The most commonly used models, for distance training environment, propose learning activities which are generally limited to online lessons consultation and submission of home works to be marked by Graders before they are sent back to the learner. The tutors, seldom, intervene at the time when the learner is in the process of performing his/her learning activities. We present in this paper the design of an agent based interactive environment to assist the learner. This environment offers to the learner self-training resources, tutor service and distance assistance. In this context, the assistant can play the role of a substitute teacher and companion to the learner. We propose the use of an agent-based approach, which is a suitable for two main reasons: agents are a natural metaphor of human and it is convenient for complex distributed systems. Therefore, our aim is to create a personalised artificial assistant agent who is able to communicate with learner and help him/her to perform the tasks. The MaSE (Multi-agent Systems Engineering) methodology is used to develop this system.