AI-Powered Financial Email Parser & Data Entry Agent (Gemini/Groq to Sheets)

Unlock Automated Financial Data Entry with this AI Agent Template

This n8n workflow template acts as an AI Agent designed to liberate you from manual financial data entry. It diligently monitors specified Gmail labels for incoming invoices, payment confirmations, and bank statements. Using the power of Google Gemini and Groq Large Language Models, it intelligently parses email content—whether it's in the email body, HTML, or attached PDFs—to extract crucial financial data. The agent then structures this information according to a predefined schema and seamlessly logs it into your designated Google Sheet, ensuring your financial records are always up-to-date with minimal effort.

Key Features & Benefits

• AI-Powered Data Extraction: Leverages Google Gemini and Groq LLMs for accurate extraction of details like date, service, amount, category, currency, and payment method from diverse email formats.
• Gmail Integration: Automatically triggers on new emails in specified Gmail labels, downloading attachments for processing.
• PDF Content Extraction: Capable of extracting text from PDF attachments, including password-protected files (if password provided).
• Handles Various Email Structures: Differentiates and processes emails containing single transactions, multiple transactions within the body (e.g., daily summaries via HTML parsing), or standard invoices.
• Customizable Schemas: Define precisely what data to extract and how it should be structured using Langchain's Structured Output Parsers.
• Direct to Google Sheets: Automatically appends extracted and structured financial data to your Google Sheets for easy tracking, reporting, and analysis.
• Time-Saving Automation: Drastically reduces manual data entry, minimizes errors, and frees up valuable time for founders and finance teams.
• Versatile LLM Options: Provides paths for both Google Gemini and Groq, allowing flexibility based on your preference or specific needs.

• An n8n instance (Cloud or self-hosted).
• Gmail account with OAuth2 credentials configured in n8n.
• Google Gemini API Key (via Google AI Studio or Google Cloud Vertex AI, ensure it's compatible with the n8n-nodes-langchain.lmChatGoogleGemini node).
• Groq API Key.
• Google Sheets API credentials configured in n8n.
• (Optional) Default password for encrypted PDF attachments if consistently used.

1. Download the n8n workflow JSON file.
2. Import the workflow into your n8n instance.
3. Configure Gmail Triggers ('Get invoice', 'Get payment'):
   • Connect your Gmail OAuth2 credentials.
   • Update the 'Label Ids' in the 'Filters' section to match the Gmail labels you use for invoices and payments (e.g., 'invoices-to-process', 'payment-confirmations'). Ensure these labels exist in your Gmail.
   • Verify 'Download Attachments' is enabled.
4. Configure PDF Extraction ('Extract invoice', 'Extract payment'):
   • If your financial PDFs are commonly password-protected with a standard password, enter it in the 'Password' field. Otherwise, this can be left blank or handled dynamically if passwords vary.
5. Configure AI Models & Parsers:
   • Google Gemini Path: In the 'Google Gemini Chat Model1' node, select or create your Google Gemini API credentials. Review and customize the JSON schema in the connected 'Structured Output Parser1' node to define the exact fields you want to extract (date, service, amount, category, etc.) and their types.
   • Groq Path: In the 'Groq Chat Model' node, select or create your Groq API credentials. Review and customize the JSON schema in the connected 'Structured Output Parser' node similarly.
6. Customize Email Routing ('Switch' node):
   • Adjust the conditions in the 'Switch' node. The default rules route emails based on sender addresses (e.g., service@pxbillrc01.cathaybk.com.tw for multiple payments, regex for HSBC/Crypto.com/Fubon for single payments). Modify these to match the sender patterns of your financial emails.
7. Refine AI Prompts ('Extract details', 'Extract details1' LLM Chain nodes):
   • Edit the system prompts within these nodes. Tailor the instructions to the AI to improve extraction accuracy based on the typical content and format of your financial emails. The current prompts guide the AI to look for transaction details for bookkeeping.
8. Configure Google Sheets Nodes ('Send', 'Send1'):
   • Connect your Google Sheets OAuth2 credentials.
   • Select the 'Document ID' (your target Google Sheet) and the 'Sheet Name' where data should be appended.
   • In the 'Columns' section, map the output fields from the AI (e.g., {{ $json.output.date }}, {{ $json.output.amount }}) to your corresponding Google Sheet column headers. Ensure your sheet has these columns.
9. Test each path of the workflow with sample emails to ensure correct data extraction and routing.
10. Activate the workflow.