<p>Children with autism spectrum disorder often have severely limited speech and language development and these children might therefore benefit from alternative and augmentative communication (AAC) intervention. While there is an extensive literature demonstrating effective procedures for teaching children with developmental disabilities and severe communication to use AAC, studies in this area have mainly focused on teaching relatively short and single-function communicative exchanges, such as teaching the person to request access to a preferred object by producing one communication act, such as by selecting a single symbol on a speech-generating device. While requesting access to a preferred object is useful, it would also seem important for children to be able to engage in more extended and multi-functional communication exchanges. However, there is little research into the development and evaluation of procedures for teaching more extended, multi-function communication sequences, such as teaching the child to (a) first greet his or her listener, (b) then make a request for a general object (“I want a snack.”), (c) then make a request for a specific object (“I want popcorn.”), and (d) then thank the listener for providing the requesting item. Furthermore, there is no published literature teaching communicative behaviours on a speech-generating device using the antecedent prompting procedure. The aim of this research is to determine whether children with developmental disabilities and severe communication impairment can learn to produce such extended and multi-functional communication sequences by systematically implementing the antecedent prompting procedure with an iPad-based speech-generating device. Study 1 used a multiple baseline design to evaluate the effectiveness of the antecedent prompt to teach a 4-step communicative exchange on an iPad®-based speech-generating device set with a progressive display. Participants were required to discriminate between two icons at a time, and the progressive display guided them through four screens, or the four steps, of the communicative exchange. All five participants (4 boys, 1 girl, aged 6 to 10 years) who participated in this study mastered the 4-step communicative exchange in four sessions on average. Study 2 was identical to Study 1, except the iPad®-based speech-generating device was set with a static screen. Participants (4 boys, aged 6 to 10 years) were required to discriminate between 12 icons simultaneously and select the appropriate icon for each step of the exchange. Of the participants, three were performing the sequence correctly after one session on average, one student did not reach mastery criteria but the average number of his communicative exchanges per interaction increased. Considering the positive outcomes and rapid rate of acquisition from both studies, future research should consider implementing the antecedent prompt when teaching communication interventions to minimally verbal children with autism on a speech-generating device.</p>