Google Connects Gemini Finance Agents to D&B Data, Adding 50-Plus Skills for Regulated Work

Financial institutions can now put external commercial data directly into Gemini agents built for regulated work. Gemini Enterprise for Financial Services combines Dun & Bradstreet’s Commercial Graph connection with a managed research agent, more than 50 specialized skills and enterprise data connectors.

Google Cloud announced the product on August 25. It uses Model Context Protocol integrations to connect Gemini Enterprise with D&B’s Commercial Graph, placing commercial information on the path an agent uses to answer a request.

The design shifts the product beyond a model working from a static dataset. Google says its Financial Research agent can fetch external data streams as needed, then cross-reference a customer’s data with D&B’s registry before it produces a response. Retrieval-augmented generation, or RAG, is the method used to bring that retrieved material into the model’s response process.

The data connection changes the research workflow

D&B positions Commercial Graph as a context layer for understanding business identity, relationships and risk globally. The intended result is less reliance on a one-time manual review of a business client: an institution could use automation for approval and receive alerts if that client later becomes a financial or legal risk.

That promise depends on both the external registry and an institution’s own customer data. D&B says the integration can make agent outputs more consistent, auditable and explainable, but those are product aims rather than an indication that the agent can independently make regulated decisions.

8%Financial institutions saying their enterprise data is ready for the technology

A Dun & Bradstreet survey cited in the article found that 8% of financial institutions say their enterprise data is ready for the technology.

Current data does not remove the control problem

Google positions the service for regulated processes including know-your-customer and anti-money-laundering work. Those uses raise the stakes of an agent’s output: financial institutions must still put safeguards in place to manage model risk and regulatory exposure.

The distinction is material when an agent moves from summarizing records to recommending a credit line or flagging risk. A verified external source can improve the context available to an agent, but it does not settle the institution’s responsibility for the architecture and controls around the resulting workflow.

What the finance package includes

• A managed Financial Research agent that retrieves external information and cross-references customer data with D&B’s registry.
• More than 50 specialized skills, enterprise data connectors and an expanding third-party agent ecosystem for financial roles and workflows.
• Support within Google Workspace and Microsoft 365, where analysts can generate, analyze and export AI-powered insights.

A vertical template extends beyond finance

Finance is launching alongside Gemini Enterprise for Legal, aimed at law firms and corporate legal departments. Google says legal agents can support contract review, legal research, regulatory monitoring and court-filing preparation.

Google describes finance and legal as the first packaged industry solutions on Gemini Enterprise. Rather than offer only a general-purpose assistant, the company’s template combines tailored agents, specialized shortcuts for particular workflows, data connectors and sector-specific model optimizations.

The unresolved test is whether institutions can pair those packaged capabilities with data that is ready for agent use and controls suitable for regulated decisions. The new D&B connection addresses the availability of external business context; it does not remove either operational requirement.