Monitor ecommerce reviews with MrScraper, GPT-4o-mini, Slack and Notion

Description

This n8n template gives ecommerce brands a fully automated review intelligence system — running every morning to scrape, analyze, and report on what customers are actually saying across every platform. It uses MrScraper to collect reviews from Tokopedia, Shopee, Lazada, Bukalapak, Amazon, and more, then GPT-4o-mini to extract 15 brand intelligence signals per review including sentiment, emotion, viral risk, competitor mentions, and CS response suggestions.

The result is a daily Brand Awareness Score (BAS) delivered to Slack, every review archived in Notion, and urgent alerts fired the moment a critical review is detected — before it goes viral.

---

How It Works

Phase 1 – Trigger & Config: A Schedule Trigger fires daily at 6AM. The workflow reads your product list from Google Sheets — each row is one product SKU across any platform — and loops through them one by one. Phase 2 – Review URL Discovery: For each product, the Map Agent crawls the product page and discovers review section URLs or paginated review pages. A smart fallback ensures reviews embedded directly on the product page (common on Tokopedia and Shopee) are still captured even when no separate review URL exists. Phase 3 – Review Extraction & Filtering: Each review URL is processed by the General Agent, which extracts the full review text, star rating, reviewer name, date, photo count, helpful votes, verified purchase status, and any existing seller reply. Short reviews under 10 words are skipped — unless the rating is 1 or 2 stars, where even brief negative feedback is treated as a valuable signal. A deduplication hash is generated per review to prevent double-processing on re-runs. Phase 4 – AI Brand Sentiment Analysis: Every valid review is sent to GPT-4o-mini with a structured prompt that returns 15 brand intelligence fields: sentiment label and score, emotion tags (frustration, delight, anger, loyalty, etc.), the most impacted product dimension (quality, delivery, packaging, pricing, authenticity), a CX score out of 10, competitor brand mentions, viral risk assessment, urgency level, and a ready-to-use customer service response suggestion. Phase 5 – Storage, Alerts & Daily Digest: Reviews flagged as action-required trigger an immediate Slack alert to your brand-alerts channel. Every review is saved to Notion with all 27 metadata fields. At the end of each run, a Daily Brand Health Digest is compiled — including the Brand Awareness Score (BAS) out of 100, sentiment breakdown, top praises and complaints, emotion trends, competitor mentions, viral risk list, and action items — then posted to your brand-monitoring Slack channel.

---

How to Set Up