91% believe they can rapidly contain compromised AI agents.

10% of organizations do not track AI-agent risk separately.

Organizations allocate an average of about 6% of security budgets to AI-agent risk.