Generate Personalized Promotion Emails with GPT-5 and Gmail Context Analysis

This n8n workflow automates the process of generating and sending personalized promotion emails. It leverages customer data from a Google Sheet, analyzes past email interactions from Gmail (or a local file for demonstration), and uses an OpenAI Chat Model to craft highly relevant promotional content. The workflow includes a human-in-the-loop step for approval before sending.

What it does

1. Triggers Manually: The workflow is initiated manually.
2. Loads Customer Data: Reads customer information (e.g., name, email, past purchases) from a specified Google Sheet.
3. Loops Through Customers: Processes each customer's data individually.
4. Retrieves Contextual Data:
   • Option A (Gmail): Searches the user's Gmail account for past email interactions with the customer to gather context.
   • Option B (Local File - for demonstration): If Gmail context retrieval fails or is not desired, it falls back to reading a predefined context from a local JSON file.
5. Analyzes Context: Extracts relevant information from the retrieved email context (or local file content).
6. Generates Personalized Promotion: Uses an OpenAI Chat Model (GPT-5, as per the directory name, though the JSON specifies OpenAI Chat Model) to generate a personalized promotional email draft based on the customer's data and the analyzed context.
7. Parses Output: Extracts the structured email content (subject, body, etc.) from the AI's response.
8. Human-in-the-Loop Approval: Presents the generated email draft for manual review and approval.
9. Sends Email: If approved, sends the personalized promotion email via Gmail to the customer.
10. Handles Rejections: If the email is not approved, it proceeds without sending.

Prerequisites/Requirements

• n8n Instance: A running n8n instance.
• Google Sheets Account: To store customer data.
  • A Google Sheets credential configured in n8n.
• Gmail Account: To search for past interactions and send emails.
  • A Gmail credential configured in n8n.
• OpenAI API Key: For the OpenAI Chat Model to generate email content.
  • An OpenAI credential configured in n8n.
• Local File (Optional/Demonstration): A local JSON file named customer-email-context.json containing sample email context, used if Gmail context retrieval is bypassed or fails.

Setup/Usage

1. Import the Workflow:
   • Download the provided JSON file.
   • In your n8n instance, go to "Workflows" and click "New".
   • Click the "Import from JSON" button and paste the workflow JSON or upload the file.
2. Configure Credentials:
   • Locate the "Google Sheets" node and select/create your Google Sheets credential.
   • Locate the "Gmail" node (for context retrieval) and select/create your Gmail credential.
   • Locate the "OpenAI Chat Model" node and select/create your OpenAI API key credential.
   • Locate the "Gmail" node (for sending emails) and select/create your Gmail credential.
3. Configure Google Sheets Node:
   • Specify the "Spreadsheet ID" and "Sheet Name" where your customer data is located.
4. Configure Gmail Context Retrieval (Optional):
   • The workflow is set up to attempt Gmail context retrieval first. Ensure your Gmail credential has the necessary permissions.
   • If you prefer to use the local file for context, you can disable the first "Gmail" node and ensure the "Read/Write Files from Disk" node is configured correctly.
5. Configure Local File (Optional):
   • If using the local file fallback, ensure a file named customer-email-context.json is accessible by the n8n instance at the specified path (e.g., /tmp/customer-email-context.json or adjust the path in the "Read/Write Files from Disk" node). This file should contain JSON data representing email context.
6. Activate the Workflow: Toggle the workflow to "Active" in the top right corner.
7. Execute Manually: Click "Execute Workflow" to run it. The workflow will process your customer data, generate emails, and pause at the "If" node (acting as the human-in-the-loop approval) for each email.
8. Approve/Reject Emails: When the workflow pauses, review the generated email content. You can then manually route the item through the "True" branch (to send) or "False" branch (to skip sending).