Compliance work has long been one of the most labour-intensive corners of business operations. Know your customer (KYC) and know your business (KYB) checks require pulling data from dozens of sources, cross-referencing records, screening against sanctions lists, resolving false positives, and documenting every decision. For regulated industries, that process can take days per entity.

Dun & Bradstreet announced on June 18 that it has embedded agentic AI directly into these workflows, with results that sound almost implausible: up to 96% reduction in processing time, up to 20 times more review capacity, and a 50-90% reduction in false positives. The capabilities are live now inside the D&B Risk Analytics platform and via Model Context Protocol (MCP) integration in Anthropic’s Claude.

What the Integration Actually Does

The mechanism is worth understanding, because it illustrates how the current wave of enterprise AI is different from what came before.

Rather than requiring compliance teams to log into a separate platform to check business data, D&B’s MCP integration lets an AI assistant like Claude reach directly into D&B’s Commercial Graph, the company’s database of business identities and relationships, during a live conversation or automated workflow. An analyst can ask a question, or a workflow can trigger a check, and the AI agent retrieves verified D&B data in real time, applies compliance logic, and surfaces a finding, all without leaving the workflow.

The automation extends across the full compliance cycle: entity identification, sanctions and watchlist screening, prioritisation of alerts, false positive resolution, due diligence documentation, and ongoing monitoring. Tasks that previously required a compliance analyst to manually assemble information across multiple systems now happen in a single automated step.

A Broader Ecosystem Play

D&B is not betting on a single AI platform. The company simultaneously announced it is becoming a trusted data partner for agentic workflows across Anthropic’s Claude, OpenAI’s GPT models, and Microsoft Copilot. Additional integrations with Google’s AI ecosystem are planned.

This multi-platform approach reflects something important about where enterprise AI is heading. AI agents need reliable, structured data to act on. General language models are good at reasoning, but they do not know whether a specific company is on a sanctions list or whether a supplier’s ownership structure has changed. That gap is where data companies like D&B are positioning themselves, as the verified ground truth that agents can act on without second-guessing.

The Model Context Protocol, which Anthropic originally developed and has since become an open standard, is the plumbing that makes this work. MCP lets AI agents call external data sources with defined schemas, get back structured results, and incorporate them into their reasoning. It is quickly becoming the standard interface between AI agents and enterprise data.

The Compliance Problem Is Not Small

To understand why this announcement matters, consider the scale of the problem. Any regulated business, from banks to law firms to insurance companies to fintech startups, must verify the identity and risk profile of entities they do business with. A mid-sized financial institution might onboard hundreds of new business clients per month. Each one requires a KYB check. A single check involving multiple corporate entities can take a compliance analyst the better part of a day.

Scale that across an organisation, factor in the cost of analyst time, and the cost of delays in onboarding new clients, and the numbers become significant. D&B’s claim of 20 times more review capacity is not just an efficiency metric. It means a compliance team can handle the same workload with a fraction of the headcount, or handle far more volume without adding staff.

The 50-90% reduction in false positives matters for a different reason. False positives, legitimate businesses flagged as potential risks because their names resemble a sanctioned entity, consume enormous analyst time to resolve. Reducing them directly reduces the operational cost of compliance.

What This Means for Business

Compliance as a bottleneck is becoming optional. For years, regulated industries have accepted slow client onboarding as the cost of doing things properly. Agentic AI is challenging that assumption. The same rigour is now achievable in seconds, not days. Businesses that move first on this will be able to onboard clients faster than competitors while maintaining the same risk standards.

Data quality is the new competitive advantage. The D&B announcement highlights a structural shift: the value of AI agents is limited by the quality of the data they can access. Organisations that have invested in clean, structured, verified data are in a position to automate faster and more reliably than those still working from spreadsheets and siloed systems. If your business data is fragmented, AI agents cannot help you at full speed.

MCP is becoming the integration standard. If you are evaluating enterprise AI tools, look at whether they support MCP. It is rapidly becoming the standard way AI agents connect to external systems. Vendors that support it are interoperable with each other. Those that do not are building walls around themselves that will be harder to maintain as the standard matures.

The regulatory side of AI agents is real. Compliance is one area where AI agents need auditable decision trails. D&B’s announcement specifically mentions full traceability and policy alignment as part of the offering. As AI agents take on more consequential decisions, the ability to show regulators exactly what the agent did, and why, will not be optional. Businesses adopting agentic AI for regulated workflows should treat auditability as a core requirement, not an afterthought.

Enterprise DNA works with businesses at every stage of their AI adoption journey. Whether you are just starting to understand what AI agents can do for your operations, or ready to deploy them across specific workflows, the practical question is always the same: where does automation create the most value with the least risk? Compliance is one of the clearest answers to that question that has emerged in 2026.