Learning and Information Technology are majors worries in the knowledge society. Education, especially drop out prevention is one of the most important challenges for the future. The scope of our research is to create a non pedagogical intelligent assistant that supports students during their schooling and prevents them from dropping-out of school. Our solution is to implement a personal intelligent assistant with communications capabilities in natural language. Our assistant I-CAN (Intelligent Coach and Assistant to New way of learning) will have a complementary role to the teacher, it will be used to help student during their homework time outside the classroom. The helps of our personal assistant will focus on the conative aspects of learning. I-CAN support students to manage: the affective strategies, resource strategies, meta-cognitive strategies, educational guidance and technical issues related to the use of the educational platform. The aim of our approach is not to build a new virtual teacher; we want to introduce a tool that collaborates with existing pedagogical systems. Our conative assistant will motivate the student to be more involved in the learning. As our assistant intelligent involves several areas of research, we will explore in this thesis cognitive science, educational science, graphical design, information technology, information systems, and artificial intelligence
Key words: Non pedagogical agent, Intelligent Personal Assistant, conative support, Educational Agent, Tutor, Coach, Student support, School help, Dropout prevention. schooling support.