A structured representation of named entities and their typed relationships that search engines and AI answer systems use to resolve identity, classify categories, and select sources for retrieval.
Definition: An entity graph is a structured representation of named entities and the typed relationships between them. Nodes in the graph represent real-world things — organizations, people, products, locations, concepts — while edges represent the relationships that connect them: a company employs a person, offers a product, belongs to a category, or is identified by a public profile.
Definition: The entity graph separates the entity from the pages that mention it. A page is a document. An entity is the person, organization, product, or concept that the graph represents independent of any single document.
Premise: Google introduced its Knowledge Graph in 2012 to organize search around entities rather than keyword strings. The Google Knowledge Graph Search API classifies entities using schema.org types and returns structured data that identifies people, places, and things through names, descriptions, and relevance scores (Google Knowledge Graph Search API).
Premise: Google Cloud defines an enterprise knowledge graph as a system that "organizes siloed information into organizational knowledge, which involves consolidating, standardizing, and reconciling data" (Google Cloud Enterprise Knowledge Graph). The reconciliation process resolves duplicate entities by matching through "any combination of fuzzy text, common relationships, entity types, and its attributes."
Premise: Search engines and large language models use entity graphs at multiple stages of retrieval and answer generation.
Premise: During indexing, structured data on a page — JSON-LD markup using schema.org vocabulary — helps the system identify which entities a page describes and what relationships it asserts. Schema App research found that schema markup has "evolved from supporting individual search features into the semantic foundation that AI systems use to interpret entities, relationships, and meaning at scale" (Schema App, 2025).
Premise: During query resolution, the system matches a search query to an entity node rather than a keyword string. When a user searches for a company name, the entity graph disambiguates between the company, a person with the same name, and a product that shares the term. Google calls this entity resolution.
Premise: During answer generation, AI answer engines draw on entity relationships to decide which sources to retrieve. A source connected to a recognized entity through multiple relationship types — coverage by a known publication, structured data on an official page, consistent naming across third-party profiles — may provide the system with more resolution confidence than an isolated mention.
Premise: Google Search Central recommends adding Organization structured data to help Google "better understand your organization's administrative details and disambiguate your organization in search results" (Google Search Central: Organization markup). The elements a brand should document include:
The record should distinguish a verified external fact from an intended positioning claim. A brand's preferred category description is not automatically an externally established classification until independent sources confirm it.
Definition: Entity Graph names the relationship structure. Entity Clarity names the coherence of the public identity represented through that structure.
Premise: Within the Machine Relations model, entity-graph work belongs to the Entity Clarity layer. The graph provides a diagnostic map for finding conflicting names, missing relationships, ambiguous identifiers, and unsupported category claims across owned and third-party surfaces.
Working hypothesis (not measured): When public sources express compatible identifiers and relationships, an answer system may be more likely to resolve the intended entity correctly. A coherent entity graph — where the same name, the same relationships, and the same category associations appear consistently across the brand's website, structured data, public profiles, and earned coverage — reduces the system's disambiguation burden.
Start with a canonical entity record that lists the organization's name, key people, products, category, and public identifiers. Then compare this record against:
Mark conflicts, stale relationships, unsupported category claims, and missing identifiers. Preserve citations for externally asserted facts. The audit output is a gap list, not a score — each gap represents a specific relationship or identifier that the graph does not yet establish through independent confirmation.
Definition: An entity graph is a representation, not proof of visibility or recommendation. Adding a relationship to owned structured data does not establish that an external system accepted it. External behavior requires direct observation under a declared protocol.
Premise: A complete entity graph does not guarantee citation or recommendation by an AI answer engine. The graph reduces ambiguity and provides the system with resolution confidence, but source selection depends on additional factors including source authority, content relevance, and retrieval-model design.
Definition: A knowledge graph is the broader system — the entire database of entities, relationships, and facts that a platform like Google maintains. An entity graph is the subset relevant to a specific entity: the organization, its people, products, category, and the relationships that connect them. Brand-side entity graph work focuses on ensuring the brand's portion of the broader knowledge graph is accurate, consistent, and externally confirmed.
Premise: AI answer engines use entity relationships to resolve which entity a query refers to and which sources are authoritative for that entity. A brand with a coherent entity graph — consistent naming, confirmed relationships, structured data, and third-party coverage — gives the system higher resolution confidence, which may influence source selection during answer generation.
Premise: A brand controls its owned structured data (JSON-LD markup, schema.org types, sameAs references) and can maintain consistency across its website, public profiles, and product pages. The entity graph's strength depends on external confirmation: third-party coverage, public directory listings, and independent sources that verify the brand's identity and relationships. Owned signals alone are insufficient without earned signals.
Definition: Schema markup (typically JSON-LD using schema.org vocabulary) is the primary mechanism for communicating entity graph information to search engines. Organization, Person, Product, and other schema types let a brand declare entities and relationships in a format that search systems can parse directly. The markup is the encoding; the entity graph is the conceptual structure it represents.
Premise: Measurement happens through direct observation: does the search engine or AI system resolve the brand correctly, display accurate entity information, and cite the brand's sources for relevant queries? The Google Knowledge Graph Search API can confirm whether an entity is recognized (Google Knowledge Graph Search API). Answer-engine observation protocols can track whether the entity receives citation, attribution, and recommendation for its target query set.
Machine Relations definition. This is an established industry term. This page is Machine Relations' definition of it — not a claim to have originated the term.
Machine Relations' own methodology, dataset, and research pages related to this term. These are self-references, listed separately from Sources — they are not independent evidence.
Related concepts
Supporting research
Framework context