Graph RAG
Building a graph of entities and relationships from documents and combining graph retrieval with LLMs for complex questions.
In plain language
Vanilla RAG retrieves passages by similarity and struggles with global questions such as how events relate. GraphRAG extracts entities, relationships and community summaries, expands along the graph or uses summaries at query time, then asks the model to answer. It suits multi-document, cross-entity, global-view tasks such as due diligence, supply-chain risk and research reviews. The cost is high construction and maintenance, and entity-resolution or relation-extraction errors propagate through the graph.
Example
For “which suppliers are affected by both a port strike and chip export limits,” graph retrieval can intersect supplier-port-component-policy relationships.
Recommended tools
Related terms
Agentic Workflow
Breaking a task into steps planned by AI, executed by tools and reviewed by humans.
Retrieval-Augmented Generation
Retrieve relevant passages from a knowledge base, then have the model answer from them.
Vector Database
A database built to store vectors and retrieve them by similarity.
Agent Memory
A mechanism for an agent to store and retrieve user preferences, past tasks and key facts across sessions.