Intelligenttutoring systems are becoming increasingly popular teaching tools within the business domain. These systems carefully diagnose the behaviour of the learner in order to gain insight into the learner's knowledge state to tailor the instruction according to the needs of the individual learner. Although the artificial intelligence research within the field of diagnosis has progressed significantly, diagnostic processes are still far from perfect and doubts have emerged whether perfection can ever be achieved.This paper, therefore, explores suggestions that have been made on how to achieve effective teaching without detailed student models. It then proposes to merge these suggestions with the valuable ideas that have emerged from the research in the field of intelligent tutoring systems in order to develop effective teaching tools.Still, the development of these hybrid systems raises further questions about the concrete implementation of some of the ideas raised. This paper concludes by addressing some of these problems as a point of departure for further research towards "less intelligent tutoring systems