Scaling Databricks Genie 10x at GSK Global Supply Chain: Building a Production AI/BI Factory
Overview
| Experience | In Person |
|---|---|
| Track | Analytics & BI |
| Industry | Healthcare & Life Sciences |
| Technologies | AI/BI, Lakebase |
| Skill Level | Advanced |
Most organizations don't struggle to prototype conversational AI—they struggle to get it reliably into production at scale. At GSK's Global Supply Chain, operating across 37 GMP-regulated manufacturing sites with thousands of daily decisions, we built an AI/BI factory designed to move Databricks Genie from early experimentation to enterprise-grade production—serving analysts, planners and decision-makers across a complex, regulated environment.This talk explores the journey from pilot to platform: how we scaled Genie 10x by treating it not as a standalone tool, but as a production system embedded in a broader agent operating model—one that fuses technology, process, governance and organizational ways of working into a repeatable delivery engine.What we’ll cover:
- Genie Spaces lessons learned and best practices
- Enterprise observability dashboard for Genie
- CI/CD migration and deployment automation for DevOps
- Semantic caching for Genie
Session Speakers
Virginie Marelli
/Principal Generative AI Engineer
GSK
Kshitiz Sharma
/Sr. AI Engineer, AI FDE
Databricks