AI
Dynatrace shifts autonomous SRE from guesswork to real-time precision
New AI agents promise deterministic incident resolution by grounding actions in system facts rather than probabilistic models.
The New Stack
Observability vendor Dynatrace has expanded its AI-powered operations platform with a suite of autonomous agents designed to move beyond probabilistic analysis toward deterministic, real-time incident handling. The latest capabilities aim to bridge a critical gap in AI-driven operations: translating insights into safe, governed action.
The company, which introduced its Dynatrace Intelligence service earlier this year, is rolling out new autonomous agents for incident triage and remediation, a no-code agent builder, and expanded integrations with cloud platforms and incident management tools. The shift reflects a broader industry bet that 2027 could mark the arrival of genuinely unattended autonomous SRE functions.
From probabilistic guessing to deterministic facts#
The core distinction Dynatrace is emphasizing: instead of relying on statistical models that generate plausible-sounding suggestions, the new agents ground their actions in real-time, environment-specific system understanding. Each remediation step remains auditable and governed, addressing what IDC analyst Stephen Elliot calls "the gap between AI-generated insight and safe, governed execution"—the industry's single hardest challenge in autonomous operations.
Steve Tack, Dynatrace's chief product officer, describes the transition as gradual. Teams typically follow a "crawl-walk-run" arc: first validating that the platform accurately pinpoints root causes, then enabling human-in-the-loop remediation automation, and only eventually moving to fully autonomous action as confidence builds. Rather than removing humans, Tack argues the real metric is simple: "the percentage of incidents that never need a human at all."
The new agent lineup#
Three agents form the core expansion. The autonomous SRE agent detects problems and either flags them as part of an existing investigation or opens a new one, enriching it with additional context. A cloud SRE agent orchestrates fixes across AWS, Azure, and Google Cloud environments and centralizes findings into a single auditable record. Agent Builder allows teams to create custom AI agents without code—a recognition that each organization's operational workflows differ.
On top of these, an upgraded Dynatrace Assist now supports natural-language investigation and agent-ready workflows. Deeper integrations with ServiceNow, Atlassian, and PagerDuty extend autonomous functions into the tools teams already use daily.
