PreCon: Governed Finance AI on Databricks - From Data to Variance Agents

This full-day pre-conference workshop shows how to build an end-to-end financial intelligence solution on Databricks using a realistic financial variance-analysis use case. Participants will begin with ingestion of source financial and operational data, then build a medallion architecture that produces trusted, reusable finance data products for reporting, analytics, and AI workloads. The session will cover Delta Lake design patterns, Delta Uniform for broader data interoperability, and Unity Catalog controls that protect sensitive financial information through attribute-based access control (ABAC). From there, the workshop moves beyond the platform foundation. Participants will create governed business metrics for actual-versus-budget and period-over-period variance analysis, expose those metrics through Genie Spaces for natural-language exploration, and build AI agents that can investigate material variances, retrieve supporting evidence, and prepare analyst-ready explanations. The session will also cover how to maintain traceability, apply approval boundaries, and ensure that AI-generated insights remain tied to authorized data and approved financial definitions. Attendees will leave with an implementation blueprint and practical patterns for moving from fragmented finance data to governed, AI-assisted variance analysis on Databricks.