Difficulty with social communication is regarded as one of the most significant features of autism. Social communication involves an initiation-response sequence where children take turns communicating for social purposes (Stanton-Chapman & Snell, 2011). However, many children with autism initiate or respond less frequently than other children (Bauminger, Shulman, & Agam, 2003). In addition, the quality of this initiation-response sequence is affected by their difficulties with remaining on topic (Capps, Kehres, & Sigman, 1998; Paul, Orlovski, Marcinko, & Volkmar, 2009; Siegel, 1996; Sng, Carter, & Stephenson, 2017; Tager-Flusberg & Anderson, 1991). The ability to communicate with peers to initiate and respond is particularly important because of its effect on the development of positive peer interactions (Bauminger-Zviely, Karin, Kimhi, & Agam-Ben-Artzi, 2014). Researchers recommend that interventions for children with autism occur in the child’s natural context and that they utilise naturally occurring activities and materials to improve the generalisation of intervention outcomes (Bellini, Peters, Benner & Hopf, 2007; Cowan & Allen, 2007; Gresham, Sugai, & Horner, 2001; Hansen, Frantz, Machalicek, & Raulston, 2017). As schools are a natural social context for primary school-aged students, a systematic review was conducted to evaluate interventions in mainstream schools that targeted the initiating and responding behaviours of students with autism (Sutton, Webster, & Westerveld, 2018). Results indicated that school-based social communication interventions are more likely to be conducted by researchers and are conducted away from the classroom in empty rooms. As a consequence, teachers did not receive the necessary training or experience to implement social communication interventions in their own classroom. One explanation for this is that few social communication interventions are designed specifically for classroom settings (Bellini et al., 2007; Ostmeyer & Scarpa, 2012). To address this, a social communication intervention known as Social Stations was developed to teach the behaviours of initiating and responding on topic. Social Stations is a peer-mediated intervention that utilises the iPad to provide motivation and visual structure in order to engage students with autism in conversation with typical peers as part of a literacy lesson. A study to evaluate the Social Stations intervention was set in a Queensland mainstream primary school with five verbal students with autism (aged six to nine years). This study evaluated the implementation of this intervention to assess the degree of change in (a) the frequency and quality of specific social communication behaviours of students with autism, and (b) teachers’ self-efficacy for teaching social communication. The study used a multiple baseline across participants design to evaluate changes in the students’ initiating and responding behaviours, both on and off topic. The frequency of initiating and responding behaviours was measured across the intervention phases and changes were analysed for both significance and effect size. The study also aimed to qualitatively explore teachers' perspectives regarding their self-efficacy for teaching social communication. The students’ teachers were interviewed before and after the intervention. Their responses were inductively and deductively evaluated to determine if changes to their self-efficacy occurred following their training and implementation of the Social Stations intervention. In addition, teachers evaluated the social validity of the intervention and utilised a goal attainment scale to assess their capacity to overcome barriers to teaching social communication. All students with autism showed significant improvements in the frequency and quality of targeted social communication behaviours at the end of the intervention. These positive outcomes were maintained over time and generalised to a new context and new topic of conversation. The teachers in this study reported that the Social Stations intervention was socially valid, and they believed they could commit more time to teaching social communication in their classroom. Teachers changed their self-efficacy for teaching social communication following their involvement in the intervention, as they observed the positive outcomes for their students. Teachers reported increased confidence and believed they could embed the intervention into their classroom program. Their experience overcoming barriers, such as a lack of time to teach social communication, enabled them to develop self-efficacy by formulating plans to utilise the intervention in the future. This study has the potential to make a significant contribution to the research literature on classroom-based interventions and for shaping practice for students with autism in mainstream school settings. The insights gained from this study show that teachers can develop self-efficacy for teaching social communication in their classrooms, if they are provided with an intervention that can be embedded into the curriculum.