PagerDuty launches Paige AI agent for site reliability engineering

What Paige brings to incident response

Paige is positioned as PagerDuty’s next-generation SRE agent, capable of acting as a "virtual responder" on call. It operates alongside human teams, with permissions configurable at a team level to ensure it only takes autonomous actions in low-risk, well-understood scenarios. According to PagerDuty, Paige’s recommendations include evidence and confidence levels, allowing teams to assess its suggestions before acting.

The company claims Paige outperforms leading AI coding agents in diagnosing simulated IT incidents. It can also integrate with collaboration platforms like Slack and Microsoft Teams, enabling end-to-end incident management within these tools.

Paige is built to address unique failure modes introduced by AI-generated software, such as model drift, token runaway, and data leaks. These issues, according to PagerDuty, require new approaches to reliability as AI adoption accelerates.

PD Reliability Platform and PD Labs

The renamed PD Reliability Platform, formerly PagerDuty Operations Cloud, now provides a unified console for incidents, on-call rotations, and AI agents. This consolidation aims to simplify workflows for operations teams managing complex systems.

PD Labs, a new initiative, offers early access to features still in development. Design partners in the program can test experimental tools like Topology (for system mapping), Deep Investigation (for root-cause analysis), and SRE Workbench (for reliability testing). Paige’s Deep Investigation feature allows the agent to test multiple root-cause hypotheses simultaneously while displaying its reasoning as new evidence emerges.

PagerDuty emphasized that PD Labs is built in collaboration with design partners, who receive direct access to engineers and early versions of features.

Pricing and adoption

The PD Reliability Platform is offered with a pricing model tied to operational scale, starting with an allotment of ingested events, automation actions, and user seats. Customers exceeding this allotment can purchase flexible credits without needing to forecast usage in advance.

Telus Corp., a Canadian telecommunications provider, has already integrated the platform into its internal incident management tool, signaling early adoption in enterprise environments.

How this connects to earlier reporting

PagerDuty’s expansion into AI-driven reliability tools aligns with broader trends in AI agent infrastructure, such as IBM and CoreWeave’s collaboration on workload controls for AI agents. While IBM and CoreWeave focused on foundational infrastructure like security and identity management, PagerDuty’s Paige targets the operational layer, specifically incident response and reliability.

The integration with Slack also intersects with earlier reporting on Slack’s post-Salesforce evolution. While Slack’s acquisition by Salesforce has faced challenges, its role as a platform for AI-driven workflows—like Paige—highlights its continued relevance in enterprise operations.

What this means

Lazyfounder analysis — our interpretation, not reported fact.

For founders and operators, Paige represents a shift in how AI can be embedded into reliability engineering. Unlike generic coding assistants, Paige is purpose-built for incident response, offering transparency in its reasoning and the ability to operate autonomously in constrained scenarios. This could reduce the cognitive load on engineering teams, particularly as AI-generated software introduces new failure modes that traditional monitoring tools may not catch.

The launch of PD Labs also signals a growing trend of companies involving customers directly in feature development. For startups, this model could provide a way to validate ideas earlier and build tighter feedback loops with users. However, it also requires a level of transparency and trust that may not be feasible for all products.

The pricing model, tied to operational scale with flexible credits, reflects an attempt to align costs with usage. This could be attractive for teams with variable workloads but may require careful monitoring to avoid unexpected expenses.

Related on Lazyfounder

• Slack
• IBM and CoreWeave co-design workload controls for AI agent infrastructure
• Slack co-founders detail operational and cultural clashes post-Salesforce acquisition