Dun & Bradstreet Deploys Agentic AI Workflows on Databricks

The integration allows finance departments to bypass manual analysis, which previously required days or weeks to complete. In a recent anonymized test case, these AI-driven workflows successfully raised bad debt capture rates to 38%, resulting in over $6 million in avoided losses. The system relies on the D-U-N-S Number, a long-standing identifier that provides the necessary context for AI agents to reason accurately about business relationships and risk.

Scott Spencer, General Manager for Finance and Credit at Dun & Bradstreet, noted that the platform focuses on streamlining credit origination and policy optimization. By functioning within the Databricks Marketplace and OpenSharing, the tools enable companies to combine their own internal data with D&B's verified commercial information. This approach covers three primary areas: faster credit decisioning through automated prompts, adaptive policy optimization for testing financial impacts, and continuous portfolio risk monitoring. Sarah Branfman, Global VP at Databricks, emphasized that the collaboration moves beyond experimental AI, offering finance teams a practical, governed environment to improve precision in risk management.