Report by Optro
Authority without oversight: The 2026 agentic enterprise report
Key Findings
34% of enterprises allow AI agents to modify controls or policies.
34% of organizations report they have experienced an inaccurate AI agent decision or output that was acted upon.
96% of GRC and audit leaders agree that workflows should adapt to leverage AI capabilities.
Only 9% of GRC and audit leaders have redesigned workflows to safely accommodate AI agents.
38% of enterprises authorize AI agents to approve or reject transactions.
30% of organizations have seen an AI agent take an unintended action.
25% of organizations have experienced a control failure involving an AI agent.
46% of organizations report employees are spending more time reviewing, validating, or correcting AI outputs.
27% of enterprises permit AI agents to change user permissions or access rights.