What is Local entity graph?

The local entity graph is the set of a business’s location entities — one per physical location — as an AI engine resolves them across map data, business directories, data aggregators, and the open web, before it will consider naming any of them in a local answer.

It is the local-scoped analog of the brand entity graph: where brand disambiguation resolves a company to a single knowledge-graph node, the local entity graph resolves each location to a single, unambiguous, place-anchored node. Resolution is the first and most under-appreciated failure point in local AI search — when an engine sees four slightly different names, two stale addresses, and an old phone number for what is really one location, it treats them as several weak, low-confidence entities rather than one strong one it is willing to name, and the business silently never enters the candidate set regardless of review quality. A clean local entity graph requires one canonical name per location, LocalBusiness (or the right subtype) schema on a distinct indexable page per location, and reconciled map, directory, and aggregator presence all agreeing on the same canonical facts. For multi-location brands this is the single highest-leverage local fix, because a single ‘locations’ page listing many addresses resolves to nothing specific and gets no location named.