Robotic Process Automation (RPA) has significantly transformed how companies handle repetitive tasks by automating routine, rule-based processes. However, traditional RPAs struggle with complex, unstructured tasks requiring decision-making and adaptability. This thesis explores the transition from RPA to Intelligent Process Automation (IPA), which combines RPA with advanced technologies like Artificial Intelligence and Machine Learning to address these limitations. The research focuses on developing and implementing an IPA solution for the Charging Station Operators’ invoicing process at EDP Comercial. Through a structured five-phase methodology, including a comprehensive literature review, detailed work plan, design and development, implementation, and evaluation, the study demonstrates the substantial benefits of IPA. The developed IPA system utilizes UiPath’s Robotic Enterprise Framework, Azure services, and a GPT-powered API for data extraction and processing. Evaluation results highlighted a 65.2% reduction in processing time, significant cost savings, and a notable increase in accuracy from 90% to 97%. The automation not only streamlined the invoice validation process but also allowed the billing team to focus on higher-value tasks, enhancing overall productivity and job satisfaction. Despite the promising outcomes, the project encountered limitations such as initial investment costs, ethical challenges, and integration with legacy systems. Future work should focus on expanding IPA to other processes, continuous improvement of AI models, enhancing data analytics capabilities, and exploring emerging technologies. This project underscores the transformative potential of IPA in optimizing business processes, aligning with global trends toward digital transformation and intelligent automation, and contributing to the development of smarter and more adaptive business environments.

Dissertation presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data Science