Dun & Bradstreet adds agentic AI to compliance workflows, targets 70-90% faster processing

Dun & Bradstreet has rolled out new agentic AI capabilities inside its compliance workflow stack, with the company saying the system can cut processing times by 70-90%. For PSPs, banks, FinTechs, and crypto firms, the part that matters is not the AI label; it is whether onboarding, screening, and due diligence can move faster without turning compliance into a manual relay race.

1. The new offering sits within the D&B Risk Analytics platform and uses its model context protocol (MCP) server to work “fully agentically,” according to the release. In practice, Dun & Bradstreet says organizations can embed verified D&B data, models, and workflows into AI assistants and custom agents.
2. The company says the setup is aimed at speeding up onboarding, screening, and due diligence, while continuously monitoring know your customer/know your business (KYC/KYB) compliance. The data source behind that workflow is the D&B Commercial Graph, which the release describes as verified business data used to make faster decisions.
3. Alex Zuck, general manager of risk at Dun & Bradstreet, framed the product as a response to a compliance function under pressure from “the pace and complexity of regulations” and the cost of fragmented tools, data, and services. He also argued that companies have struggled to implement the data needed to make AI useful in compliance, which is the usual catch with these launches: the model is not the hard part, the data plumbing is.
4. Zuck said traditional KYB onboarding can take days or even weeks, and that determining the ultimate beneficial owner of a company often means searching registries, validating documents, and manually piecing ownership together. Dun & Bradstreet says it can now complete that process “in seconds” and deliver it into client onboarding workflows without human intervention.
5. The release lands in a compliance market where alert volumes and false positives already eat into operational budgets. PYMNTS also cited Flagright co-founder and CEO Baran Ozkan, who said the larger value of AI is on the investigation side, because detection alone does not remove the work that follows when teams have to prove a hit is not a real risk.

For high-risk operators, the practical question is simple: does this reduce manual review without weakening the audit trail? Dun & Bradstreet is betting that verified data plus agentic workflows can do both, which is exactly the pitch PSPs and onboarding teams will be asked to judge against their own risk tolerance and internal controls.