PagerDuty Paige

• Adding Paige to the first escalation level of a payments service so it has checked recent deploys and past incidents before the on-call engineer acknowledges
• Asking Paige in a Slack incident channel to search Datadog or Splunk logs for the error in the alert and summarize what changed
• Running the rollback incident workflow Paige recommends, after reviewing its reasoning, from the incident details page
• Uploading a service runbook so Paige's future triage on that service follows your team's own procedures
• Generating a playbook after a recurring incident is resolved and saving the approved version to the team knowledge base
• Redacting outdated or sensitive details from a service's Paige memory through the Memory API

Paige's advantage is that it sits where the page already lands. Teams on PagerDuty can put it on an escalation level and let it start investigating before anyone acknowledges, without wiring a separate AI SRE into their alert flow. The docs are candid that much of the agentic surface is still maturing: the escalation-policy trigger, Teams, and skills are Early Access, per-team agent profiles are only planned, and remediation means recommending incident workflows you have already built rather than writing new fixes. That makes Paige most valuable for teams with good runbooks and automation in PagerDuty. A practical evaluation is to connect your main log and metrics tools on a few noisy services, compare Paige's root-cause suggestions against recent incidents, and watch how quickly AI Actions are used before turning it on broadly.

About PagerDuty Paige

Paige is PagerDuty's SRE agent, generally available since November 3, 2026, when PagerDuty also renamed its flagship product (formerly PagerDuty Operations Cloud) the PD Reliability Platform. It started as "SRE Agent" earlier in 2026. It lives inside PagerDuty, not alongside it. You talk to Paige from the incident details page, the PD Reliability Console, Slack, or Microsoft Teams (Early Access). You can also put it on call: a toggle on an escalation level, or an incident workflow template, starts Paige investigating as soon as an incident triggers, in parallel with the paged human rather than instead of them. Both setups are labeled Early Access in the docs. During an investigation Paige reads the event payload and notes, recalls similar past incidents, checks change events, and queries connectors you allow. These cover logs, metrics, and traces in Datadog, Dynatrace, Grafana, New Relic, Splunk, Elasticsearch, Sentry, AWS CloudWatch, Azure Monitor, GCP, and others, plus runbooks in Confluence, Notion, ServiceNow, and GitHub, with GitHub code search and a revert-PR tool. Each connector tool is set to Always Allow or Blocked, with read-only and write/delete tools grouped separately. Skills (Early Access) add custom instructions and scripts. For remediation, Paige ranks your existing incident workflows, explains its choice, and puts a Run recommended workflow button in front of the responder. PagerDuty says Paige can act autonomously in well-understood scenarios, but the public docs describe the approval path, and agent profiles that scope which integrations and actions Paige can use per team are still only planned for Early Access. After an incident is resolved, Paige saves what it learned to per-service memory, which you can view, edit, or redact through a Memory API, and it can draft playbooks and update runbooks. It coordinates PagerDuty's Shift, Scribe, and Insights agents, which handle schedules, call transcription, and operational insights. On pricing, Paige is included in every PD Reliability Platform plan, from Starter at $2,800 a year to custom Ultimate. It is an add-on (PagerDuty Advance) on the Professional incident management plan and not available on Free. Usage draws on a plan's AI Actions allowance, and each request or nudge click uses four. Paige is a natural first AI SRE to evaluate for teams already running PagerDuty, but how useful it is depends on the connectors, runbooks, and workflows you already have.