Case Study: BASF achieves governed enterprise AI at scale with Databricks

Multi-Agent Supervisor Architecture Orchestrating Enterprise AI at Scale

BASF Coatings, part of BASF, needed a way to turn fragmented structured and unstructured data into actionable insights for sales and marketing, while improving collaboration, governance, and real-time decision-making at global scale. The challenge was to manage growing complexity across data sources, teams, and workflows using a more modular and controlled AI agent system, with Databricks Mosaic AI and AI/BI Genie as key components.

Databricks implemented a governed multi-agent “supervisor” architecture that combined Genie for structured data, vector search for unstructured content, and agent orchestration in Microsoft Teams. The solution powered the Marketmind use case, helping sales teams surface personalized recommendations and faster follow-ups; it completed a 25-user pilot, is moving toward rollout to more than 1,000 sales representatives, and has already improved prep, customer conversations, and time spent selling.

BASF

Adrian Fierro

Head of Global Market Intelligence