AI-powered observability promises faster correlation, smarter triage, and automated root cause analysis, but only if the data underneath it is ready. In real production environments, AI agents face the same messy, high-stakes, interconnected problems that engineers wrestle with daily. A poorly structured Datadog deployment doesn't disappear when you add AI to the picture, it gets amplified. If your Datadog deployment is messy today, AI may help in some cases, but it can also make that mess louder.

The effectiveness of tools like Bits AI to correlate signals, surface root causes, and guide incident response depends entirely on the quality and coherence of the foundation they operate on: your infrastructure telemetry, APM instrumentation, log pipelines, and the relationships between them.

This session will draw on real customer experiences to walk through the most common missing prerequisites seen before AI enablement and what it looks like to get them right first. Attendees will leave with a practical framework for evaluating their own Datadog maturity and a clear path to getting the most out of AI-assisted troubleshooting.