AI Agents Are Creating a New Cybersecurity Blind Spot

**Executive summary:** The briefing warns that autonomous AI agents deployed across identity systems, SaaS, cloud, and business processes create a new attack surface and governance problem by accelerating identity sprawl, excessive permissions, unmanaged tokens, and opaque runtime behaviour; organisations should prioritise an AI asset inventory, aggressive least-privilege, continuous runtime assurance, and updated governance to contain risk and restore provable assurance.