What is a knowledge graph?

A knowledge graph is a structured database of real-world entities — people, places, organizations, concepts — and the relationships connecting them, which search engines and AI systems use to understand facts directly rather than inferring them from unstructured text alone. Instead of treating a webpage as a wall of text to be parsed for keywords, a knowledge graph stores facts as discrete nodes linked by explicit relationships, such as a person node connected to an employer node by a "works at" relationship. This lets a system answer factual questions with confidence instead of guessing from context, and it's why the same entity can be recognized consistently across billions of pages rather than treated as a fresh string of text each time. Wikidata, a public structured dataset maintained independently of any single search engine, is one of the sources these graphs draw on to cross-check facts about an entity before treating them as reliable.

A concrete example

Google's Knowledge Graph is what powers the information panel that appears alongside search results for a well-known company or public figure. Search a major brand name and the panel showing its founding date, headquarters, and key people isn't pulled from a single ranked webpage — it's assembled from structured facts Google already holds about that entity, cross-referenced against multiple sources for confirmation. The underlying challenge, known as entity resolution, is deciding that mentions of a name across different sites all refer to the same real-world thing rather than a coincidence of spelling; when resolution fails, two unrelated people or companies with similar names can get merged into one incorrect panel, which is why consistent naming and identifiers across your own pages matter.

How facts get into a knowledge graph

Adding schema.org markup in JSON-LD format to your own pages — tagging your organization, people, and products as explicit entities with properties like name, founding date, and sameAs links to authoritative profiles — is the direct mechanism for feeding facts into a knowledge graph, rather than hoping a system infers them correctly from prose. The structured data SEO guide to JSON-LD walks through implementing this markup step by step. If you're new to the underlying concept of an "entity" itself, the structured data glossary entry is a useful starting point before tackling markup.

What Is a Knowledge Graph? A Quick Definition