Triage JMA weather delays with GPT-5-mini, Microsoft Teams and Gmail

## Alert shipping teams to JMA weather warnings, approve delay notices and email affected customers Heavy rain, storms, and snow in Japan often delay pickups and deliveries. This workflow watches official Japan Meteorological Agency (JMA) warnings for the prefectures you ship from or to, finds the pending shipments that may be affected, and emails those customers after a staff member approves the notice. ### How it works 1. Every 30 minutes, it reads the public JMA warning feed (no API key) and the warnings in effect for each watched prefecture. 2. It compares them with the state saved in an n8n data table, so only changes continue. 3. Find Affected Shipments matches the prefecture against pending orders in the open_shipments table, including all orders when your warehouse is in that prefecture. 4. OpenAI rates severity and drafts an internal alert and a keigo customer email with placeholders. 5. Teams gets a card with the affected orders and a Review and send button. Staff edit and approve the email in an n8n form. If nobody answers in time, a reminder is posted, and the request expires after a second wait. 6. After approval, one Gmail message is sent per order. Each order is marked as notified or failed, and a summary card reports the result. Lifted warnings get an all-clear card. ### Setup steps 1. Create two n8n data tables: weather_alert_state (prefecture, warnings, reported_at) and open_shipments (order_id, customer_name, email, ship_to_prefecture, ship_date, status, notice_sent, notified_at). All columns are text. 2. Run Load Sample Shipments once to try it, then fill the table from your order system. 3. Edit Set Areas and Settings, including the Teams webhook URL. 4. Connect OpenAI and Gmail credentials.