Pass rates in first-year university Calculus courses are typically low. In this work, we present an updated version of our e-learning platform, GIEMATIC, designed to support students in mastering these subjects. The platform offers interactive learning tracks and self-assessment exercises that allow students to practice and evaluate their understanding. In the new version, we have focused on enhancing diagnostic capabilities for professors through more detailed server-side analytics, expanding the number of chained exercises in areas where students commonly face difficulties, and integrating AI-powered chatbots as virtual learning assistants to provide additional motivation and support.

The platform has been implemented in several engineering Calculus courses. Rather than replacing traditional instruction, it serves as a complementary framework for learning activities. With the new analytical tools, we are able to examine student performance and its relationship to final grades and pass rates. Additionally, we conducted a survey to evaluate the use and perceived usefulness of the chatbot feature.

Results show that the logging system successfully supports meaningful analysis. While outcomes vary depending on the teaching approach to the use of the platform, a consistent pattern emerges: students who achieve higher final grades are also the ones more actively engaged with the platform. However, the number of highly committed students remains low across courses. The motivational effect of AI chatbots has proven less significant than anticipated. The main challenge moving forward is to encourage deeper and more sustained student engagement with these tools so they can fully benefit from them, rather than using them only superficially.

Keywords: Interactive exercises, chained exercises, AI chatbots, student engagement, learning tracks.