Collect and theme customer complaints with Apify, OpenAI and Notion

Quick Overview

This weekly workflow uses Apify to collect public Reddit posts, OpenAI to extract and theme real product complaints, and Notion to store them as a structured complaints board with severity, a verbatim quote, and a “leaving” flag.

How it works

1. Runs every Monday morning on a schedule.
2. Searches Reddit via Apify using the configured query, subreddit, and maximum item count.
3. Discards posts that are too short to contain meaningful complaint content and skips items already processed in previous runs.
4. Sends each remaining post to OpenAI to decide if it is a complaint and to return a theme, verbatim quote, severity score, and whether the user is switching away.
5. Normalizes theme wording into consistent buckets and filters out complaints below the configured minimum severity.
6. Creates a new page in a Notion database for each kept complaint, storing the theme, quote, severity, source, and leaving flag.

Setup

1. Add credentials for Apify, OpenAI, and Notion.
2. Create a Notion database with properties matching Quote (rich text), Severity (number), Source (rich text), and Leaving (checkbox), then select its database ID in the Notion node.
3. Review and update the listening settings (product, query, optional subreddit, maxItems, and minSeverity) in the configuration step.
4. If you enable optional outputs, add Slack and/or Google Sheets credentials and set the target Slack channel, Google Sheet document ID, and sheet name.