Evaluation of a low-code intelligent invoice processing prototype

Abstract

This paper investigates the development of a Robotic Process Automation (RPA)-based prototype for automated invoice data extraction and its potential implications for financial data analysis. We address the challenge of manual invoice processing by utilizing Microsoft’s Power Automate and AI Builder tools. The prototype employs Optical Character Recognition (OCR) and Natural Language Processing (NLP) to extract relevant financial data from invoices received via email and populate an Excel workbook. Achieving a 96% accuracy rate in data extraction is assessed in the light of technical reliability and economic feasibility within an integrated performance–cost framework. The limitations of the prototype are discussed, and avenues for future development are proposed, including integration with accounting software via eXtensible Markup Language (XML) conversion.