Invoice Matching with Level of Confidence

Abstract

3-way matching is an important step within the purchase-to-pay process. In this step received invoices are matched to purchase orders and their corresponding goods receipts. Data on these document can be missing or incorrect, and corresponding values on a matching invoice and purchase order can vary. For this thesis we have researched methods for efficiently matching invoices with purchase orders, based on the data matching process, as well as determining a level of confidence for found matches. We have proposed an algorithm and implemented a proof-of-concept, which is tested using a dataset containing invoice and purchase order data.