Protecting sensitive business data in preparation for the organization's Gen AI

This report analyzes the data-protection challenges posed by enterprise generative AI (e.g., Microsoft Copilot, Google Gemini), arguing that traditional, infrastructure-focused controls fail to address the risks of unstructured sensitive data. It emphasizes the need for automated inventorying, data-centric defenses, integrity verification, and visibility into who can access and how answers derived from data are used, using Controlled Unclassified Information (CUI) as an illustrative example.