← back
Agentic GraphRAG: Simplifying Retrieval Across Structured & Unstructured Data — Zach Blumenfeld
Takeaway
For aggregations, similarity and relationship questions, give your agent a knowledge graph and Cypher-generating MCP tool — not just a vector index.
Summary
- Demos an employee-skills assistant: starts with plain vector search over resume PDFs in Neo4j — fails on aggregations ('how many Python devs') because semantic search returns top-k=5.
- Replaces it by extracting an entity graph (Person-knows-Skill, Person-publishes/builds/manages-Thing-belongs-to-Domain) using Pydantic enums and LangChain.
- Agent built with Google ADK gets an MCP tool that reads schema and generates Cypher; same questions now return precise aggregations (28 developers) and explainable similarity (Sarah by skill overlap).
- Argues knowledge graphs let agents decompose questions into multiple structured queries — vector search alone can't aggregate or traverse relationships.
- Walks Jupyter notebooks against Cyberdyne Systems demo data.
graphragneo4jmcp
Original description
Agentic workflows often become complex, brittle, and hard to maintain when they need to retrieve and reason across both structured data (typically requiring precise query execution) and unstructured data (commonly handled via vector search in RAG). In this talk, we’ll explore how mapping key information into a knowledge graph can simplify these workflows and improve retrieval quality. You’ll learn core concepts behind GraphRAG, how to integrate it into agent tools, and get access to end-to-end code examples so you can start building right away. About Zach Blumenfeld Zach Blumenfeld is an AI/ML graph specialist at Neo4j who helps engineers, data scientists, and business leaders leverage graph technology for analytics and intelligent applications. His expertise spans several dynamic fields, including GraphRAG and AI systems, criminal fraud detection, entity resolution, and recommendation engines. Recorded at the AI Engineer World's Fair in San Francisco. Stay up to date on our upcoming events and content by joining our newsletter here: https://www.ai.engineer/newsletter