Autonomy boundaries in IT automation
You're building an internal IT-services automation in Python: ingest tickets from the ITSM API, classify them, auto-resolve a safe subset (password resets, license grants), and escalate the rest. You plan to use an AI agent across several steps. Decide which steps an AI agent should run autonomously, which need a human checkpoint, and which shouldn't be AI-driven at all — and justify the boundaries.
Implement
route_tickets(ticket_rows: list[str], allowed_actions: list[str], confidence_threshold_pct: int, auto_action_budget: int) → list[str]Examples
in
[["t1|password_reset|95|false","t2|license_grant|60|false","t3|vpn_access|99|true"],["password_reset","license_grant"],80,5]out["auto","review","escalate"]in
[["t1|password_reset|99|false","t1|password_reset|99|false","t4|delete_mailbox|99|false"],["password_reset"],80,5]out["auto","duplicate","escalate"]in
[["a|password_reset|99|false","b|password_reset|99|false","c|password_reset|99|false"],["password_reset"],50,2]out["auto","auto","review"]What a strong answer looks like
Treat the AI’s output as a draft to verify, not an answer to trust. Name the specific flaw and the input that triggers it, say how you’d catch it (tests, edge cases, reading critically), and how you’d re-prompt or decompose to get it right.
0:00 of about 18 min
Vibe & agentic: describe the solution in plain language (or narrate it) and the coach grades your approach.
Which questions mattered is sealed until you submit. Telling you now would just be handing over the edge cases.
Run or narrate your approach, then ask the coach.