Ryan Knight

Databricks Genie gives analysts a natural language way to ask questions of governed Lakehouse data. But some of the highest-value questions are relationship questions: which accounts are connected to suspicious activity, which merchants sit inside unusual communities, and which patterns only appear two or three hops away? In this 20-minute lightning talk, Neo4j shows a graph pattern for enriching Delta Lake data with graph-derived context that Genie can query like ordinary dimensions. Using a financial crime investigation, we'll show how connected scores and communities can turn flat results into explainable, relationship-aware answers without changing the analyst workflow.