HR Process Automation: A Case Study on Document Workflows at Amadori

This thesis addresses the problem of automating repetitive and time consuming HR processes, focusing on document workflows in large organizations. Manual management of HR documentation, such as medical certificates, often leads to inefficiencies, human errors, and limited scalability. This challenge is increasingly relevant in the context of digital transformation, where Business Process Automation (BPA) and Artificial Intelligence (AI) play a crucial role in improving accuracy and freeing employees from low-value administrative tasks. The main objective of this work is to design and implement an AI-driven solution to optimize the handling of medical certificates within the Payroll Office of Amadori, a leading Italian agri-food company. The study aims to reduce manual workload, increase data reliability, and integrate the automated process into the company’s existing digital ecosystem. The problem was tackled through a structured methodology: process mapping and quantitative assessment of current HR activities, feasibility analysis of automation opportunities, design of a middleware architecture and integration with Google Document AI for intelligent data extraction. The system was connected to the corporate Oracle Database and HR management platforms, ensuring end-to-end automation and compliance with data-protection standards. Experimental results showed that the Document AI model achieved 95.2% extraction precision, reducing the time required for certificate management by up to 100%, equivalent to a recovery of 0.36 FTE. Overall, the project demonstrates how AI-based document processing can transform HR operations by enhancing efficiency, reliability, and employee experience. The proposed framework provides a replicable model for intelligent automation across other administrative processes, supporting Amadori’s broader digital-transformation strategy.