PagerDuty, known for orchestrating incident response and alerting workflows, is moving towards autonomous operations with the announcement of its AI Agent Suite. The new generation of tools aims to shift PagerDuty beyond being an orchestration layer toward being an active partner in diagnosing, remediating, and optimizing operations. The change builds on a foundation of machine learning and automation that PagerDuty has been layering into its platform for years.
PagerDuty’s roots lie in helping organizations detect and respond to disruptions across complex systems. Customers use it to centralize alerts, coordinate on-call rotations, escalate issues, and marshal response teams. As Jennifer Tejada, PagerDuty’s CEO, put it in an interview with Knowledge at Wharton, “You become very reliant on the convenience or the 24/7 on-demand availability of something … when your brand experience isn’t available … you don’t blame the technology, you blame the brand.” That pressure of seamless uptime is core to PagerDuty’s mission.
Over time PagerDuty added features such as AIOps to reduce alert noise, correlate incidents, triage events, and automate parts of the response. Its AIOps offering claims reductions of up to 87 percent in alert noise and accelerates resolution by embedding intelligence into workflows. Its generative AI and embedded intelligence functions, branded under “PagerDuty Advance,” already support tasks like drafting status updates, summarizing diagnostics, and automating repeatable steps..
PagerDuty’s agents may classify issues, surface relevant historical context, propose or execute remediation, and continuously learn from operational data. Among the initial agents are the SRE Agent (to support site reliability workflows), the Insights Agent (to analyze operational patterns), and a Shift Agent (to optimize on-call coverage). The SRE Agent, for example, can find related incidents from the past, guide responders toward root causes, and propose next steps. The Shift Agent addresses another perennial pain point: scheduling conflicts and shift handoffs. The Insights Agent brings cross-tool data into a unified analysis that surfaces trends and recommendations.
PagerDuty emphasizes safety, guardrails, and context-awareness as essential. The agents are intended to operate under constraints, escalate uncertain cases to humans, and make recommendations (rather than fully automated moves) in more complex scenarios. PagerDuty claims that its agent designs draw on over 15 years’ worth of operational data from billions of events, giving the system a contextual advantage. In its “AI-First Operations” framework, PagerDuty outlines a path for enterprises to evolve: evaluating where AI should be applied, matching human involvement to issue complexity, embedding governance, measuring impact, and iterating continuously.
In the Wharton interview, Tejada spoke to culture and readiness more than technical mechanisms, stressing an “ownership mindset” and agility to adapt. She argued that organizations must prepare not just technologically, but culturally, to integrate AI deeply into their operations. The shift is about shifting how teams work, how they trust AI, and how workflows are governed.
In the broader market, PagerDuty is not alone in pressing forward with AI in incident and operations. In the AIOps and observability space, companies like BigPanda, Moogsoft, and Rootly are active competitors. BigPanda is often cited as a peer for correlation and noise reduction, while Moogsoft has long been known for anomaly detection and event correlation (though lacking PagerDuty’s strength in on-call orchestration). Rootly, in contrast, positions itself as AI-native for incident response, offering features like incident summaries, recommended responders, and postmortem drafts. Some newer entrants, like Ciroos (a startup launched in 2025), are building multi-agent AI systems specifically to support SRE workflows, aiming to rival such suites.
Splunk, now under Cisco, also competes in observability and security domains; its capabilities for log analytics, correlation, and alerting overlap with parts of what PagerDuty aims to extend. Cisco’s acquisition of Splunk underscores the belief that data, security, and observability will be foundational in AI-driven operations. The consolidation signals intensified competition for any platform wanting to be central to mission-critical operations.
Where the field stands is uneven. Many AI and AIOps tools today excel in noise suppression, anomaly detection, or correlation, but fewer deliver autonomous agents with safe action capabilities. Academic research (for example, frameworks like AIOPSLAB) illustrates both the aspiration and difficulty of designing agentic systems that can handle complex operational tasks reliably. Some research prototypes such as “Nissist” show how co-pilot systems can reduce time to mitigate by extracting structured decisions from sets of manuals and past incidents. But production-grade deployments in live, heterogeneous systems remain rare and carefully guarded.
PagerDuty’s advantage is its installed base, deep integration into operational workflows, and existing AI and automation capabilities. Its bet is that many enterprises will prefer to embed agents into a known, trusted platform rather than bolt new agentic systems across stacks. The Agent Suite is a stepping stone to making PagerDuty a proactive operational partner.
With AI agents now generally available (or in early access), the pressure will be on PagerDuty to demonstrate reliability, safety, measurable gains, and cultural acceptance. The company’s path may mirror its earlier transitions: from alerts to correlation to embedding intelligence, but the stakes are higher: failures, missteps, or unpredictable behavior in autonomous agents could erode trust faster than they build it.
