Agentic AI is the most dramatic change to software development since the advent of cloud-native development and, going back farther, the introduction of higher-level programming languages. In this case study, we detail how Microsoft uses agentic processes to develop its Azure site reliability engineering (SRE) agent and show tech leaders how agentic software development is as much a process, skills, and governance shift as it is a tooling change. While Microsoft operates at a massive scale, its core lessons of human oversight, spec-driven guardrails, blended roles, and outcome-based metrics are practical patterns that even smaller teams in large enterprises can easily adopt.