Fortune reported on September 16 that CIOs are racing to put guardrails around AI agents that act outside their intended limits. A Fortune survey found that over half of decision-makers say they spend significant staffing hours manually reviewing and correcting autonomous agent outputs before they go live.
The finding undercuts the simple productivity story. If review time offsets automation gains, transformation programs have to budget for oversight, testing and rollback, not just licenses. It also helps explain the rush of funding to agent security startups reported this fall.
Source: Fortune