A Wikidata item that fixed entity confusion: a measured before/after

A Wikidata item that fixed entity confusion: a measured before/after

The question: when two people or brands share a name, does creating a disambiguating Wikidata item change how often the right entity is surfaced and cited?

A documented case: a consultant shared a name with a better-known athlete, and AI Overviews plus knowledge results consistently surfaced the athlete on the consultant's branded queries. The consultant created a Wikidata item with occupation, employer, and sameAs links to LinkedIn, a personal domain, and two published bylines. Reported result within ~8 weeks: correct entity resolution on branded queries rose from roughly 30% to 80% of sampled results, and citations to the consultant's site in AI answers became consistent.

The mechanism is entity disambiguation — Wikidata is a confirmed input to the Knowledge Graph, and clean identifiers reduce ambiguity.

Counter-evidence: the consultant also published more first-party content in this window, which independently strengthens the entity.

Caveat: 'correct resolution rate' was measured by manual sampling, not at scale, and AI Overview outputs are nondeterministic — the same query varies run to run.

What we still don't know: how much weight Wikidata carries versus other identifiers, and whether the effect holds for entities with no other distinguishing signals to anchor.