What's Actually Happening Here
Dun & Bradstreet, the 182-year-old commercial data giant, is doing something that sounds deceptively simple: rebuilding how its flagship database works from the ground up. The target? Making its Commercial Graph — a sprawling network of data covering 642 million businesses worldwide — readable, navigable, and actionable for AI agents rather than just human analysts.
This isn't a cosmetic update. The Commercial Graph contains not just company names and addresses but deep relational data: corporate hierarchies, ownership structures, risk profiles, payment histories, and business relationships that took nearly two centuries to compile. The problem is that all of it was architected for a human brain to interpret — not for a large language model or autonomous AI agent to parse and act upon in real time.
Why This Is Trending Right Now
The timing isn't accidental. Enterprise AI adoption has hit an inflection point in 2024 and into 2025, with companies deploying agentic AI systems — tools that don't just answer questions but autonomously execute multi-step business workflows. Think AI that doesn't just tell a credit analyst about a company's risk profile, but actually initiates the credit review, pulls related entity data, flags anomalies, and drafts a recommendation.
For that to work, the underlying data infrastructure needs to speak the same language as these agents. Most enterprise databases, including D&B's, don't. They were built for keyword searches, dropdown filters, and human-readable reports. Feeding that kind of structured-but-siloed data to an AI agent produces hallucinations, missed connections, and broken logic chains.
D&B's move is part of a broader industry reckoning: the data layer hasn't kept pace with the AI layer.
Key Details of the Rebuild
From Records to Relationships
Traditional commercial databases store information as discrete records. D&B's rebuild centers on making the relationships between those records machine-traversable. An AI agent investigating a supplier's risk exposure needs to instantly understand that a subsidiary in one country shares ultimate beneficial ownership with a sanctioned entity in another — a connection a human might spend days tracing manually.
The Role of the D-U-N-S Number
D&B's proprietary D-U-N-S numbering system, which assigns unique identifiers to business entities globally, becomes the backbone of this transition. By structuring graph data around these identifiers with AI-compatible APIs and semantic layers, the company is essentially creating a Rosetta Stone that lets AI agents cross-reference business entities at scale without human mediation.
API and Agentic Integration
The rebuild includes new API architectures specifically designed for agentic workflows — meaning AI systems can query, reason over, and return contextual business intelligence dynamically, rather than pulling static reports.
The Broader Impact
The implications touch multiple industries simultaneously. In financial services, AI-powered credit underwriting could become significantly more accurate when connected to a properly structured commercial knowledge graph. In procurement and supply chain, autonomous agents could monitor third-party risk across thousands of vendors in near real time. In sales intelligence, AI tools could map entire account ecosystems — identifying decision-makers, relationships, and timing signals — without manual research.
For D&B specifically, this is also a competitive survival move. Data providers that can't plug into the agentic AI ecosystem risk becoming irrelevant as companies build or buy AI-native alternatives. The company is essentially future-proofing a business model that dates back to 1841.
What to Expect Next
D&B's rebuild signals a shift that other legacy data infrastructure companies — think credit bureaus, government data aggregators, and industry-specific intelligence platforms — will be forced to follow. The companies that move fastest won't just sell data; they'll sell reasoning-ready intelligence. Expect partnerships between major data incumbents and AI platform providers to accelerate, as well as a new wave of "AI-ready" data certifications and standards emerging across the enterprise sector.
The real question isn't whether AI agents will run on commercial data — they already are. The question is whether the data they're running on is structured well enough to be trusted. D&B's gamble is that the answer to that question determines who controls the next chapter of enterprise intelligence, and they intend to write it themselves.