At a federal space agency, I led DevOps teams managing Kubernetes clusters with manual scaling decisions. Every Friday, the same SREs made the same scaling calls—burning budget and burning out. When budget pressure hit, two senior engineers were on the line for termination. The breakthrough wasn't technical. It was realizing that cost visibility and intelligent automation weren't luxuries—they were survival. By giving teams visibility into cost impact, automating repetitive scaling decisions, and building trust in the system, we cut infrastructure costs significantly. Those two engineers stayed. This talk is about the leadership moves that work: how to introduce automation without losing your team's trust, and why cost optimization can be the difference between keeping or losing your best people.