Enkrypt AI Data Risk Audit evaluates datasets used in AI systems before deployment to identify security, privacy, and governance risks. The tool produces sign-off packets for security and legal teams that document what data the AI system uses, associated risks, and approval status. The platform generates a Data Bill of Materials that tracks every dataset and source feeding training, fine-tuning, and RAG indexing. It creates a risk register with ranked issues including severity levels, due dates, and remediation guidance. The tool maintains an approved/blocked registry with explicit permissions based on intended use cases such as TRAIN versus RAG. The audit covers security and privacy risks including PII, PHI, PCI detection, cross-tenant data mixing, prompt injection attacks, and data poisoning. It evaluates governance controls by checking conflicts with AI governance policies and generating evidence for security, legal, and audit requirements. The tool supports intent-based approvals where the same dataset can be blocked for training but approved for RAG with conditions. It provides role-based review workflows for security, legal, product, and ML teams with versioned approvals that include reviewer information and timestamps for audit trails. The platform supports multimodal and multilingual analysis across text, audio, and images. It can be triggered to re-run audits when datasets, connectors, or regions change, treating data updates as release gates. Artifacts are exportable in PDF, CSV, and JSON formats for internal sharing and governance documentation.