AI Agents Still Need Heavy Human Supervision

A new Collibra-sponsored Harris Poll reveals that over 75% of decision-makers faced roadblocks moving AI agents from pilots to full production in the past year. More than half reported spending significant staff hours reviewing and correcting agent outputs before deployment. The survey, covering 306 U.S. data management, privacy, and related professionals, highlights a growing gap between the automation promise of AI agents and the real-world oversight they still demand.