Approve Employee Expenses: OpenAI + Google Sheets + Telegram

Instantly route employee expense submissions for approval using OpenAI, Google Sheets, and Telegram

This n8n workflow automatically receives expense submissions via webhook, uses OpenAI to clean and categorize them (including whether manager approval is required), logs everything to Google Sheets, and sends Telegram notifications to either your manager or the employee.

What this workflow does

• Receives expenses via Webhook: Accepts a POST request containing fields like employee_name, amount, vendor, description, currency (and optionally expense_id).
• Normalizes incoming data: Standardizes values such as timestamp, expense ID, employee name, amount, vendor, description, and currency for consistent processing.
• Analyzes with OpenAI: Sends the expense details to OpenAI to clean the amount, assign a category, and determine if manager approval is needed.
• Structures the final output: Parses the AI response and prepares the final fields for logging and notifications.
• Logs to Google Sheets: Appends each record to the “Expenses” sheet with an AUTO_APPROVED or PENDING_APPROVAL status (including fields like ai_reason).
• Notifies via Telegram:
  • Sends a message to the manager when approval is required.
  • Sends an auto-approval confirmation to the employee when no approval is needed.

Use cases

• Centralize travel, meals, and vendor receipts submitted from an internal expense form.
• Reduce manual triage by auto-categorizing expenses and flagging edge cases for review.
• Keep an audit trail in Google Sheets while using Telegram for real-time approvals.

Technical details

• n8n nodes/integrations: Webhook, OpenAI (chat model for expense analysis), Google Sheets (OAuth), Telegram, plus supporting nodes (e.g., if, set, sticky note).
• Setup requirements: Configure the webhook URL for incoming fields; add OpenAI credentials and select the chat model; configure Google Sheets OAuth and ensure an “Expenses” sheet with matching columns (e.g., expense_id, received_at, employee_name, amount, currency, vendor, category, description, status, ai_reason).