Supporting Agile Roles and Practices with GAI

Generative Artificial Intelligence (GAI) is reshaping how software development is learned and practiced. In the last two years, our research has focused on integrating agile roles and practices with GAI in both educational and professional contexts. This work investigates how a Large Language Model (LLM) can support not only core development activities but also agile practices such as test-driven development, refactoring, and retrospectives, as well as roles like the Product Owner and Scrum Master. By treating LLMs as linguistic collaborators, we explore their potential to enhance efficiency, collaboration, and decisionmaking while also highlighting challenges related to prompt engineering, rhetoric, and responsible use. The resulting insights lead to recommendations for training future software professionals to work effectively with GAI, leveraging its strengths while critically addressing its limitations.