ManualTrigger Automate

For the ManualTrigger Automate workflow, automate the extraction and analysis of Trustpilot reviews to generate detailed customer insights. This workflow scrapes up to 1,500 reviews, clusters similar feedback using K-means, and summarizes key insights, including sentiment and suggested improvements. Results are seamlessly exported to Google Sheets for easy access and reporting, enabling businesses to understand customer opinions and enhance their services effectively.

7/8/2025
37 nodes
Complex
manualcomplexsplitoutlangchainfiltergooglesheetsexecuteworkflowexecuteworkflowtriggersticky noteadvancedapiintegration
Categories:
Complex WorkflowManual TriggeredData Processing & AnalysisBusiness Process Automation
Integrations:
SplitOutLangChainFilterGoogleSheetsExecuteWorkflowExecuteWorkflowTriggerSticky Note

Target Audience

  • Marketing Teams: Looking to analyze customer feedback and sentiment from Trustpilot reviews to enhance their products and services.
    - Data Analysts: Who need to extract, process, and visualize customer insights from large datasets.
    - Business Owners: Seeking to understand customer satisfaction and areas for improvement based on reviews.
    - Developers: Interested in integrating automated workflows with third-party APIs for data collection and analysis.

    • Problem Solved

    This workflow automates the process of gathering and analyzing customer reviews from Trustpilot. It addresses the challenge of manually sifting through large volumes of feedback by providing an efficient method to extract insights, sentiments, and suggested improvements. The results can help businesses make informed decisions to enhance customer satisfaction and drive growth.

    Workflow Steps

  • Step 1: Clear Existing Reviews - This step ensures that any previous records related to the company are removed from the Qdrant vector store to start fresh.
    - Step 2: Scrape TrustPilot for Company Reviews - The workflow collects reviews from TrustPilot, fetching the most recent pages to gather a comprehensive dataset.
    - Step 3: Store Reviews in Qdrant - Extracted reviews are stored in Qdrant, a vector database optimized for similarity searches.
    - Step 4: Trigger Insights SubWorkflow - A subworkflow is initiated to analyze the reviews based on the company ID and specified date range.
    - Step 5: Apply K-means Clustering Algorithm - Reviews are clustered to identify common themes and sentiments, allowing for better analysis of customer feedback.
    - Step 6: Fetch Reviews by Cluster - The workflow retrieves reviews grouped by their clusters for detailed insights.
    - Step 7: Generate Customer Insights - Using a language model, the workflow summarizes reviews, providing insights and identifying suggested improvements.
    - Step 8: Export Results to Google Sheets - Finally, the insights and relevant data are exported to a Google Sheet for easy access and reporting.

    • Customization Guide

  • Adjust Company ID: Modify the company ID in the Set Variables node to target a different business.
    - Change Review Date Range: Update the date range in the Set Variables1 node to fetch reviews from a specific period.
    - Modify Clustering Parameters: In the Apply K-means Clustering Algorithm node, adjust the number of clusters based on the volume of reviews for more tailored insights.
    - Customize Insights Generation: Alter the system prompt in the Customer Insights Agent node to refine how insights are summarized based on specific business needs.