Agentic GraphRAG: Multi-Agent Knowledge Graph Construction for Research Teams

This session will explore a multiagent approach to GraphRAG that democratizes knowledge graph construction for research teams. Watch as four specialized AI agents collaborate: analyzing research focus to suggest entities, extracting knowledge from documents in parallel, merging duplicate entities, and optimizing graph structure—all while keeping researchers in control.

The session showcases a production-ready platform where teams upload PDFs, define their research domain, and receive AI-generated suggestions they can refine with domain expertise. Built with Neo4j, LangChain, and deployed on Cloudflare Workers, this agentic architecture dramatically improves extraction quality while maintaining accessibility for non-technical users.

You will learn advanced techniques for agent coordination, streaming processing with real-time progress tracking, and hybrid human-AI decision making. The live demonstration processes research documents in real-time, showcasing how agent specialization produces higher quality knowledge graphs than traditional single LLM approaches.

Speaker: Akhil Hemanth

Resources:
Get Started with Aura - https://bit.ly/3LOLrjh
Deployment Center - https://bit.ly/4jOelM3
Ground AI Systems and Agents with Neo4j - https://bit.ly/4oVsnyb

• #nodes2025
• #neo4j
• #graphdatabase
• #graphrag
• #knowledgegraph