Beyond Retrieval: Adding a Reasoning Layer to RAG
summary
A RAG stack augments vector retrieval with Neo4j so the system does entity extraction and Entity Linking into nodes then uses Cypher for Multi-Hop Traversal to surface hidden relationships and enrich queries before the LLM. Implementation uses a vector DB plus Neo4j, Vortex AI and Gemma. Result: search-to-answer drops from days or weeks to hours and 70% of internal support questions become automated using grounded, traversable knowledge paths and reasoned prompt injection.
AI-written summary · about
chapters (9)
- 0:00Introduction and speaker intro
- 1:58Limitations of traditional RAG
- 3:58Case study overview and challenges
- 6:57Architecture and reasoning layer
- 9:27Hybrid retrieval with reasoning
- 11:38When to use graphs in RAG
- 12:46Outcomes and business impact
- 14:56Building user trust with grounded answers
- 15:18Q&A session
related talks
39mEnhancing RAG with Neo4j Knowledge Graph
30mAgentic RAG with Knowledge Graphs Explained
27mWhen Vectors Aren’t Enough: Unlocking AI’s Hidden Connections with Graphs
36mAgentic GraphRAG: AI’s Logical Edge
20mHybridRAG: A Fusion of Graph and Vector Retrieval
27mEventKernel: Multi-Hop Reasoning and GraphRAG for AI-Powered Event Intelligence
15mRAG Beyond Chatbots
51mDeveloping a Custom RAG System