Augmentative and Alternative Communication (AAC) systems can also be used to assist people with disabilities to develop their intellectual functions. Some of these systems provide facilities to customize its vocabulary. With these facilities, mediators can take advantage of theoretical foundation like Meaningful Learning and Stimulus Equivalence to create adapted learning activities for AAC systems. In this sense, this work proposes a framework with six types of activities that can be adapted in AAC systems. The main features of these activities are: 1) to present a degree of increasing difficulty; 2) to be based on Meaningful Learning, and Stimulus Equivalence, by using conditional discrimination to explore the relations of reflection, symmetry, and transitivity; and 3) to employ multi-model correspondence to work order and sequence between discriminative stimuli. We show these activities in a sketch of a general AAC system, and as a proof of concept, we adapted two of these activities into the aBoard Platform.