Detect fake product reviews with OpenAI and send alerts to Slack via Airtable

AI-Powered Fake Review Detection Workflow Using n8n & Airtable

This workflow automates the detection of potentially fake or manipulated product reviews using n8n, Airtable, OpenAI and Slack. It fetches reviews for a given product, standardizes the data, generates a unique hash to avoid duplicates, analyzes each review using an AI model, updates the record in Airtable and alerts the moderation team if the review appears suspicious.

Quick Implementation Steps

1. Add Airtable, OpenAI and Slack credentials to n8n. 2. Create an Airtable Base with a reviews table. 3. Connect the Webhook URL to your scraper or send sample JSON via Postman. 4. Test the workflow by passing product and review URLs. 5. Activate the workflow for continuous automated review screening.

What It Does

This workflow provides an automated pipeline to analyze product reviews and determine whether they may be fake or manipulated. It begins with a webhook that accepts product information and a scraper API URL. Using this information, the workflow fetches associated reviews.

Each review is then expanded into separate items and normalized to maintain a consistent structure. The workflow generates a hash for deduplication, preventing multiple entries of the same review. New reviews are stored in Airtable and subsequently analyzed by OpenAI. The resulting risk score, explanation and classification are saved back into Airtable.

If a review's score exceeds a predefined threshold, a structured Slack alert is sent to the moderation team. This ensures that high-risk reviews are escalated promptly while low-risk reviews are simply stored for recordkeeping.