Entity: definition and role in SEO and AI engines

A entity is a thing, a concept, a person or a place that can be identified uniquely, independently of the words used to name it. "Paris" the city, "Paris" the first name and "Paris" the plaster brand are three different entities behind one word; "GA4", "Google Analytics 4" and "Analytics" all name the same entity. Search engines have gradually replaced word analysis with the recognition of entities and their relationships, stored in the Knowledge Graph. The language models behind AI assistants work the same way: they describe the world as a network of entities and relationships.

Why engines reason in entities

For Google, a page isn't "about the word SEO" but "about the entity SEO, written by the entity Sylvain Vidal, of the entity GreenRed, covering the entities GA4, Google Ads and Search Console". That reading is what lets it judge a page's relevance and its author's expertise, in line with the E-E-A-T.

The entities that concern a business

• The organisation: the company itself, with its attributes (name, founding, headquarters, directors, activities, website, social profiles).
• The people: the founder, the experts, the authors of your content; a Person entity connected to the organisation strengthens perceived expertise.
• The place: the location, with its address, opening hours and service area; this is the LocalBusiness entity of your Google Business Profile listing, whose consistency is covered in NAP consistency.
• Products and services: each named offer, with its features and its price.
• Areas of expertise: the concepts (SEO, GA4, digital marketing) the company must be associated with to be cited on those themes.

Establishing your entity

1. A unique, constant name: the same spelling and the same form (with or without the legal suffix) across every source.
2. A reference "About" page: describing the organisation, its people, its history and its expertise, with verifiable facts; the method is in optimising your About us page.
3. schema.org markup: Organization (or LocalBusiness) with a stable identifier (@id), Person for your authors, sameAs pointing to official profiles (LinkedIn, Wikidata, directories), Article with author connected to Person. Local markup is covered in schema.org markup for local.
4. Links between entities: each article signed by its author, the author connected to the organisation, the organisation connected to its subjects through regular, cited content.
5. External confirmations: Google Business Profile, Wikidata, registries, sector directories, press, partners — all confirming the entity's attributes.
6. Consistency over time: a change of name, address or director is reflected everywhere, and quickly.

Entities and AI engines

When an AI assistant answers "which agency for a GA4 audit in Nantes", it assembles entities: the service, the place, the organisations connected to both, and the sources confirming them. A company whose entity is clear (activity, place, people, content on the subject) gets picked; a company described inconsistently, or absent from the sources, is ignored. Entity work is therefore the foundation of AI visibility, covered in structuring your pages for AI.

The frequent mistakes

• An inconsistent name: "GreenRed", "Green Red", "GreenRed SAS" depending on the page; three possible entities as far as the engines are concerned.
• Anonymous authors: articles with no named author connect the expertise to nobody.
• Isolated markup: an Organization schema on the home page with no link to the Person or Article entities.
• Contradictory sources: an old address on a directory, a former director on a registry.
• The keyword rather than the entity: repeating "SEO agency Nantes" establishes nothing; expert content on SEO, signed, connected to the organisation and cited by others, does.