Data & retrieval
Knowledge Graph
A structured representation of entities and the relationships between them — used to give AI systems explicit, queryable facts.
01 ——
In plain English
A knowledge graph is a database where information is stored as a network of entities (nodes) and relationships (edges). Where a normal database has rows and columns, a knowledge graph has facts like "Anthropic — founded — 2021" or "Claude — made by — Anthropic."
Why pair it with LLMs:
- Explicit facts — the model can look up rather than recall (more reliable)
- Multi-hop queries — follow chains of relationships the model can't hold in context
- Provenance — every fact has a source you can cite or update
- Reasoning over structure — graph queries answer "which of X are also Y" questions instantly
Common implementations:
- Neo4j, Memgraph, ArangoDB — graph databases
- Google Knowledge Graph — the original public knowledge graph behind search
- Wikidata — open, community-edited
- Microsoft GraphRAG — combine a knowledge graph with retrieval-augmented generation
Knowledge graphs are most useful where facts and relationships matter more than nuance — compliance, healthcare, enterprise search, fraud detection.
02 ——
Related terms
Knowledge Base
A structured collection of documents an AI system can search and quote — the source-of-truth corpus that grounds RAG and many AI agents.
RAG
Retrieval-Augmented Generation — a technique that gives an AI model access to external documents before it answers, so it can cite real, up-to-date sources.
Vector Database
A database optimised for storing and searching embeddings (numerical representations of text or images) by similarity.
Semantic Search
Search that finds results by meaning rather than exact keyword matches — so "car" finds results about "automobile" too.
AI Search
A search experience that returns a synthesised AI answer with citations — instead of a ranked list of links — powered by retrieval plus an LLM.
Embeddings
A way of converting text (or images) into lists of numbers so an AI can measure how similar two pieces of content are.
Back to glossaryLast reviewed May 2026