Many students with Autism Spectrum Disorder (ASD) struggle to appropriately interact and play with their peers at recess. In this pilot feasibility study, we tested the efficacy of practitioner-implemented, peer-mediated Pivotal Response Training (PRT) with 11 elementary and middle school students with ASD. Participants were randomly assigned to a treatment or control group. We measured outcomes at multiple time points, and analyzed data using multi-level modeling with time nested within student. We demonstrated large and statistically significant increases in peer interaction (d = 1.13). Appropriate play with peers also increased substantially (d = 0.89). Practitioners and students provided positive feedback. These findings suggest school staff can feasibly facilitate peer-implemented PRT that improves social outcomes for students with ASD at recess.