AI Invoice Data Extraction Agent: Google Drive to Sheets using Mistral & Gemini

Unlock Automated Invoice Data Entry with this AI Agent

This AI Agent acts as your diligent accounts payable assistant. When you drop an invoice (PDF or image) into a designated Google Drive folder, it automatically springs into action. Using advanced OCR technology powered by Mistral AI, it reads the invoice content. Then, leveraging the intelligence of Google's Gemini model within an AI Agent setup, it smartly extracts key information—like invoice number, vendor, amount, due date—by dynamically matching it to the column headers you've set up in your Google Sheet. Finally, it appends this data as a new, neatly structured row in your sheet, ready for your accounting or record-keeping.

Key Features & Benefits

• AI-Powered OCR: Utilizes Mistral AI via its API to accurately digitize text from both PDF and image-based invoices, in multiple languages.
• Intelligent Data Extraction: Employs Google's Gemini model (via Langchain AI Agent node) to understand invoice content and extract relevant data fields.
• Schema-Agnostic: Dynamically adapts to the fields defined in Row 1 of your target Google Sheet. No need to pre-configure fixed templates for different invoice layouts.
• Seamless Google Workspace Integration: Monitors Google Drive for new invoices and writes structured data directly to Google Sheets.
• Eliminates Manual Entry: Frees up valuable time and reduces human error associated with manual data input.
• Versatile: Handles various invoice formats and is fully modifiable for power users to tweak or extend.

• An n8n instance (Cloud or self-hosted).
• Google Workspace account with Google Drive and Google Sheets access.
• Credentials for Google Drive and Google Sheets nodes in n8n.
• Mistral AI API Key.
• Google API Key with access to the Gemini API (for the Langchain Google Gemini Chat Model node).

1. Download the n8n workflow JSON file.
2. Import the workflow into your n8n instance.
3. Configure the 'Google Drive Trigger' node: Authenticate and specify the Google Drive folder to monitor for new invoices.
4. Configure the 'Mistral OCR' node (likely an 'HTTP Request' node): Enter your Mistral AI API Key in the authentication section and ensure the API endpoint and request body are correctly set up to send the file and receive OCR text.
5. Configure the 'Google Gemini Chat Model' node (connected to the 'AI Agent' node): Select or create your Google Gemini credentials using your API key.
6. Review and customize the prompt within the 'AI Agent' node or the connected 'Code' node (if used for prompt engineering). This prompt guides Gemini on how to extract data based on your Google Sheet headers.
7. Configure the 'Google Sheets' node: Authenticate, specify the Spreadsheet ID, and ensure the first row of your target sheet contains the desired column headers (e.g., Invoice ID, Vendor Name, Total Amount, Due Date).
8. Test the workflow with a sample invoice to ensure correct data extraction and population.
9. Activate the workflow for ongoing, automated invoice processing.