🤝 Predict customer churn with AI analysis of HubSpot and Google Sheets data

⚡ 1,294 views · 🤝 CRM & Sales Operations

Description

Who it’s for

Built for Customer Success and Account Management teams focused on proactive retention. This workflow helps you automatically identify at-risk customers – before they churn – by combining CRM, usage, and sentiment data into one actionable alert.

What it does

This end-to-end workflow continuously monitors customer health by consolidating data from HubSpot and Google Sheets. Here’s how it works:

• Fetch deals from HubSpot.
• Collect context — linked support tickets and feature usage from a Google Sheet.
• Run sentiment analysis on the tickets to generate a customer health score.
• Evaluate risk — an AI agent reviews deal age, sentiment score, and usage trends against predefined thresholds.
• Send alerts — if churn risk is detected, it automatically sends a clear, data-driven email to the responsible team member with next-step recommendations.

How to set it up

To get started, configure your credentials and parameters in the following nodes:

1. Credentials:
   • HubSpot: Connect your account (HubSpot: Get All Deals).
   • LLM Model: Add credentials for your preferred provider (Config: Set LLM for Agent & Chains).
   • Google Sheets: Connect your account (Tool: Get Feature Usage from Sheets).
   • Email: Set up your SMTP credentials (Email: Send Churn Alert).
2. Tool URLs:
   • In Tool: Calculate Sentiment Score, enter the Webhook URL from the Trigger: Receive Tickets for Scoring node within this same workflow.
   • In Tool: Get HubSpot Data, enter the Endpoint URL for your MCP HubSpot data workflow. (Note: This tool does call an external workflow).
3. Google Sheet:
   • In Tool: Get Feature Usage from Sheets, enter the Document ID for your own Google Sheet.
4. Email Details:
   • In Email: Send Churn Alert, change the From and To email addresses.

Requirements

• HubSpot account with Deals API access
• LLM provider account (e.g. OpenAI)
• Google Sheets tracking customer feature usage
• n8n with LangChain community nodes enabled
• A separate n8n workflow set up to act as an MCP endpoint for fetching HubSpot data (called by Tool: Get HubSpot Data).

How to customize it

Tailor this workflow to match your business logic:

• Scoring logic: Adjust the JavaScript in the Code: Convert Sentiment to Score node to redefine how customer scores are calculated.
• Alert thresholds: Update the prompt in the AI Chain: Analyze for Churn Risk node to fine-tune when alerts trigger (e.g. deal age, score cutoff, or usage drop).
• Data sources: Swap HubSpot or Google Sheets for your CRM or database of choice — like Salesforce or Airtable.

✅ Outcome: A proactive customer health monitoring system that surfaces risks before it’s too late — keeping your team focused on prevention, not firefighting.

🔗 Nodes Used

Send Email, Webhook, HubSpot, Markdown, AI Agent, Basic LLM Chain

📥 Import

Download workflow.json and import into n8n: Workflow menu → Import from File

📖 Importing guide · 🔑 Credential setup