In agile product development, the Daily Scrum Meeting is often used to improve communication, coordination and collaboration. However, practice shows that simply introducing this event is not enough. This paper presents an AI-based analysis method that automatically evaluates communication and interaction patterns using large language models and helps to systematically improve the effectiveness of daily scrum meetings. This makes it possible to systematically analyze key influencing factors such as time management or team participation, identify potential for improvement and derive targeted measures that promote transparency, flexibility, self-organization and the ability to learn.