Modeling is often indispensable in tackling the increasing complexity of software being developed, and the skillset that a good modeller needs to master is ever expanding. Modellers need to handle a multitude of specific abstractions, their use across the software life-cycle, and their relationships with other abstractions to ensure global consistency of the modelling activities and resulting modelling artifacts. Effective teaching of such essential modelling skills is crucial. As enrolment numbers in Computer Science and Software Engineering classrooms continue to increase, the use of modelling-specific teaching tools is becoming a must. This talk reviews some of the current work in this area, presents RF-IMA, a recently published reference framework for intelligent modelling assistants, and then explores the idea of intelligent modelling assistants (IMA) for teaching of the future, i.e., tools that automatically derive modelling knowledge, learn from student experiences, and provide context-aware teaching assistance tailored to the individual student’s expertise and learning style.