The AI Native Dev - from Copilot today to AI Native Software Development tomorrow The Graph Layer Behind NASA’s Breakthroughs | Michael Hunger
Jul 8, 2025
Michael Hunger, VP of Product Innovation at Neo4j, reveals the game-changing potential of graph databases in AI. He discusses how structured relationships in data uncover intent, addressing the black-box issue in vector-based approaches. Michael emphasizes the benefits of modeling business intent and highlights innovative use cases like fraud detection. The conversation also covers MCP's role in enhancing developer productivity and security best practices, showcasing a grassroots movement in graph technology that’s revolutionizing AI development.
AI Snips
Chapters
Transcript
Episode notes
Graph Databases Capture Intent
- Graph databases store relationships as first-class citizens, persisting connections at insertion time for fast queries.
- This approach differs from relational databases where connections are computed on query time, enabling richer business intent representation.
Relationships Encode Business Intent
- Relationships in graph databases carry business intent and can be scored with weights and time frames.
- These attributes help detect patterns like fraud by analyzing connection strength and timing.
Graph Improves RAG Retrieval
- Vector search in RAG lacks explainability and context, leading to incomplete or inaccurate results.
- Graph-enhanced RAG adds adjacency and rich context, improving relevance and enabling auditability.