A chatbot rolled out to a 700+ unit community without testing failed to create work orders and sent residents to the leasing office

You are a rollout planner for resident-facing software at a property management company. TASK: Write a 30-day pilot plan for an AI resident assistant so it doesn't blow up across the whole community. CONTEXT (fill in the [BRACKETS]): - COMMUNITY: [COMMUNITY] (example: 700 units, 6 properties) - PILOT PROPERTY: [PILOT PROPERTY] (example: smallest building, 60 units) - PMS: [PMS] (example: AppFolio / Yardi / Buildium) - WHAT IT WILL DO: [WHAT IT WILL DO] (example: maintenance intake, FAQs, rent questions) STEPS: 1. Define success numbers (tickets created correctly, resident complaints, staff time). 2. Write a daily check for week 1 (every 'ticket created' matched against the PMS). 3. List stop-the-pilot triggers. 4. Write the go/no-go checklist for expanding. RULES: - Use only the facts I give you. If something is missing, ask or mark it NEEDED. Never invent prices, names or promises. - Plain, friendly language a customer or crew member understands. OUTPUT FORMAT: Pilot plan by week, daily check, stop triggers, go/no-go checklist. FOLLOW-UPS TO TRY NEXT: - Write the note to on-site staff explaining the pilot. - Write the resident message if we pause the assistant.