Extract marketing insights from Google Reviews using Dumpling AI + GPT-4

📄 What this workflow does

This workflow helps you analyze Google reviews of any business to generate powerful marketing insights. By simply submitting a business name and its Google Place ID, it fetches the top 30 reviews and uses GPT-4 (via LangChain Agent) to extract valuable customer insights such as marketing angles, customer motivations, product pain points, and voice of customer (VOC) quotes. The output is stored automatically in a connected Google Sheet.

👤 Who is this for

• Marketing teams looking for messaging inspiration
• Founders or product managers exploring customer feedback
• Brand strategists gathering real-world insights
• Agencies running VOC or sentiment analysis

🛠️ Requirements

• Dumpling AI API key
• OpenAI GPT-4 or GPT-4o access
• Google Sheets connection
• A form or manual input with:
  • Business Name
  • Google Place ID

⚙️ How to set up

1. Connect Credentials

   • Dumpling AI (via HTTP Header Auth)
   • OpenAI (GPT-4)
   • Google Sheets (OAuth2)

2. Prepare your Google Sheet

   • Create columns:
     Business Name, Place ID, Marketing Angles, Customer Motivations,
     Frictions and Barriers, Product Opportunities, VOC Snippets

3. Update Nodes

   • Replace the Google Sheets Document ID and Tab Name with yours
   • Check that the Dumpling API node is linked to your credential
   • Optional: tweak the prompt in the LangChain Agent node to fit your tone or goals

🤖 How it works (Workflow Steps)

1. User submits business name + Google Place ID
2. Dumpling AI fetches top 30 reviews
3. Workflow aggregates review text
4. GPT-4 via LangChain analyzes the reviews
5. Insights are parsed and logged to Google Sheets

💡 Customization Ideas

• Push output to Notion, Airtable, or Slack
• Add sentiment scoring to prioritize themes
• Create summaries for each insight category
• Schedule insights to be emailed weekly

This is a plug-and-play VOC research workflow — great for founders, marketers, and product teams who want actionable data from real customers without doing manual review scraping or summarizing.