A new free open standard, EntityMap, aims to help organisations make their facts, relationships and evidence easier for AI systems to retrieve, understand and cite. The project, which has entered a 33-day public consultation, gives organisations a way to publish a structured, machine-readable map of what they do, what they offer, how their key entities relate to one another and where the supporting evidence sits on their website.
The aim is to reduce the need for AI systems to infer meaning from fragmented web pages, making it easier for search engines, retrieval systems and large language model applications to access factual information directly from the source. The specification is available at entitymap.org/spec/v1.0. The consultation runs until 30 June 2026, with the official launch scheduled for 1 July 2026.
Developers, publishers, structured-data specialists, AI retrieval practitioners, SEO professionals and data-quality experts are invited to review the specification, test implementation and contribute feedback through the EntityMap community forum and GitHub repository at github.com/entitymap.
Fred Laurent, CTO of InLinks and Waikay, said: “Where a sitemap tells search engines which pages exist on a website, EntityMap tells AI systems what an organisation is, what it does and how its knowledge connects. AI systems are increasingly being asked to summarise, recommend and explain organisations. If the underlying information is fragmented, incomplete or ambiguous, machines are forced to infer relationships. EntityMap gives them a structured source of truth to work from.”
AI systems are now being used to answer questions that would historically have been asked through search engines, websites, professional advisers or customer-service teams. Yet organisations have limited control over how those systems interpret their websites. A company’s products, services, expertise, locations, leadership, accreditations and relationships may be spread across many pages. AI systems often retrieve small fragments of this content and reconstruct meaning probabilistically, leading to incomplete answers, weak attribution or inaccurate representations.
EntityMap addresses this problem by allowing organisations to publish a single structured file that declares key entities, defines relationships and links each claim back to its source evidence. The file can be reviewed by humans before publication, then read by machines in a consistent format.
Dixon Jones, co-founder of Waikay and a long-standing specialist in search, entities and AI visibility, said: “The web was built around pages, links and prose. AI retrieval needs a clearer layer of meaning and evidence. EntityMap is designed to help organisations say: these are the things we know, these are the relationships between them, and this is the evidence that supports those claims. This consultation is about opening the standard up to scrutiny. We want people to test it, challenge it, implement it and help improve it before the formal launch.”
EntityMap is published as a structured file at a predictable location on a website. It identifies important entities associated with an organisation, such as products, services, people, topics, locations, claims or areas of expertise. It then maps the relationships between those entities and links them to supporting pages, allowing machines to retrieve an evidence-backed view of the organisation rather than relying only on isolated page fragments.
The project includes a specification, documentation, examples and validation tools. It is published under CC BY 4.0, with no subscription, vendor lock-in or proprietary software requirement. The 33-day consultation is intended to give the technical community time to review the structure, test practical implementation and identify improvements before the standard is finalised.
The project team is particularly seeking feedback from developers and AI retrieval specialists, structured-data and schema practitioners, technical SEO professionals, publishers and website owners, data-quality and governance experts, organisations concerned about AI misrepresentation, and tool builders interested in creating generators, validators or integrations.
R.V. Guha, one of the founders of Schema.org, has reviewed the project and said: “This is a good thing for the world.” The first phase of the consultation focuses on technical review, early implementation and community feedback. Wider adoption, sector-specific applications and further research will follow after the consultation period.
EntityMap is relevant to any organisation that needs AI systems to understand its information accurately. Potential use cases include healthcare organisations publishing accurate service, treatment or professional information; financial services firms clarifying products, risks, advice boundaries and regulated information; legal, professional-services and B2B organisations with complex expertise; publishers that want clearer attribution for their knowledge and editorial content; brands concerned about how AI systems describe their products, people or services; and technology teams building retrieval-augmented generation systems that need cleaner source data.
The project is not designed to replace existing web standards. Instead, it is intended to add a structured evidence layer for AI systems that need to understand not just what pages exist, but what an organisation knows and how that knowledge connects.
