AI agents are set to take over root cause analysis at most enterprises by 2028

For years, software engineers have shouldered the responsibility of sifting through massive volumes of logs, monitoring systems, and application data to identify what's broken and why. This reactive, manual approach demands extensive engineering resources and limits how quickly organizations can detect and fix problems.

A fundamental shift is underway. Generative AI and autonomous agents are moving enterprise IT observability from labor-intensive human investigation to machine-driven diagnosis and remediation.

The scale of AI adoption in observability#

According to Elastic's 2026 research on observability trends, 85% of organizations are already leveraging some form of generative AI for monitoring their systems. That adoption is expected to accelerate to 98% within two years. The transition represents more than just incremental improvement—it signals a move toward predictive and self-healing infrastructure, where machines catch problems before humans even notice them.

How agents change the game for IT teams#

Rather than human operators manually forming hypotheses and investigating each incident, AI agents can autonomously examine logs, traces, and metrics to understand what went wrong and suggest—or directly execute—fixes. "I see this as attention economics," says Thaddeus Walsh, Principal Solutions Architect at Elastic, in an interview with The New Stack. "If developers didn't have to worry about the underlying components that are necessary to ensure an application runs properly and could just focus on delivering functional capability, that would be perfect for them."

This shift benefits multiple roles within organizations:

• Site reliability engineers (SREs) and IT operations teams can move beyond reactive incident response toward understanding root causes
• DevOps and security teams gain deeper insights into system dependencies and failure chains
• Product teams can use AI tools to analyze A/B test data across regions and user segments
• Finance and compliance departments can automatically audit whether external service providers meet their obligations

Root cause analysis: where agents shine brightest#

Pinpointing why a system failed—not just restoring it to working order—has always been the harder problem. It requires mapping dependencies across complex, distributed applications, yet most organizations lack adequate documentation of these relationships. GenAI helps bridge this gap by correlating signals across logs, metrics, and traces to identify causality chains that would otherwise remain hidden.