This talk presents an innovative AI feedback tool developed to enhance the learning process of students and support teachers. As students increasingly use AI when writing their work, the focus in education is shifting from assessing end products to guiding learning processes. The tool responds to that shift by providing students with direct, targeted feedback during their writing process, based on assessment criteria set by teachers, such as content, argumentation, writing style and structure.

What distinguishes this tool from existing AI tools is that it does not just give a single, general assessment, but works according to an agentic AI approach. This means that multiple specialised AI agents analyse different assessment aspects in parallel and independently. This creates a rich, multifaceted and balanced picture, in which students can clearly see both their strengths and their areas for improvement.

In addition, the feedback is not generic or difficult to interpret (as is often the case with standard AI solutions), but explicitly linked to specific text passages. Students know exactly what the feedback refers to and how they can improve their work. The generated feedback can also be exported and used to monitor the learning process, for example in self-reflection or guidance interviews. For teachers, this provides valuable insights into how students deal with feedback, without necessarily using it as an assessment tool.

During the session, the tool will be demonstrated and dialogue with you and the other participants will be actively sought: how can we integrate AI into education in a responsible and meaningful way? And how can we strengthen the further development of this tool, precisely through collaboration with educational practitioners?