Guide

Building entity-rich sites.

Search and AI systems do not read your site as text. They read it as entities: the people, products, and organizations they can match to something they already know. When the match fails, the fact goes with it.

Why entities, not keywords

A keyword is a string. An entity is a thing with an identity, a Wikidata node, a set of relationships. When a machine resolves your founder to a real person and your company to a real organization, it can carry your claims with confidence and connect them to everything else it knows. When it cannot, it guesses or drops the claim. Entity coverage is the difference between being understood and being approximated.

Find your gaps first

Start by seeing what a knowledge graph already resolves about you and what it cannot. Entity Gap X-Ray maps your domain's entities and marks the missing nodes and the dangling connections. Work the critical gaps first, usually a founder or an organization with no node at all, because those anchor the relationships that would resolve the entities around them.

Mark up what is really there

Then state the entities explicitly in your markup: an Organization for the company, Person nodes for the people, and sameAs links to authoritative profiles so the match is declared rather than inferred. Keep it honest, only markup that agrees with the visible page. Most sites deploy far less schema than the documentation implies matters; our State of Schema.org Deployment report shows what the web actually ships, which is a useful floor to beat.

Pull clean data, with receipts

When you enrich your entities from a knowledge graph, keep the provenance. Wikidata MCP returns every answer with its receipt: the statement node, its rank, its references, and when it was retrieved. That lets you cite the source of each fact you add and re-check it later, instead of copying a number that was true once and may not be now. Entity-rich is not a one-time pass; it is markup you can stand behind and keep current.