An entity is more than its name: what matters is the thing itself, not the string of characters used to name it—hence the phrase “things, not strings.” One name can stand for several entities: “Jaguar” refers to both a big cat and a car brand. Conversely, one company often appears under several names, such as its full legal name and a short name.
Attributes, relationships, and identifiers
An entity is described by attributes such as its name, headquarters, or offering, and by relationships to other entities: a product belongs to a company, a person runs it. So that machines can recognize an entity reliably, knowledge bases give it a unique identifier; in the free knowledge base Wikidata, that is a Q followed by a number.
Structured data describes entities, too. In schema.org, a widely used vocabulary for this, the most generic type is Thing; more specific types include Organization, Person, Product, and Place. The sameAs property points to pages that unambiguously indicate an entity’s identity, such as official profiles or a Wikidata entry.
Why entities matter for GEO
In GEO, entities are central because many questions put to AI systems revolve around them: which provider offers something, what a product can do, where a company is based. A good answer requires the system to recognize which company is meant and to attach the right facts to it. Otherwise it can confuse namesakes or mix up details, and the answer may contain convincing but false statements: so-called hallucinations.
This can particularly affect lesser-known companies: a study published at a research conference in 2023 found that language models reproduced facts about less popular entities from their parametric knowledge far less reliably than about well-known ones, and that systems retrieving relevant content performed considerably better for exactly these entities. For a mid-sized company, this is a reason to provide clear information that AI systems can retrieve.
In Bing’s Webmaster Guidelines, which also apply to Copilot, Microsoft recommends naming people, organizations, products, and locations clearly and consistently. According to Microsoft, clear entity definition improves both the chance of content being selected as the basis for AI answers and the accuracy of AI citations.
What makes an entity unambiguous
A company becomes clearly identifiable when its information matches everywhere; how to achieve that is covered under brand fact consistency. It helps to have one official name each for the company, its products, and key people, used consistently; a factual About page with the core facts; and structured data of the type Organization (Organization markup) with sameAs links to official profiles. How AI systems recognize entities is up to their providers and rarely documented, but consistent information is a reliable foundation for any system and a building block of your AI visibility.