Machine Learning-Enhanced Automation for Invoice Reconciliation: OCR-Based Solutions for Accuracy and Efficiency in Logistics Industry

Invoice reconciliation, a pivotal process in logistics, involves verifying and comparing data on freight bills to ensure accuracy. Manual reconciliation is time-consuming and error-prone, taking up to 10 days. Leveraging OCR technology and automated reconciliation mechanisms is proposed to streamline this process. Challenges include varied invoice formats and field name discrepancies. A robust comparison algorithm is vital for an effective reconciliation engine. Literature review reveals innovative solutions in Regular Expression Pattern Matching, such as FREME, offering fast and scalable results. Optical Character Recognition studies, like OCRMiner, demonstrate the potential for automated invoice processing but face challenges in handling varying formats. Another approach, the Digitization Process, focuses on transforming invoice text into usable formats but lacks extensive discussion on handling diverse formats. This research aims to address these challenges by developing an OCR-based automated freight invoice reconciliation system. The study explores improvements in OCR accuracy, adaptable algorithms, and effective handling of diverse invoice layouts.