Support queue breach predictor
Your support queue occasionally blows through its service-level targets before anyone notices, and the first you hear of it is an angry escalation. Design an agent that watches the queue and predicts breaches early enough to matter. What can it do on its own — reprioritize, reassign, page someone — and what must it never do? Describe the rollout that earns the team's trust step by step, and the evidence you'd show your VP that it's working.
decide_queue_action(phase: str, minutes_to_breach: int, minutes_since_alert: int, human_acked: bool) → str["shadow",10,30,false]out"log"["assist",40,5,false]out"reprioritize"["autonomous",10,30,false]out"page"Turn the fuzzy goal into a bounded agent task. Say what the agent reads, what it may do on its own versus draft for a human, and what it must never touch. Then earn the trust: how you’d verify it before rollout, and the pitch (the problem, what you built, why it matters).
Agent build: design the agent in plain language (or narrate it). The coach grades the decomposition, the guardrails, the verification plan, and the pitch.