Triage training survey comments with Google Sheets, TypeSafe Jev and Slack

Quick overview

This workflow triages new training survey comments from Google Sheets using TypeSafe AI (Jev) to score sentiment, urgency, and meaningfulness, posts urgent issues to Slack, and writes a routed record for every response back to a results tab to avoid double-processing.

How it works

1. Starts from a manual run, a form button trigger, or a daily schedule at 9:00.
2. Reads survey responses from a Google Sheets “responses” tab and previously processed items from a “results” tab, then keeps only responses with new response_id values.
3. Skips comments shorter than the configured minimum length and routes them as “Skipped.”
4. Sends each remaining comment to TypeSafe AI (Jev) to classify sentiment and score urgency and whether the feedback is meaningful.
5. Routes each response as “Urgent,” “Needs Review,” or “Normal” based on the configured thresholds and a rating-vs-sentiment mismatch check.
6. Posts “Urgent” comments to a specified Slack channel and appends or updates the routed record in the Google Sheets “results” tab.

Setup

1. Connect Google Sheets OAuth credentials and set the spreadsheet ID plus the “responses” and “results” sheet names in the workflow settings.
2. Ensure your Google Sheet has the required columns in the responses tab (response_id, course, rating, comment) and matching columns in the results tab (including route, sentiment, scores, and processed_at).
3. Add a TypeSafe AI credential (API key) and confirm the Jev model ID used in settings (for example, “jev-latest”).
4. Connect Slack OAuth credentials and set the target Slack channel ID for urgent alerts.
5. Review and adjust the minimum comment length and the urgency, sentiment-confidence, and meaningfulness thresholds to fit your triage criteria.