New Relic Prepares Observability for AI Agents That Skip the Dashboard

New Relic announced on June 25, 2026, the introduction of New Relic Autopilot and New Relic Ground Truth, marking a significant advancement in how observability platforms are adapting to the rise of AI agents. These new offerings are positioned to support an operating model where artificial intelligence, rather than solely human engineers, takes the lead in investigating incidents and retrieving system context. This development follows New Relic's ongoing efforts to streamline enterprise adoption of OpenTelemetry, which has been instrumental in providing a less disruptive pathway towards open, mixed-mode observability. The core idea behind this innovation is that once telemetry data is standardized and properly governed, AI agents can effectively utilize it. Camden Swita, Head of AI at New Relic, noted that "Operations are going headless. AI agents won't log in to view dashboards. They'll pull what they need through APIs, reason about it, and act." This statement underscores a fundamental shift where observability is transforming into a data substrate that powers automated SRE tasks, including incident triage, comprehensive root-cause analysis, and AI-assisted remediation. New Relic Autopilot is designed as an automated Site Reliability Engineering (SRE) agent that can autonomously handle first-response SRE tasks. It leverages runbooks and operational history as foundational context, reducing the reliance on human review of dashboards. Complementing this, New Relic Ground Truth facilitates the exposure of standardized observability data to external AI agents and orchestrators, such as GitHub Copilot and AWS DevOps Guru. This is achieved through a Model Context Protocol, ensuring that telemetry is accessible to various AI tools without mandating New Relic as the primary agent. The efficacy of agentic observability heavily depends on the quality and governance of telemetry data. The article stresses that for AI agents to operate reliably, telemetry must adhere to strict quality standards, be well-governed, and function as a structured data foundation. This approach differentiates enterprise-grade platforms like New Relic from more general-purpose AI tools that might lack the necessary auditability and operational grounding required for critical SRE functions. The integration of OpenTelemetry is crucial in this context, as it provides the standardized instrumentation and healthier data pipelines that give AI systems a cleaner and more trusted basis for analysis and remediation support.