AI Visibility Definition
The Technical Standard for Website-to-AI Communication
This document establishes the canonical definitions for AI Visibility, AI Visibility Checking, AI Discovery Files, and related terminology. It provides a precise technical framework for understanding how websites communicate their identity, scope, and permissions to AI systems including large language models, AI search engines, and retrieval-augmented generation systems.
§1 AI Visibility
AI Visibility
AV-001AI Visibility is the degree to which a website or digital entity can be discovered, correctly interpreted, accurately represented, and safely cited by AI systems including large language models, AI search engines, and retrieval-augmented generation systems.
AI Visibility is established through AI discovery files, structured data, consistent machine-readable identity signals, and clear technical infrastructure that enables AI systems to understand and trust a website's content.
AI Visibility is distinct from traditional search engine visibility. A website may rank well in conventional search results whilst having poor AI visibility, or vice versa. The mechanisms, signals, and evaluation criteria differ substantially between the two domains.
§2 AI Visibility Checking
AI Visibility Checking
AV-002AI Visibility Checking is the process of technically verifying whether a website's infrastructure — including AI discovery files, structured data, and identity consistency — enables AI systems to discover, interpret, trust, and safely use that website as an information source.
AI Visibility Checking validates capability, not outcomes. It answers the question: "Can AI systems technically understand and trust this website?" rather than "Do AI systems currently mention this website?"
AI Visibility Checking is deterministic and verifiable. Results can be reproduced, inspected, and validated independently. This distinguishes it from AI Retrieval Testing, which produces variable results dependent on prompt phrasing, model version, and sampling parameters.
What AI Visibility Checking Validates
An AI Visibility Check typically examines whether a website provides clear signals that enable AI systems to:
Discover the site and its content. Identify the business or entity accurately. Understand the scope of services, products, or information offered. Recognise what is explicitly excluded or out of scope. Trust the information as authoritative and current. Use the site safely as a citation or recommendation source.
What AI Visibility Checking Is Not
AI Visibility Checking is not brand mention tracking, prompt-based testing, ranking simulation, sentiment analysis, search visibility metrics, AI traffic analytics, or historical monitoring. Those activities fall under AI Visibility Tracking, AI Visibility Monitoring, or AI Retrieval Testing. They measure outcomes. AI Visibility Checking validates capability.
§3 AI Discovery Files
AI Discovery Files
AV-003AI Discovery Files are machine-readable files published on a website specifically to communicate information to AI systems. These files enable websites to declare their identity, scope, services, permissions, and preferred representation to AI crawlers and large language models.
Core AI Discovery Files include:
- llms.txt
- AI-readable business identity and context
- llm.txt
- Compatibility variant (should redirect to llms.txt)
- llms.html
- Human-readable reference version of AI identity information
- ai.txt
- AI usage permissions and intent signals
- ai.json
- Machine-parseable AI interaction guidance in JSON format
- brand.txt
- Brand naming, terminology, and representation rules
- faq-ai.txt
- Factual question and answer source formatted for AI consumption
- developer-ai.txt
- Technical, platform, and integration context
- robots-ai.txt
- AI crawler-specific access directives
- identity.json
- Structured canonical identity data in JSON format
Additional supporting discovery files include:
- robots.txt
- General crawler directives, including AI-specific user agents
- sitemap.xml
- Site structure used by AI crawlers for content discovery
- security.txt
- Security contact information, increasingly referenced by AI systems
- humans.txt
- Human-readable site information that AI systems may parse
- Structured data markup
- Schema.org and similar embedded metadata
§5 Terminology Taxonomy
The relationship between these terms forms a clear hierarchy. AI Visibility is the goal; AI Discovery Files are the mechanism; and the various activities represent different approaches to validation, measurement, and observation.
This taxonomy positions AI Visibility Checking as foundational. Infrastructure must be validated before outcome measurement becomes meaningful. A website cannot reliably improve its representation in AI systems without first ensuring those systems can technically discover, interpret, and trust it.
§6 Reference Implementation
365i AI Website Visibility Checker
The reference implementation of AI Visibility Checking is the 365i AI Website Visibility Checker. This tool performs a read-only technical audit of a website's AI-facing infrastructure, validating the presence, accessibility, and consistency of AI discovery files without querying AI models or simulating prompt responses.
AI Visibility Checker§7 Version History
Phase 6 standardisation release. Added /specifications/roadmap/ (theme-pegged forward plan with Active/Next/Future/On hold status flags), /specifications/extensions/ (rules for experimental x- prefixed files and the promotion path), and /specifications/i18n-a11y/ (multi-language publication, locale-tagged identity fields, RTL handling, accessibility of llms.html). Added the Discovery: directive to the robots-ai.txt specification (publishers MAY advertise AI Discovery Files on the same host). Added a formal media-type stance to the HTTP behaviour page (existing IANA types, no bespoke registrations). Expanded the file integrity and signing section on the security and privacy page with four candidate mechanisms, cross-cutting concerns, and interim publisher / consumer guidance. The Discovery: directive is the only normative addition to publisher behaviour; all other additions are forward-looking documentation.
Phase 5 standardisation release. Added /specifications/related-standards/ (positioning vs llmstxt.org, IETF AI Preferences, robots.txt, Schema.org, BCP 14, JSON Schema 2020-12, SemVer) and /specifications/implementations/ (public record of conformant implementations, IETF-style). Added an explicit llmstxt.org backward-compatibility statement to the llms.txt specification. Added a formal multi-domain and subdomain scoping rule to both the llms.txt and identity.json specifications (host-scoped files, cross-host identity asserted via sameAs). No normative requirements changed for existing publishers; the new scoping rules formalise behaviour the specification already implied.
Phase 4 standardisation release. Added /specifications/processing-model/ (seven-stage algorithm for conformant consumers), /specifications/consumer-guidance/ (what AI systems should do with AI Discovery Files), /specifications/test-vectors/ (canonical test suite framing), and reference-implementation framing on the AI Visibility Checker. No normative requirements changed.
Phase 3 standardisation release. Added /specifications/versioning/ (Semantic Versioning 2.0.0 commitments, deprecation timeline, lifecycle), /specifications/governance/ (proposal lifecycle, editorial process, working principles), /specifications/security-privacy/ (trust model, content-injection patterns, GDPR considerations, integrity primitives roadmap), and /specifications/http-behaviour/ (status codes, redirects, soft-404 detection, caching, rate limits). No normative requirements changed.
Phase 2 standardisation release. Added formal conformance specification (Essential / Recommended / Complete classes). Published machine-readable registry at /specifications/registry.json, spec meta-schema, and validator-output schema. Introduced versioned JSON Schema URLs (/v1/) alongside unversioned 'latest' aliases. Added optional BCP 47 language declaration field across all applicable AI Discovery Files. No normative requirements changed.
Phase 1 standardisation release. Added 'Status of This Document' block (Stable). Normalised normative requirement keywords to uppercase per RFC 2119 and RFC 8174. Added References section linking to /specifications/conventions/ and /licensing/. No normative requirements changed.
Data-files-only sync release. Removed incorrect $schema declarations from machine-readable definition and specification data files (these are data documents, not JSON Schemas). Synced 8 spec data files (JSON and YAML) with their published page versions. Refreshed dateModified. Reordered the coreFiles list so the canonical llms.txt precedes its llm.txt compatibility alias, and clarified the same point in the FAQ.
Added AI Visibility Directory registration guidance across all specification and content pages. Dynamic last-modified dates.
Added Implementation Guides: Quick Start, Interoperability, and Validation Framework.
Converted site to PHP-based architecture with external CSS and dynamic schema generation.
Added AI Discovery File Specifications section (§11) with comprehensive format specifications for all 10 core AI Discovery Files.
Added ai.json to core AI Discovery Files.
Removed Wikidata links (entities deleted for insufficient notability).
Added Frequently Asked Questions section (§10).
Added visible Wikidata links to Machine-Readable Formats section.
Added ETSI Technical Specification draft documents.
Initial publication.
§8 Citation & Licensing
Creative Commons Attribution 4.0 International (CC BY 4.0)
This definition document is licensed under CC BY 4.0. You are free to share and adapt this material for any purpose, including commercial use, provided appropriate attribution is given.
Suggested Citation
365i. (2026). AI Visibility Definition (Version 1.9.0). https://www.ai-visibility.org.uk/definition/
§9 Machine-Readable Formats
These definitions are available in structured, machine-readable formats for integration into automated systems, documentation, and tooling.
These definitions are structured to align with ETSI Technical Specification formatting conventions. The formal specification on GitHub contains the complete normative document, JSON Schemas, a command-line validator, and test vectors.
§10 Frequently Asked Questions
The following questions address common points of clarification regarding AI Visibility and its related concepts.
What is the difference between AI Visibility Checking and AI Visibility Tracking?
AI Visibility Checking validates whether a website's technical infrastructure enables AI systems to discover, interpret, and trust it. It examines inputs: the presence and consistency of AI discovery files, structured data, and machine-readable identity signals.
AI Visibility Tracking measures outcomes: how often a website is mentioned in AI responses, the sentiment of those mentions, and changes over time. Checking validates capability; tracking observes results.
Does AI Visibility Checking query AI models like ChatGPT or Claude?
No. AI Visibility Checking is a deterministic, read-only technical audit. It examines publicly accessible files and infrastructure without querying AI models or simulating prompts. This distinguishes it from AI Retrieval Testing, which does query AI systems and produces inherently variable results.
What are AI Discovery Files and which files are included?
AI Discovery Files are machine-readable files that communicate a website's identity, scope, and permissions to AI systems. Core files include llms.txt and llm.txt for AI-readable business context, ai.txt for usage permissions, ai.json for machine-parseable AI interaction guidance, brand.txt for naming conventions, and identity.json for structured identity data.
Supporting files such as robots.txt, sitemap.xml, and Schema.org structured data also contribute to AI discoverability. See how AI Discovery Files compare to existing web standards.
Is AI Visibility the same as AI SEO or generative engine optimisation?
No. AI Visibility refers to whether a website can be technically discovered and correctly interpreted by AI systems. It is infrastructure-focused and deterministic. Terms like "AI SEO" or "generative engine optimisation" typically describe strategies for influencing AI-generated responses — an outcome-focused activity closer to AI Visibility Tracking.
AI Visibility establishes the technical foundation that makes any subsequent optimisation meaningful.
Why might a website rank well in search but have poor AI Visibility?
Traditional search engines and AI systems evaluate websites differently. Search engines primarily assess relevance, authority, and user engagement signals. AI systems require clear, unambiguous identity signals, explicit scope declarations, and machine-readable context that enables safe citation.
A website optimised for search engine ranking may lack AI discovery files, present inconsistent identity information, or fail to declare its scope in ways AI systems can parse.
Are AI Visibility Checking results deterministic and reproducible?
Yes. AI Visibility Checking produces deterministic, verifiable results. The same website analysed with the same methodology will yield the same findings. Results can be independently reproduced, inspected, and validated.
This contrasts with AI Retrieval Testing, where results vary based on prompt phrasing, model version, temperature settings, and other factors outside the tester's control.
§11 AI Discovery File Specifications
Comprehensive format specifications are available for each core AI Discovery File. These specifications define the canonical structure, required and optional fields, validation rules, and provide annotated examples using a consistent fictional business (Horizon Strategic Consulting).
Each specification is versioned independently from this definition document and includes machine-readable versions in JSON and YAML formats.
Core File Specifications
- llms.txt Specification
- AI-readable business identity and context (Markdown format, llmstxt.org convention)
- llm.txt Specification
- Compatibility variant (should redirect to llms.txt)
- llms.html Specification
- Human-readable HTML presentation with Schema.org structured data
- ai.txt Specification
- AI usage permissions, restrictions, and attribution requirements
- ai.json Specification
- Machine-parseable AI interaction guidance (JSON Schema validated)
- identity.json Specification
- Structured canonical identity data aligned with Schema.org Organization
- brand.txt Specification
- Brand naming conventions and representation guidance
- faq-ai.txt Specification
- Structured Q&A content optimised for AI consumption
- developer-ai.txt Specification
- Technical context for AI systems assisting developers
- robots-ai.txt Specification
- AI crawler-specific access directives (robots.txt syntax)
Canonical Examples
All specifications include examples using a consistent fictional business profile. Complete example files are available in the examples directory.
JSON Schemas
Validation schemas are provided for JSON-based formats: