During performance reviews, before major launches, when SLA targets are at risk, or as part of growth preparation. Essential before scaling events or architectural changes affecting latency. Free template — copy and paste into ChatGPT, Claude, or Copilot.