Gartner predicts that AI will autonomously handle 25% of IT operations work by 2030, but warns that the transition will come with a sharp rise in critical system failures.
The shift toward autonomous decision-making is dramatic. By 2029, 60% of enterprises will have deployed agentic AI in their operations, and only 20% of AI-recommended actions will require human approval—down from 80% today. This reduction in human oversight accelerates deployment but also amplifies risk.
The risk is real and measurable. Organizations scaling up agentic systems in infrastructure and operations face a steep failure curve: 40% are projected to suffer a critical outage by 2028, compared to less than 1% in 2026. Early adoption of AI agents may actually worsen system stability before improvements materialize, creating a difficult transition period where enterprises must manage both increased automation and increased vulnerability.
