PagerDuty launches Paige AI agent for site reliability engineering

A new era for operations

PagerDuty has officially launched Paige, an artificial intelligence agent specifically designed for site reliability engineering (SRE). The company is positioning Paige as a functional member of engineering and operations teams rather than just a standalone tool. It is built to run on the company's flagship product, which has been renamed the PD Reliability Platform.

The launch addresses a growing challenge where AI coding tools are increasingly used to ship production software with minimal human oversight. PagerDuty notes that applications built on AI models can break in unique ways, such as experiencing model drift or leaked personal data. To combat these issues, the company aims to use AI to manage the very systems that AI is helping to build.

As part of this rollout, PagerDuty also opened PD Labs. This initiative allows customers to test early-stage features while they are still under development. It provides a direct line to the engineers writing the code, allowing for a collaborative approach to building new reliability tools.

How Paige handles incidents

Paige functions as a virtual responder that can be added to escalation policies. Depending on the permissions set by a team, the agent can either act autonomously in well-understood, low-risk scenarios or provide recommendations for human approval. When it suggests a fix, it attaches a confidence level and the evidence it used to reach that conclusion.

One of the primary goals of Paige is to rule out possible causes before a human responder even sees the alert. This allows teams to focus on high-priority issues while the agent handles the initial triage. By filtering out the noise, the tool helps responders avoid the tedious task of manually chasing through logs to find the source of a problem.

A significant part of the agent's intelligence comes from a shared memory system. Once an incident is resolved, the details are written into this memory, which Paige consults whenever a similar failure occurs. This means the agent becomes more effective at diagnosing recurring issues as it gains more experience with the specific infrastructure of a company.

Unified visibility and tools

The renamed PD Reliability Platform includes a new console for incident commanders and team managers. This dashboard puts incidents, on-call rotations, and AI agents into a single view. It allows managers to see exactly which agents are active and what specific actions they are performing in real time.

To streamline communication, responders can manage incidents directly within Slack or Microsoft Teams. When a trigger is activated, the platform opens a dedicated channel and a conference bridge automatically. This setup allows human and agentic teams to operate side-by-side in the same workspace, providing better visibility for everyone involved.

Automation also extends to the post-incident phase. The system automatically generates a post-incident review as soon as a problem is resolved. These reviews include follow-up actions that come with pre-assigned owners and due dates, removing the manual burden of documentation from the engineering team.

Previewing the future

Through PD Labs, PagerDuty is currently testing three specific experiments with design partners. One is a topology feature that creates a live map of an organization’s services based on code and infrastructure. This map automatically identifies the blast radius of a problem as conditions change without requiring a manual configuration database.

Another experiment, called Deep Investigation, allows Paige to test multiple root-cause hypotheses simultaneously. It shows its reasoning for each hypothesis as new evidence surfaces during the diagnostic process. A third tool, the SRE Workbench, provides a shared workspace where humans and agents can tackle a problem together while saving all findings to the shared memory.

The company is already seeing early adoption of these types of tools. For example, Telus Corp. has integrated the platform into its internal incident tools. This allows their responders to move through incidents without having to stop and manually document every step of the process.

Business and scalability

PagerDuty is offering the PD Reliability Platform and Paige as a fee-based service tied to operational scale. The pricing model includes a starting allotment of user seats, ingested events, and automation actions. This structure is designed to accommodate growing organizations that need to scale their incident response capabilities.

To provide flexibility, customers who exceed their initial limits can purchase extra credits. This allows companies to handle unexpected spikes in traffic or incidents without having to forecast their exact usage needs in advance. The goal is to provide a predictable way for IT teams to measure the return on their AI investments.

Industry analysts suggest that the value of these agents lies in measurable outcomes like better uptime. By handing back time to responders who would otherwise be assembling the right people for an incident, companies can maintain more stable systems. Paige is intended to be a core part of that strategy by automating the most repetitive parts of the reliability cycle.