New Relic Unveils Three Pillars to Build Trust in AI-Driven Production

The problem: faster creation, harder operation

New Relic argues that AI helps engineering teams create software faster than ever, but that speed does not automatically mean more reliable production. As AI-generated code, applications and agents move into the digital estate, they bring new dependencies, operational blind spots and added complexity for the teams who keep systems running, the company says.

The company framed the event around three challenges that stand between the speed of AI and confidence in production. Platform engineering teams need to see everything they run. AI needs additional ways to be proven in production. Incidents need answers grounded in evidence.

The shared foundation, in New Relic's description, brings together telemetry, dependencies, ownership, organizational knowledge and business context. The aim is to give both people and AI stronger evidence to understand what is happening, why it matters and what should happen next.

AI Observability and Autonomous Operations

The AI Observability pillar covers the AI lifecycle from development to production. New Relic lists four capabilities:

• AI Coding Observability: shows how AI coding assistants are used across engineering teams, including adoption, usage, cost and potential security risks.
• AI Evaluation: assesses AI application responses for quality, relevance and performance so teams can find issues before users are affected.
• AI Agent Monitoring: gives visibility into agent behavior, tool calls, handoffs and workflows.
• AI Observability for OpenTelemetry: uses OpenTelemetry GenAI telemetry to monitor AI applications and agents across models and tools.

For Autonomous Operations, the company points to the gap that follows an alert. Alerts rarely explain what changed, which parts of the environment are affected or what to do next. Three offerings target that gap. New Relic Ground Truth and New Relic Ground Truth CLI bring operational context, including telemetry, changes, dependencies and organizational knowledge, into the developer's terminal. New Relic Autopilot investigates operational issues, evaluates the available evidence and helps teams understand what happened, why it matters and where to investigate next.

New Relic stresses that people remain in control of the response. It describes a longer-term vision in which, as trust, governance and AI capabilities mature, teams can progressively coordinate more operational work while keeping human oversight.

Platform Engineering: live catalogs and infrastructure maps

The third pillar addresses platform engineering teams, which New Relic says are expected to help developers ship quickly and reliably while handling fragmented infrastructure, growing observability costs, disconnected inventories, governance requirements and an expanding AI footprint. Many still rely on spreadsheets, CMDBs and custom portals to answer basic questions about what exists and who owns it.

New Relic offers two connected parts. Service Architecture Intelligence is described as an internal developer platform included with New Relic at no additional cost. It is built from live telemetry and provides catalogs of services and their relationships, teams and ownership, and scorecards that turn standards such as tagging and alerting coverage into an automated maturity model.

Infrastructure 360 gives a unified view of infrastructure resources. According to the company, it discovers public cloud assets and flags missing coverage, maps app-to-infrastructure dependencies across Kubernetes, networks, load balancers and application gateways, and tracks configuration changes so teams can correlate them with performance issues.

What it means for technology leaders

New Relic says the three pillars answer one question for leaders deciding how to run AI in production: can teams trust what AI helps them build, run and operate? Platform Engineering shows what exists, how it connects, who owns it and how it is changing. AI Observability is meant to prove how AI behaves from development through production, including what it costs.

For developers, the company expects less time spent searching for context and navigating internal tooling. For engineering leaders, it promises a measurable view of excellence, benchmarked against their own goals and their industry. These are the company's own claims, and the source does not provide pricing for most of the offerings or performance data.