Dun & Bradstreet partners Anthropic on compliance AI

Dun & Bradstreet has agreed to collaborate with Anthropic to bring its risk data into Claude, with a focus on business onboarding and compliance work.

The arrangement will let users build automated know-your-customer and know-your-business workflows inside Anthropic's AI assistant using Dun & Bradstreet data.

The integration will connect Dun & Bradstreet's Commercial Graph to Claude through Model Context Protocol server technology. Clients will be able to design workflows that verify business identities, assess risk and produce documentation for compliance reviews.

The focus is on regulated sectors, where companies need to check counterparties, map ownership chains and maintain records for later review. Financial institutions were cited as one example, using the system to assess new corporate customers and generate the paperwork required for onboarding.

At the centre of the system is Dun & Bradstreet's D-U-N-S Number, the identifier it uses to track commercial entities. The identifier and the surrounding data give Claude a verified basis for examining ownership, control and risk across business networks.

How it works

Users will be able to give natural-language instructions to create onboarding agents that draw on Dun & Bradstreet's business records and risk logic. Those agents are intended to replace a sequence of manual checks and disconnected tools with a single workflow inside Claude.

The integration is designed to support several common tasks, including verifying the identity of businesses across complex ownership structures, evaluating exposure across third-party and supplier networks, automating decision-making that uses risk intelligence, and creating risk-decision documentation.

The collaboration enters a growing market for AI systems tailored to compliance and due diligence, where companies are trying to reduce manual work without losing traceability. Banks, insurers and large corporates have been testing AI tools for document review and process automation, but adoption has been tempered by concerns over reliability, accountability and the need for clear records.

Alex Zuck, General Manager of Risk at Dun & Bradstreet, outlined the rationale for the integration.

"What makes this integration powerful is that Claude isn't just being given more data; it's being given the verified context and decision logic required to act. That means outputs that are not only personalized to the user and situation, but also explainable, auditable, and consistent, all essential components for organizations to act with confidence in high-stakes, regulated environments," Zuck said.

The emphasis on audit-ready output reflects a wider challenge for companies deploying generative AI in regulated settings. Businesses may be willing to use AI to gather information and draft assessments, but they still need to show how decisions were reached and what underlying data was used.

Dun & Bradstreet's Commercial Graph is intended to address part of that problem by supplying structured business identity data, rather than leaving the model to rely on general web content or user-provided material alone. In practice, a company using Claude for onboarding can pull in records tied to a standard business identifier and use them to support checks on legal identity, ownership and risk.

Compliance Push

The collaboration also highlights how data providers are trying to secure a larger role in the AI software stack. Rather than serving only as external sources for analysts or compliance teams, companies such as Dun & Bradstreet are seeking to embed their data directly into AI tools used in operational workflows.

For Anthropic, the arrangement adds a specialist business data source to Claude in an area where customers often demand strong controls. Compliance and onboarding tasks have emerged as an early target for enterprise AI because they involve repeatable steps, large volumes of documentation and clear rules, even if the underlying ownership structures can be complex.

Zuck said the business-identity layer was central to the offering.

"Agents in onboarding workflows must understand who they're dealing with. D&B gives Claude a persistent, verified view of business identity through the D-U-N-S Number, as well as the context required to reason about ownership, control, and risk. By bringing this business-verification layer into Claude, we're helping organizations move faster on onboarding without compromising safety, accountability, or trust," he said.

Dun & Bradstreet traces its history to 1841 and says the D-U-N-S Number became a global standard for identifying commercial entities after its introduction in 1963.