Observability as a foundation for AI-enabled platforms

AI can also improve how platforms are operated, but only when observability is treated as a core capability. Structured telemetry helps teams understand agent behavior, troubleshoot issues faster, and reduce mean time to recovery (MTTR).

This session explores why observability is critical for running AI workloads on internal developer platforms. Using OpenTelemetry and emerging GenAI semantic conventions as examples, we look at the tools and signals available today, the kinds of insights they enable, and how connecting these pieces can help platform teams evolve their internal developer platforms.