Robotic Process Automation and Artificial Intelligence at Peab: Next step towards Intelligent Automation

Sammanfattning

Robotic Process Automation (RPA) and Artificial Intelligence (AI) are two techniques that can beused when automating processes and, when combined, transitions to Intelligent Automation (IA).IA can bring a paradigm shift to complex automated business processes. With the assistance ofPeab’s RPA team, this thesis addresses the existing gap in effectively implementing IA within or-ganizational workflows and specifically Peab’s workflows. It evaluates the current state of RPA andAI integration at Peab and outlines strategic recommendations for future development. Throughsemi-structured interviews with Peab’s RPA developers, this study answers questions regardingchallenges, including the difficulty of processing unstructured data, managing system deviations,and the limited application of AI beyond specific use cases such as invoice validation. Furthermore,an experiment was conducted investigating the use of AI models to detect interactive elements inweb interfaces. The result indicated that while AI was sturdy and accurate in identifying elements,the accuracy in mapping the coordinates of these elements was heavily influenced by the inter-face design’s simplicity and consistency. The findings highlighted the potential for AI to improvedecision-making in automated systems emphasizing that proper usage of AI is crucial for achiev-ing better outcomes. However, they also point out the fact that more AI training, comprehensivevalidation protocols, and user-friendly system design are needed. This research presents practicalinsight for Peab’s RPA team to move from rule-based automation to a more cognitive, intelligentapproach. It also emphasizes the importance of strategic investment in educating and trainingemployees in AI technologies while addressing ethical and security concerns.