Dun & Bradstreet's Commercial Graph—a trusted database containing verified information on more than 500 million business entities—is being woven into three of the most widely used AI platforms. Starting June 16, 2026, users of Microsoft 365 Copilot, OpenAI’s ChatGPT, and Anthropic’s Claude will be able to pull D-U-N-S numbers, corporate hierarchies, credit scores, and firmographic data directly into their workflows without switching applications. The integration, announced via PRNewswire from Hong Kong, marks one of the largest injections of vetted B2B data into the generative AI ecosystem, and it leans heavily on the Model Context Protocol (MCP) to make the connection seamless.
The move transforms how enterprise users interact with AI assistants. Instead of asking a chatbot to draft a generic market analysis, they can now instruct it to incorporate live, authoritative company intelligence from Dun & Bradstreet. For Windows users, the immediate impact lands inside Microsoft 365 Copilot, where the D&B data will appear in Word, Excel, Teams, and Outlook—turning each app into a live business research terminal.
The Data Powering Enterprise AI
Dun & Bradstreet’s Commercial Graph isn’t a static list. It’s a living map of global commerce, continuously updated with legal names, trade names, physical and digital addresses, industry codes, revenue estimates, employee counts, family trees, and financial stress scores. At its core is the D-U-N-S Number, a unique nine-digit identifier used by governments, banks, and supply chain systems worldwide to verify a business’s existence and ownership.
For years, accessing this data meant separate logins, API calls, or dedicated portals. The new integrations collapse that friction. A sales representative drafting a proposal in Microsoft Word can highlight a company name and ask Copilot to enrich it with D&B data—revenue, key contacts, risk indicators—without ever leaving the document. A supply chain analyst using Excel can pull entire supplier portfolios into a spreadsheet and immediately stress-test them against Dun & Bradstreet’s predictive scores.
The same capability extends to ChatGPT and Claude. OpenAI and Anthropic have both adopted the Model Context Protocol, allowing their assistants to securely reach out to external data sources. With the D&B integration, a user of ChatGPT Enterprise can query “Give me the ultimate parent of this subsidiary in Germany, plus its credit limit recommendation,” and receive an answer grounded in Dun & Bradstreet’s latest records.
How the Integration Works
The technical backbone is MCP—the Model Context Protocol, originally open-sourced by Anthropic in late 2024. MCP standardizes how AI models connect to external tools and data sources, much like USB-C did for hardware. Instead of every vendor writing bespoke plugins, they build a single MCP server that any compliant client can query.
Dun & Bradstreet has built an MCP server that exposes curated endpoints from the Commercial Graph. When a Microsoft 365 Copilot user prompts for business information, Copilot sends a structured request to that server, which returns the relevant data in a token-efficient format the AI can digest and cite. The exchange respects user authentication and entitlement, so a company only sees data tiers it has licensed—Enterprise, Advanced, or Standard.
Crucially, the integration does not train the underlying large language models on proprietary business data. The MCP calls are ephemeral; the AI receives the information for that specific interaction and does not retain it beyond the session. This design preserves Dun & Bradstreet’s strict data governance while giving end users a responsive, context-rich experience.
Inside Microsoft 365 Copilot
For the Windows community, the most tangible benefit appears inside Microsoft 365 Copilot, which is deeply embedded in the Office suite that runs on over a billion devices. The D&B integration surfaces through a new skill that Copilot can invoke. When a user types a natural language request—such as “Tell me about Contoso Ltd.’s revenue, employee count, and family tree”—Copilot detects the intent, calls the D&B MCP server, and displays the information with clickable citations back to the Dun & Bradstreet record.
• In Word: Sales teams can build customized proposals that auto-populate with live company overviews, risk ratings, and executive names pulled from D&B.
• In Excel: Finance departments can construct dynamic models that refresh supplier financial health scores each month. A single cell formula like =D&B.RISK(A2, “StressScore”) becomes possible through the integration.
• In Teams: During a video call about a potential partner, a participant can type a command into the chat pane and receive a one-page D&B snapshot of that partner’s global operations, visible to all meeting members.
• In Outlook: When drafting an email to a new vendor, Copilot can suggest an opening line that references the vendor’s most recent D&B-reported revenue, helping establish context without manual research.
Microsoft has positioned Copilot as a reasoning layer across the Microsoft Graph. Adding Dun & Bradstreet extends that graph from internal organizational data to the wider business universe. It’s a logical expansion for the “Copilot ecosystem,” which already supports plug-ins from Jira, SAP, and Adobe.
ChatGPT and Claude Get Business-Ready
OpenAI’s ChatGPT Enterprise and Team plans, as well as Anthropic’s Claude Enterprise, also gained the D&B connector on June 16. Both platforms have embraced MCP as a way to differentiate from consumer-grade chatbots. For ChatGPT, the D&B integration appears as a connected app within the platform’s app store. Administrators enable it for their workspace, set data tier permissions, and then users invoke it by mentioning the specific company in a prompt.
Claude, known for its long-context capabilities, benefits particularly from the integration because users can now feed in a list of hundreds of companies and ask Claude to, for example, rank them by Dun & Bradstreet’s Supplier Evaluation Risk (SER) rating. The AI maintains the entire list in memory, calls the MCP server for each entity, and produces a scored table in seconds.
Both platforms enforce enterprise-grade security. All data in transit is encrypted, and the MCP calls never expose raw company data to the AI provider—only the specific user who is authenticated against Dun & Bradstreet’s entitlement system gets the result. This architecture addresses the concern many compliance teams have about proprietary information leaking into public models.
Real-World Scenarios
Consider a multinational manufacturer re-evaluating its supplier base after a natural disaster. A supply chain manager opens Excel, imports a list of 300 suppliers, and runs a D&B-powered Copilot skill that checks each supplier’s financial viability score, ultimate parent, and disaster exposure index. Suppliers flagged as high risk are highlighted automatically, with a suggested action: “Request updated Certificate of Insurance” or “Activate alternate source.” The entire process, which previously took days of cross-referencing portals and spreadsheets, now completes in under ten minutes.
A private equity analyst using ChatGPT can similarly upload a target list and receive a comparative table with D-U-N-S Numbers, employee counts, and revenue ranges, then ask the model to identify outliers based on Dun & Bradstreet’s growth percentile. The AI generates the report with embedded citations, ready for a partner meeting.
In the legal sector, a lawyer drafting a due diligence memo in Word can highlight a client name and ask Copilot to produce a D&B business background report, including litigation indicators and liens. The document remains editable, and all data refreshes with a single button.
Security and Trust
Dun & Bradstreet built its reputation on data accuracy and trust, and the company has repeatedly emphasized that these integrations maintain the same data integrity and security standards required by global financial regulators. The MCP server is hosted within Dun & Bradstreet’s own infrastructure, so the company retains full control over data access and usage.
All requests are logged. All responses are watermarked with source timestamps. Users cannot accidentally expose premium data to unauthorized colleagues because entitlements are verified against the organization’s D&B contract at every call. For Microsoft 365 Copilot, the integration inherits the tenant’s existing compliance policies, including data residency, Microsoft Purview labels, and conditional access rules.
Anthropic’s Claude platform similarly supports SOC 2 Type II environments, and the D&B connector operates within the customer’s private cloud boundary when configured that way. OpenAI’s ChatGPT Enterprise already holds ISO 27001 certification, and the D&B integration passes through its compliance framework.
Why Hong Kong?
The press release distributed from Hong Kong underscores Dun & Bradstreet’s deepening commitment to the Asia-Pacific region. Hong Kong serves as a strategic hub for the company’s supply chain and trade data operations, and the simultaneous announcement with Chinese language materials suggests a concerted push to capture the fast-growing enterprise AI market in Greater China and Southeast Asia. For multinational corporations with dense supplier networks across the region, having D&B data inside Copilot, ChatGPT, and Claude can streamline cross-border compliance checks and local partner verification.
What This Means for Windows and Enterprise Users
For the Windows-focused IT professional, this announcement reinforces Microsoft’s vision of Copilot as an extensible, data-agnostic assistant. The operating system itself becomes a vessel for enterprise intelligence, where a user’s workflow—spreadsheet, document, email, or chat—is the only interface needed to access world-class business data. IT administrators will find the setup straightforward: enable the D&B connector in the Microsoft 365 admin center, map user licenses, and roll it out through existing group policies.
The integration also highlights a broader industry shift. The earlier wave of generative AI tools often hallucinated company information or pulled outdated web results. By directly connecting to authoritative, paid data sources via MCP, enterprises can finally use AI for business-critical decisions with a level of trust that was previously missing.
Competitors like ZoomInfo and LexisNexis will likely follow with similar integrations, but Dun & Bradstreet’s head start—and the sheer breadth of its Commercial Graph—gives it a significant moat. The company’s data already underpins millions of credit decisions and compliance checks daily; making it available inside the AI tools that knowledge workers already use could accelerate its adoption dramatically.
The Road Ahead
Dun & Bradstreet has indicated that this is only the beginning. The MCP server’s architecture allows for rapid, ongoing updates, and the company plans to expose more granular data products over time—such as beneficial ownership structures, environmental sustainability scores, and cyber risk ratings. As Copilot, ChatGPT, and Claude evolve from assistants to autonomous agents, having a trusted data backbone will become even more critical. An agent that schedules meetings or places orders on a user’s behalf must be able to verify the counterparty’s identity, financial standing, and regulatory status in real time. Dun & Bradstreet is positioning itself as that verifier.
For Windows enthusiasts and enterprise IT leaders, the message is clear: the AI tools you already use are now business-grade. The days of copying and pasting company data from a separate website are numbered. The Commercial Graph, representing decades of curated business intelligence, has arrived inside your flow of work.