Dun & Bradstreet has spent more than 180 years developing a detailed commercial database known as the Commercial Graph, which includes information on 642 million businesses and their interconnections. Originally designed for human analysts, the system struggled to meet the demands of AI agents used in modern credit, procurement, and supply chain processes. The original database was fragmented and static, making it unsuitable for AI’s need for dynamic, real-time data access and relationships.
To address this, D&B consolidated its various systems into a unified, cloud-based knowledge graph with continuous AI-driven updates. They developed a new access layer tailored for AI agents, complete with entity resolution to ensure accuracy. Recognizing that AI agents require distinct authentication, D&B created a new “Know Your Agent” system to verify machine identities securely.
This transformation highlights four critical lessons for enterprises implementing AI: prioritize solid data foundations, design for changing relationships, enforce entity consistency in multi-agent workflows, and embed data lineage from the start to ensure traceability and confidence in the results.
