Automated Invoice Payment Tracking with OCR, Claude AI, Slack & Notion

This workflow streamlines invoice and payment management by automatically processing uploaded invoices, receipts, or JSON files. It uses OCR and Claude AI to extract key details like invoice number, vendor, currency, totals, and payment amounts/dates. The system then canonicalizes data for de-duplication and matches against an existing Notion database, supporting primary matches on invoice number (with vendor/currency) and fallbacks on amounts/dates.

Based on matches, it intelligently decides actions: creating new unpaid/paid/partial invoices, updating existing ones, marking as paid, flagging for manual review, or archiving. Slack integrations provide one-click approvals with contextual messages showing prior payments, new totals, and buttons linking to webhook resumptions for actions like Approve Partial, Mark Paid, Manual Review, or Archive.

Updates are written to Notion DBs (Invoices, Receipts, Archived), editing only safe fields like Paid Amount, Status, and Last Payment Date, while formulas auto-recompute balances. Benefits include reduced manual data entry, error-proof tracking, quick approvals, duplicate prevention, and audit-ready logs. Ideal for finance teams, small businesses, or AP departments handling high-volume invoices to save hours weekly and minimize payment errors.

Use cases: Accounts payable automation, receipt matching for reimbursements, vendor payment verification, or compliance logging in no-code environments.