96% of IT Professionals Have Adopted AI, but Validation of Results Remains the Key Challenge

Nearly 96% of IT professionals are already working with AI. A new study reveals a scale that far exceeded expectations: virtually all IT specialists have implemented AI agents into their operations. This is not the hypothetical future we discussed at conferences two years ago — this is the current reality for most IT departments.

Seven Key Operational Tasks for AI

The study identified seven areas where AI agents are being deployed by IT professionals first:

• Automation of code review and static code analysis on pull requests
• Infrastructure monitoring, log analysis, and anomaly detection
• Automated testing and QA processes on critical paths
• Security threat detection and real-time incident response
• Deployment optimization, CI/CD pipelines, and version management
• Application performance analysis, database monitoring, and infrastructure analysis
• Migration planning and change management in systems

For each of these tasks, AI handles routine work—analyzing data volumes, finding patterns, generating recommendations. Engineers gain the ability to focus on architectural decisions and strategy instead of night shifts combing through logs.

Main Blocker: How to Actually Validate AI

But the scale of implementation revealed one acute, often underestimated problem. The biggest blocker in deploying AI to production is validation of results. How does an engineer know that AI's advice is correct before running the change on a live system? AI can be very convincing and appear highly competent. But AI can suggest an elegant solution that is critically wrong. An engineer who blindly applies AI advice could disable a crucial service. Monitoring built on AI recommendations could miss a real attack. A deployment based on AI suggestions could overload a database or cause a data leak.

The study emphasizes: the main obstacle to full AI adoption in IT operations is not technology. It is the ability to verify what AI produced.

"AI result validation is becoming a new critical competency for the IT

sector"—this is the conclusion drawn by the study's authors.

Humans in the Loop—Returned, but Differently

IT specialists are coming to understand that you cannot simply launch an AI agent and trust it to manage infrastructure without a human in the loop. A system of checks is needed, similar to how code review prevents errors from reaching the main branch. This means that IT departments are seeing growing demand for people who not only work with AI but critically validate it. This is a new competency—at the level of security audit, production-grade code review, or pre-deployment risk assessment.

What This Means

The study shows: AI in IT is no longer experimenting but solving real tasks and affecting production in 96% of companies. This is a huge leap over the past year. But this means IT departments urgently need people trained to validate AI solutions before implementing them. Without this skill—no acceleration, only risk. This is not a technical problem; it is a problem of competency and processes.