IDC Survey Finds Identity Control Gaps for AI Agents
Security Boulevard, Friday, July 17th, 2026
IDC survey of 539 organizations finds 90% say identity management must improve to handle agentic AI risk.
An IDC survey of 539 North American organizations documents growing trouble managing machine identities as AI agents multiply.
Some 90% of respondents believe their identity management practices need improvement to address agentic AI risk, and 58.7% say substantial changes or a complete overhaul are required.
Organizations now risk holding more machine identities than human accounts, yet only 26.7% have deployed dynamic role-based access controls that support AI systems.
Weak governance around credential verification, task-scoped access, and identity lifecycle management leaves agents over-permissioned or still active long after their task ends. The survey frames agent identity as an emerging major security exposure.