AI workloads are putting new pressure on site reliability teams, even at organisations that already use observability tools. A Dynatrace survey of 919 IT leaders, reported by Diginomica, found that 97% of surveyed SRE teams use some form of observability, while the article describes wide variation in how deeply those tools are integrated.
Watch Desk analysis
What happened
Dynatrace regional vice-president Josh Clay told Diginomica that observability can be embedded as a control for new systems, or left as an underused dashboard. He said some AI teams reach production without adopting their organisation’s observability tools, then scramble to add monitoring when problems surface. His recommendation is to build observability in from the start.
The report also describes SRE responsibilities expanding to cover AI workloads without a formal change to job descriptions. Clay said one Dynatrace customer had created a dedicated AI SRE team; elsewhere, he said, the work is more commonly absorbed into existing roles. He also pointed to the difficulty organisations face measuring AI’s effects when they did not establish a business baseline beforehand.
Why it matters
Having a monitoring tool is not the same as having it woven into how teams build and operate systems. That gap matters as AI workloads reach production: teams may need to troubleshoot unfamiliar systems while also being asked to demonstrate their return on investment.
The findings are based on a survey reported by Diginomica, and Clay’s examples are practitioner accounts, not a universal measure of how SRE teams are staffed. Still, the combination of widespread tool use and uneven integration is a useful warning against counting installed dashboards as operational readiness.
Our read
The practical lesson is to make observability part of an AI project before it goes live, not a panicked retrofit after the first incident. And if AI work is becoming part of the SRE remit, organisations should be honest about the staffing and measurement work that comes with it. A dashboard cannot quietly become a team member.
What to watch
- Whether organisations make observability a standard part of AI projects from the start.
- Whether AI SRE becomes a dedicated role or remains added work for existing teams.
- How businesses establish baselines that let them assess AI’s effects on reliability, productivity and cost.
Discussion spark: Should AI observability become a formal responsibility with dedicated staffing, or is it reasonable to expect existing SRE teams to absorb the work?
Sources and evidence
- Want to hear more about AI adoption? (29 October 2026, 15:22 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.