We sit down with Philip Rathle, Chief Technology Officer of Neo4j, to explore a question that’s becoming urgent in the age of AI: What happens when powerful models operate without context, governance, or explainability?
As generative AI reshapes enterprise technology, graph databases are quietly becoming a foundational layer for accuracy, transparency, and data control. Philip shares why AI systems struggle without structured relationships, how graphs reduce hallucinations, and what this means for privacy teams navigating Customer 360, data subject requests, and regulatory pressure.
00:00 Introduction.
02:30 From chemical engineering to data architecture.
05:45 What a graph database actually is, and why it’s simpler than it sounds.
10:30 Why relational databases struggle with complex, connected data.
17:45 The AI tailwind: hallucinations, explainability, and governance.
23:10 Customer 360 and resolving fragmented identities.
28:15 Handling data subject access and deletion requests with graphs.
30:45 The double-edged sword: when graph power becomes surveillance risk.
37:00 AI models, privacy controls, and why not everything belongs in an LLM.