Dun & Bradstreet adds AI credit workflows to Databricks

Dun & Bradstreet has made a set of agentic credit and portfolio management workflows available through Databricks Marketplace and Databricks OpenSharing, giving finance teams access to D&B business data inside the Databricks platform. The company said an anonymized portfolio example showed the tools increasing the bad capture rate from 30% to about 38%, equivalent to more than $6 million in additional bad debt avoided.

The workflows are built around D&B’s Commercial Graph, the company’s business information network anchored by the D-U-N-S Number identifier. Dun & Bradstreet said the graph supplies context on business identity, corporate relationships and risk, allowing AI agents to generate outputs that are consistent, explainable and auditable.

The release extends the use of commercial data into credit decisions at a time when finance teams are testing ways to apply AI to underwriting, credit limits and portfolio surveillance. Databricks Marketplace lets customers discover and access third-party data and applications, while OpenSharing is used to deliver the workflows so customers can combine D&B data with their own information inside Databricks.

How the workflows are designed to operate

Dun & Bradstreet described three main use cases across the credit lifecycle. In credit origination, a user can enter a single prompt that triggers business verification, enrichment with commercial and risk data, and a recommended handling path for the case. The workflow can support decisions to approve, decline or triage an applicant, and can suggest credit limits and payment terms, according to the company.

A second workflow focuses on credit policy. Dun & Bradstreet said an AI agent can assess how an existing credit policy is performing, identify ways to raise its predictive power and estimate the financial effect of proposed changes. In practice, that means a finance team can test policy adjustments against data before changing decision rules.

The third workflow is aimed at portfolio monitoring. Dun & Bradstreet said it can help finance teams review exposure across existing accounts, identify deteriorating customers earlier and find risk or growth patterns across customers, markets or geographies.

Scott Spencer, general manager for finance and credit at Dun & Bradstreet, said finance leaders face pressure to accelerate decision-making while managing risk with more precision. He said combining D&B’s verified business information with Databricks can reduce analytical work that might otherwise take days or weeks to seconds.

Sarah Branfman, global vice president for ISV and data partners at Databricks, said CFOs and finance teams need AI applications that improve decisions and risk management rather than experiments alone. She said the partnership is intended to bring customer data and D&B’s Commercial Graph together natively within the Databricks platform.

Dun & Bradstreet did not disclose pricing, customer adoption figures or implementation timelines in the announcement. The company framed the launch as part of broader enterprise AI adoption, where business context and governance are required for agents to make informed decisions within controlled processes.

This story draws on original reporting from Finextra Research.