Skip to content
AI.info

The Pulse

Dynatrace survey finds AI ambition outpacing operational proof

A Dynatrace survey finds that organizations are targeting more autonomous operations while many still rely on people to verify AI decisions. Respondents also report gaps in monitoring agent behavior, business impact, costs and runtime security.

Dynatrace survey finds AI ambition outpacing operational proof

Dynatrace’s State of Observability 2026 research found a gap between organizations’ plans for autonomous operations and their current oversight of AI decisions. The company says the online survey covered 1,202 technology leaders at enterprises in 12 countries and was conducted in August 2026.

Autonomy targets outrun checks

The survey found that 87% of respondents’ organizations use AI agents for observability and IT operations. While 44% target fully autonomous operations within five years, 12% say their organizations have reached that level today, according to Dynatrace.

Respondents said humans verify an average of 66% of AI-powered decisions, while only 27% currently address visibility into how agents decide and act. Across nine AI observability capabilities, execution trails perceived importance by an average of 14 percentage points, the company reported.

Reliability beyond uptime

Dynatrace’s report distinguishes whether an AI system is available and responsive from whether it is producing results as intended. AI workloads can meet conventional availability, latency and error-rate measures while still producing inaccurate, unsafe, costly or otherwise unintended results, the company said.

Business impact is one area where the survey found a gap: 39% of respondents consider measuring it important, while 24% said their organizations currently address it. Dynatrace says assessing AI requires connecting its behavior with the applications, infrastructure, security systems and services around it.

Costs and runtime risks

Respondents also reported rising resource demands from AI workloads. Eighty-one percent said those workloads had increased compute and infrastructure spending, and 77% reported a greater need for cost-allocation visibility.

Among security-qualified respondents, 55% said their organizations had experienced a zero-day incident in the past year. Forty-five percent said their organizations continuously monitor runtime environments for vulnerabilities that are exploitable or actively targeted, according to the survey.

Standards and survey limits

On instrumentation, 77% of respondents said their organizations had moved OpenTelemetry beyond evaluation, but 20% had made it a strategic standard. Thirty-seven percent cited AI and large language model workload observability as a leading reason for adoption, Dynatrace reported.

The company says it commissioned the online survey of enterprise technology leaders. Unless otherwise stated, its results reflect respondents’ reports about their organizations; percentages may not total 100% because of rounding or multiple-response questions.

Source

Related stories

Explore

More articles