Based on a 2026 survey of over 400 governance, risk, and compliance (GRC) and audit leaders, as well as qualitative interviews with members of Optro’s Customer Advisory Board, the report highlights the growing disconnect between the reality of AI deployment and operational planning and readiness. While 96 percent of leaders agree that workflows should adapt to leverage AI capabilities, only 9 percent have redesigned their workflows to safely accommodate these agents.
“Enterprise AI agents are now embedded directly into workflows with real business, compliance, and regulatory consequences,” said Guru Sethupathy, GM of AI Governance at Optro. “This report underscores the need for a disciplined operating model where agent authority matches an organization’s ability to govern it.”
According to the report, enterprises are extending significant authority to AI agents, with 38 percent authorizing agents to approve or reject transactions, 34 percent allowing them to modify controls or policies, and 27 percent permitting changes to user permissions or access rights.
When agents can act without proper safeguards, the stakes escalate rapidly. In addition to the finding that over one-third of organizations have acted on inaccurate agent decisions, the report reveals further agentic governance and operational failures:
- 30 percent have seen an AI agent take an unintended action.
- 25 percent have experienced a control failure involving an agent.
- 46 percent report that employees are actually spending more time reviewing, validating, or correcting AI outputs.
This decision-making gap is escalating at a time when risk is evolving faster than ever. As AI, cyber, operational, and regulatory threats continue to collide, traditional GRC architectures rely on static, backward-looking systems of record, leaving organizations up to three months out of date and vulnerable to compounding failures. Optro addresses this gap through its GRC Intelligence Platform, uniting a trusted system of record with an intelligent, agentic system of action.








