
Entity
In the language of AI and knowledge systems, a person, organisation, product or concept that an AI platform has structured information about and can reference with confidence. The opposite of an entity is an unknown or ambiguous reference that AI cannot place.
Entity Build
The second stage of the AI Authority Ladder™. The construction of EntityCore™, KnownEntity.ai's proprietary structured, machine-readable source of truth that AI agents consult when deciding whether to recommend a business.
Entity custodian
The role KnownEntity.ai plays for its clients. Like Experian maintains the structured record that determines whether a business is trusted in financial decisions, an entity custodian maintains the structured, machine-readable source of truth AI agents consult when deciding whether to recommend the business. The custodian is paid for the record to be accurate, complete and trusted, not for one-off interventions.
Entity definition
The structured understanding an AI platform holds about what an organisation is, what it does, who it serves and what it stands for. The Diagnostic Audit assesses the current entity definition; the Entity Build refines and asserts it.
EntityCore™
KnownEntity.ai's proprietary, structured, machine-readable source of truth for a client. The content repository AI agents consult when deciding whether to recommend the business. Hosted at a subdomain of the client's existing website and served via a dedicated MCP endpoint, with provenance and verification baked in. Built specifically for AI extraction with question-first structure, complete answers and deep topical coverage. Costs £6,000 to £12,000 to build, delivered in 4 to 6 weeks.
F
Forrester forecast
Forrester predicts that by the end of 2026, AI agents will be a meaningful share of B2B research and shortlisting activity, marking the transition point at which AI Authority becomes commercially material rather than aspirational. One of the analyst data points KnownEntity.ai cites when sizing the urgency of the work for clients.
G
Gartner forecast
Gartner predicts that AI agents will mediate $15 trillion of B2B spend by 2028, with up to 90 percent of B2B buying becoming AI-agent intermediated. The headline figure KnownEntity.ai cites to anchor the commercial urgency of agent-readable infrastructure.
Gemini
Google's family of large language models, used in Google's AI Overviews and standalone Gemini products. One of the major AI platforms KnownEntity.ai monitors and optimises for.
Generative AI
A class of AI that produces new content (text, images, code, audio) rather than just classifying or retrieving existing content. The basis of all the major AI platforms relevant to AI Authority work. Distinct from agentic AI, which uses generative AI to take actions on a user's behalf.
Generative Engine Optimisation
The practice of structuring and positioning an organisation's expertise, content and online presence so that AI platforms recognise it as a credible, citable source. Where SEO aims to rank a website highly in search results, GEO aims to ensure a business is present in the synthesised answer that an AI platform delivers directly. Often abbreviated GEO. Used interchangeably with AEO.
GEO
See Generative Engine Optimisation.
GPTBot
The web crawler operated by OpenAI to gather content for ChatGPT and other OpenAI products. Businesses that block GPTBot in their robots.txt file will not be reachable by ChatGPT's indexing process. KnownEntity.ai checks for AI crawler access as part of every Diagnostic Audit. See AI crawler.
H
Hallucination
An AI-generated answer that is fluent, plausible and confidently delivered but factually wrong. AI systems hallucinate when they have insufficient or contradictory source material on a topic. One of the strongest commercial cases for AI Authority work: a business with a thin or fragmented entity definition is materially more likely to be hallucinated about, and the hallucination is then quoted to its prospective buyers.
I
Invisible Shortlist
The KnownEntity.ai white paper on the B2B buyer journey and AI's role in pre-vendor research. Sets out why 95 percent of winning vendors are already on the buyer's Day One shortlist, and what that means for businesses that are absent from AI-generated answers. Available at knownentity.ai/the-invisible-shortlist/.
J
JSON-LD
JavaScript Object Notation for Linked Data. The most widely supported format for embedding structured data (schema markup) in web pages. Lets a page state explicit, machine-readable facts about an entity, its attributes and its relationships, alongside the human-readable content. The standard format used by Google, AI platforms and the wider web for entity-level structured data.
K
Knowledge graph
A structured representation of entities and the relationships between them. Used by AI systems to ground generated answers in factual structure rather than free-form prediction. EntityCore™ functions as a knowledge graph for the businesses KnownEntity.ai works with.
Knowledge Graph Build
See EntityCore™. The service in which KnownEntity.ai constructs a client's EntityCore™ knowledge graph.
L
Large language model
A type of AI model trained on large volumes of text to predict and generate language. The underlying technology behind ChatGPT, Claude, Gemini, Copilot and other generative AI products. Often abbreviated LLM.
LLM
See Large language model.
llms.txt
A proposed standard file, served at the root of a website, that summarises the site's content and structure in a way optimised for large language models to consume. The AI-era counterpart to robots.txt, but for guidance rather than access control. KnownEntity.ai publishes its own llms.txt at info.knownentity.ai/llms.txt as part of the Open Entity Registry.
M
McKinsey six-level automation curve
McKinsey's framework for how AI delegation evolves in commerce, from Level 0 (programmed convenience like subscriptions) through Level 5 (networked autonomy, agent-to-agent commerce running continuously). The crucial insight is that delegation does not move uniformly: it accelerates in low-regret, high-repetition categories like consumables and plateaus in identity-driven categories like luxury. For B2B, the inflection point is supervised execution, where agents are authorised to act within defined policies. The curve is buyer-facing; suppliers need to read it backwards to understand what discoverability, trust and transactability are required at each level.
MCP (Model Context Protocol)
An open standard, originated by Anthropic, that lets AI agents securely access external context, tools and data sources. KnownEntity.ai serves each client's EntityCore™ via a dedicated MCP endpoint, giving authorised AI agents direct access to the canonical entity record with provenance and verification baked in. One of the foundational protocols of agent infrastructure, alongside A2A, AP2, ACP and UCP.
Measurement (in AI Authority)
The fourth stage of the AI Authority Ladder™. Monthly prompt audits using the same structured query set across all major AI platforms, tracking citation frequency, sentiment and competitor share of voice over time. Closes the loop on AI Authority work and provides evidence of progress.
O
Open Entity Registry
KnownEntity.ai's machine-readable entity record, published as structured data so AI assistants and search engines can verify who we are, what we do, and how we deliver. Includes registry.json and llms.txt files alongside human-readable pages covering services, methodology, sectors, credentials, people and case studies. A practical demonstration of the same agent-readable infrastructure we build for clients.
P
Perplexity
An AI-native search engine that synthesises answers from live web sources rather than relying primarily on pre-trained knowledge. One of the major AI platforms KnownEntity.ai monitors and optimises for. Distinctive in its emphasis on conversational, experience-based content.
PerplexityBot
The web crawler operated by Perplexity to gather content for live synthesis. See AI crawler.
Prompt audit
A structured exercise in which a fixed set of buyer-relevant prompts is run across the major AI platforms and the responses are analysed for citation frequency, sentiment, accuracy and share of voice. The core measurement instrument of AI Authority work, run once at baseline and then monthly thereafter.
Properly cited
The outcome KnownEntity.ai works towards. Not just being mentioned, but being cited for the right things, in the right context, when it matters commercially. The opposite of generic visibility, and the precise reason most AI Authority work delivers commercial value rather than vanity metrics. See Properly known.
Properly Known
KnownEntity.ai's brand essence. Captures the distinction between merely appearing in AI answers and being cited for the right things, in the right context, when it matters commercially. The opposite of generic visibility.
R
Research Reflex
The KnownEntity.ai companion white paper to The Invisible Shortlist. Examines the buyer behaviour shift from active search to AI-mediated research, and how that reshapes which businesses end up on the consideration set.
Retainer
See Authority Retainer. KnownEntity.ai's monthly service combining ongoing prompt audits, EntityCore™ expansion, citation outreach and reporting. Priced at £1,000 to £3,000 per month on a rolling three-month minimum.
registry.json
A structured JSON file published at the root of an entity's subdomain that declares its canonical attributes in a machine-readable format. KnownEntity.ai publishes its own registry.json at info.knownentity.ai/registry.json as part of the Open Entity Registry, demonstrating the agent-readable infrastructure pattern we build for clients.
Robots.txt
A plain-text file at the root of a website that tells web crawlers which parts of the site they may access. The first place to check in any AI Authority diagnostic: a robots.txt file that blocks GPTBot, ClaudeBot, PerplexityBot or Google-Extended renders the business invisible to those AI platforms regardless of how strong its content or citation graph is.
S
Schema markup
Structured data added to web pages, typically in JSON-LD format, that describes the content in a machine-readable way. Helps search engines and AI platforms understand what a page is about and which facts it asserts. A foundational technical practice for AI discoverability.
Sentiment analysis
Assessment of the tone and framing with which an organisation is described in AI answers. Positive, neutral or negative. Tracked over time as part of the monthly Authority Report; off-strategy or negative sentiment is itself a signal that AI Authority work is needed.
SEO
Search Engine Optimisation. The discipline of structuring a website to rank well in traditional search results. SEO and AI Authority Infrastructure are complementary but distinct: SEO optimises for the ten blue links, AIAI optimises for the synthesised answer above them and the agent that may transact below them. SEO agencies often add GEO or AEO as a service line; KnownEntity.ai is a specialist in the entity layer that sits underneath.
Share of voice
The proportion of relevant AI prompt responses in a given category that mention or cite a particular organisation, relative to its competitors. The most useful single number for understanding competitive position in AI search.
Source attribution
The practice of naming the original source of a fact, statistic or claim within content. Important both for human readers and for AI systems, which use attribution as a trust signal when deciding whether to extract and cite a passage.
Structured data
See Schema markup. Information presented in a machine-readable format with explicit relationships, as opposed to unstructured prose.
Supplier-side agent
An AI agent or endpoint operated on behalf of a supplier business, exposing its capabilities, attributes and offers to buyer-side agents in a structured, machine-readable form. The transactable layer of AI Authority Infrastructure. EntityCore™ serves as the foundational supplier-side surface, with future MCP, A2A and AP2 capability added as agentic commerce matures.
T
Topic cluster
A group of related pages or articles centred on a specific topic, with a hub page and supporting detail pages. The architectural unit of an EntityCore™ build, with a typical engagement covering 20 to 40 clusters mapped to the questions a client's buyers actually ask.
Tracking and Measurement
The fourth stage of the AI Authority Ladder™. See Measurement (in AI Authority).
Transactable
The third dimension of AI Authority. A business is transactable when AI buyer-agents can take action on its behalf: request a quote, place an order, initiate a contract, schedule a service. Transactability requires structured commercial data, agent endpoints (MCP, A2A, AP2) and identity and authorisation infrastructure. Today's clients buy discoverable and trusted; transactable is the roadmap and the competitive moat as agentic commerce matures.
Trusted
The second dimension of AI Authority. A business is trusted when AI platforms represent it accurately and authoritatively, supported by a strong external citation graph and an unambiguous entity definition. Trust is what moves a business from appearing in answers to being recommended in answers. Built through Citation Strategy and the consistency of an EntityCore™ record.
U
Universal Commerce Protocol
Google's open standard for agent-to-agent commerce, launched at the National Retail Federation event in January 2026. Establishes a framework for AI agents to communicate with retailer and supplier systems for product discovery, evaluation and transaction. Marker of how close agentic commerce is.
W
Wikidata
The structured-data sister project to Wikipedia, holding facts about entities in a machine-readable format. Used by AI systems as a high-trust source for verifying basic facts about people, organisations and things. Present in Wikidata is itself an AI Authority signal where the evidence supports it.
Wikipedia
Among the highest-trust sources used by AI platforms when generating answers. ChatGPT in particular draws a substantial share of its citations from Wikipedia. Inclusion in Wikipedia, where the evidence trail genuinely supports a notable entry, is one of the most valuable citation outcomes.

A term we've missed?

If there's a term you've seen in AI Authority or AI search that we haven't covered, email hello@knownentity.ai and we'll add it. We update this page as the field evolves.