Businesses Pursue Autonomous IT but Keep Humans in the Approval Loop
A survey found that 90% of companies want to move toward autonomous IT operations within two years using agentic AI, but 77% are reluctant to let AI make operational decisions without human approval. Only 41% say their IT environment is ready for autonomous operations, and just 19% of IT operations are automated. Most organizations also lack unified visibility across networks, applications, endpoints, and cloud, and many report insufficient oversight of AI agent performance.
The survey shows a gap between ambition and readiness: while nine in ten organizations aim for self-directed IT within two years, only about two in five consider their infrastructure prepared, and roughly one in five IT tasks run automatically today. Nearly all respondents value unified visibility across networks, applications, endpoints and cloud, yet only 17 percent have achieved it, leaving data fragmented.
Technical staff are more cautious than executives. Specialists favor human supervision more strongly and express less confidence in governance and data quality, while leaders rate both higher. Just over half report inadequate oversight of AI agent performance and reliability, and many lack clarity on employee AI usage and costs.
If organizations keep people at the approval step, AI adoption in IT may proceed more slowly but with fewer unchecked errors. Employees on service desks could see their roles shift toward prevention and complex cases, while specialists' caution might temper timelines set by leadership. Customers and workers alike could benefit from more reliable systems, though uneven visibility and thin oversight may leave some incidents undetected. The pace of change may depend less on AI capability than on governance and data readiness.