What is a knowledge graph?

Explained

A knowledge graph represents information as nodes (entities such as customers, products, contracts or people) connected by labelled edges (relationships such as "signed", "supplies" or "reports to"). Because the connections are explicit, a system can follow them to answer questions that span several sources, such as which suppliers are affected by a delayed part. Google popularised the term in 2012 to describe search based on "things, not strings". In enterprise AI, knowledge graphs improve retrieval and reasoning in RAG by adding structure and meaning beyond what text search finds.

How a knowledge graph works

• Entities: the things that matter to the business, each with a stable identifier and attributes.
• Relationships: typed links between entities, which can also carry attributes such as dates or sources.
• Schema or ontology: the agreed types of entities and relationships, so that different teams describe the same thing the same way.
• Queries: graph query languages or APIs that follow paths, for example from a customer to their contracts to the products in them.

Knowledge graphs and generative AI

Language models are good with text but weak at keeping track of many exact relationships. A knowledge graph supplies that structure. In GraphRAG approaches, a model extracts entities and relationships from documents to build a graph, then uses it to answer questions that need connections across many documents.

The graph can also work the other way: as a source of verified facts that the model must respect, which limits invented details in answers.

When it makes sense

• Questions depend on relationships across systems, not on a single document.
• The same entity (a customer, an asset) appears under different names in different sources.
• Answers must be explainable: the path through the graph shows why.
• Someone will own the schema and data quality over time; a graph without an owner decays quickly.

Frequently asked questions

Decision guides that use this term

• GraphRAG vs vector RAG: when does your knowledge assistant need a graph?

How to put it to work

• How to keep expert knowledge when people retire
• How to modernise legacy code with AI

Further reading

• Data Fabric: Weaving a Unified View of Your Data Landscape for Strategic Advantage

Related terms

• Enterprise AI AI built and run for large organisations: governed, secure, scalable and integrated with existing systems such as ERP, CRM and data platforms.
• RAG A technique that retrieves relevant passages from a knowledge base and gives them to a language model as context, so its answer rests on those sources.
• Retrieval The search step that fetches the documents or passages most relevant to a query, before they are shown or passed to a model.
• Big Data Data sets so large, fast-moving or varied that traditional databases and tools cannot store or process them efficiently; a key input for AI.
• Context Engineering The practice of deciding what information a model sees at each step (instructions, retrieved documents, tool results, memory) and fitting it into the context window.
• Corpus A collection of documents or texts used to train a model or to serve as the knowledge base an AI system retrieves from, as in RAG.