Likely control objectives

Keep AI-enabled work client-safe.

  • Define approved AI use for client work, internal research, documents, and knowledge management
  • Prevent client confidential information from leaking through prompts, uploads, retrieval, or generated output
  • Maintain human review, approval, and evidence for AI-assisted deliverables
  • Assess AI features inside SaaS tools used across client teams and delivery operations
Get help with AI security