This project aims to investigate the potential and the added value of automated feedback on written syntheses produced by students of German in Flemish higher education, targeting their strategies regarding cohesion building and textual borrowing. Writing from sources in a foreign language (L2) is a genre that students majoring in languages need to master. However, because their linguistic resources are limited, they often merely cut-andpaste fragments from source texts, which leads to cohesion problems and excessive textual borrowing. While there is a clear need for guidance, providing formative feedback on writing is time-consuming. It is therefore highly relevant to investigate the potential of automated systems to foster student´s awareness of the necessary strategies. In this vein, Automated Writing Evaluation has received growing interest. The advance in Natural Language Processing (NLP) technologies has paved the way for the development of tools that analyse text features based on corpus-linguistic research. Automated Writing Evaluation to date has received little attention within the German L2 research community, although the required NLP resources are at hand. The project aims to fill this gap by investigating the cohesive devices used by German L2 writers based on interlanguage theory, developing an e-tutor based on the results that specifically targets their cohesion and rephrasing strategies, and thoroughly evaluating its added value in higher education.