Enterprise teams are running GenAI pilots, declaring success on demos, and struggling to move to production. The gap is almost never the model — it is the operational wrapper: observability, guardrails, cost controls, fallback logic, and the feedback loops that let the system improve without human intervention.

This session presents five GenAI operational patterns that Naval has seen work at enterprise scale: AI-assisted incident triage, LLM-backed runbook generation, cost anomaly narration, automated PR review and policy enforcement, and intelligent capacity forecasting. Each pattern includes the architecture, the prompt engineering approach, the failure mode to watch for, and a cost-per-call estimate. No demos of chatbots answering HR questions.

Level: Intermediate–Advanced | Tags: GenAI, AIOps, LLM, Platform Engineering, SecOps