Knowledge Graph
A knowledge base that represents entities and their relationships as a graph structure, used by search engines and LLMs as a structured reference for entity identity
What is a Knowledge Graph?
A Knowledge Graph is a structured database of entities and their relationships, represented as a graph where nodes are entities (people, organizations, places, concepts) and edges are the relationships between them.
Google's Knowledge Graph, launched in 2012, is the most widely known example. Its launch slogan — "things, not strings" — marked a shift from treating search as keyword matching to treating it as entity recognition.
How does Google use the Knowledge Graph?
The Google Knowledge Graph Search API identifies entities based on the schema.org type system. When you search for a brand, Google checks its Knowledge Graph first to determine what kind of entity it is, then retrieves supporting information from the web.
This is why a brand registered clearly in the Knowledge Graph can appear in answers across Google AI Overview, Gemini, and Google Search's Knowledge Panel — without any specific page ranking highly.
How is the Knowledge Graph populated?
Google does not accept direct submissions. Instead, it harvests structured signals from multiple sources:
| Source | Weight |
|---|---|
| Wikidata | Very high — structured triples, CC0 license |
| Wikipedia | Very high — authoritative natural language descriptions |
| schema.org Organization (your site) | High — self-declared sameAs links |
| Google Business Profile | High — direct Google-managed channel |
| Authoritative press coverage | Medium–high — independent verification |
Knowledge Graph Grounding in LLMs
Beyond Google, LLMs use Knowledge Graph grounding as a third path for entity acquisition (alongside pre-training and RAG search). When generating answers, models can reference structured Knowledge Graph facts to ground their responses — which is why a brand clearly defined in Wikidata appears more consistently in AI answers than one described only in web documents.