New Relic introduced a no-code AI agent platform and expanded OpenTelemetry capabilities, signaling an aggressive push to make autonomous operations practical for enterprise observability teams. The launch targets a familiar pain point: turning floods of telemetry into faster, safer incident response without adding another layer of tooling sprawl.
A Purpose-Built Agentic Platform For Observability
The new Agentic Platform lets teams assemble and deploy observability-focused AI agents that can triage alerts, correlate anomalies across traces, logs, and metrics, and propose or even execute runbook steps when authorized. New Relic is positioning it as outcome-first rather than a general-use agent builder: think SRE copilots that isolate a bad deploy, generate a minimal repro, or draft a rollback plan tied to specific service-level objectives.
Enterprises can mix prebuilt agents with custom ones, centralize policy management, and maintain a single control plane for what agents can access across the observability stack. That design choice reflects where most companies are today—experimenting with autonomy but insisting on auditable actions, bounded scopes, and clear escalation paths.
Model Context Protocol Support and Ecosystem Integrations
The platform supports the Model Context Protocol, an emerging standard that connects models to external tools and data sources without hardwiring bespoke integrations. In practice, MCP lets New Relic’s agents securely tap into service catalogs, incident wikis, change logs, CI/CD metadata, and configuration stores, then act with context rather than relying solely on raw telemetry.
Because MCP emphasizes capability discovery and permissioning, teams can expose only the functions they want—query a deployment timeline, open a ticket, fetch a playbook—while keeping privileged operations gated behind approvals. That balance often determines whether pilot projects make it to production in regulated environments.
OpenTelemetry Moves From Sprawl To One Pane
Alongside the agent platform, New Relic is baking OpenTelemetry support directly into its application performance monitoring agents and data pipeline. The goal is to let teams ingest, enrich, and analyze OTel traces, metrics, and logs next to their existing telemetry without maintaining parallel pipelines or collectors.
For many organizations, getting value from OpenTelemetry has been less about instrumentation and more about operations—who runs the collectors, how upgrades roll out, and where data lands. New Relic’s approach effectively offers “fleet management” for OTel, reducing the overhead of running dozens or hundreds of collectors and cutting the risk of fragmented data paths that break correlation.
The timing aligns with industry momentum. The Cloud Native Computing Foundation highlights OpenTelemetry as one of its fastest-growing projects, and cloud providers such as AWS, Google Cloud, and Microsoft all support it in their observability offerings. Consolidating on a single pane helps teams preserve OTel’s vendor-neutral benefits while simplifying day-two operations.
Crowded Field Signals Real Enterprise Demand
New Relic’s move joins a wave of agent-platform releases aimed at de-risking enterprise adoption. Salesforce introduced Agentforce in 2024 to bring controllable agents into its CRM workflows. OpenAI followed with Frontier, a framework for building and governing task-oriented AI systems. Analyst firms have started to frame agent platforms as core infrastructure for scaling AI beyond chat-based assistants.
Enterprise buyers are pushing for clearly bounded autonomy, integrated observability, and cost controls. According to McKinsey’s 2024 State of AI report, 72% of organizations say they’ve adopted at least one generative AI capability, but productionizing these systems often stalls on governance and reliability concerns. Agent platforms embedded in operational tooling are an attempt to close that gap.
What This Means For SRE And Platform Teams
Short term, expect more “human-in-the-loop” patterns: agents that propose remediations, draft incident timelines, and automate noisy handoffs between monitoring, ticketing, and on-call tools. Done well, that can reduce mean time to detect and mean time to resolve while trimming alert fatigue.
Practical rollout steps look familiar. Start with read-only scopes and narrow, high-impact use cases like post-deploy regression detection or auto-generated runbooks. Define guardrails—role-based access, evidence logging, approval workflows—and measure outcomes such as MTTR, change failure rate, and cost per GB of telemetry. If those trends move in the right direction, expand privileges and surface more actions through MCP-exposed capabilities.
The strategic bet on agentic observability and OpenTelemetry
By combining an observability-first agent platform with managed OpenTelemetry, New Relic is betting that operational AI will stick where it is closest to the signals of software health. If the company can prove that agents cut toil without creating new risks—and that OTel can be run at scale without extra overhead—it will have a persuasive story for engineering leaders looking to modernize incident response and control telemetry spend.
The competitive stakes are real, with peers across APM and log analytics racing to add agentic workflows on top of standardized telemetry. For buyers, that competition is good news: the tools are converging on open standards, deeper governance, and measurable outcomes rather than AI for AI’s sake.
