Rigorous, quantitative examination of therapeutic techniques anecdotally reported to have been successful in people with autism who lack communicative speech will help guide basic science towards a more complete characterisation of the cognitive profile in this underserved subpopulation, and show the extent to which theories and results developed with the high-functioning subpopulation may apply. This study examines a novel therapy, the Rapid Prompting Method (RPM). RPM is a parent-developed communicative and educational therapy for persons with autism who do not speak or who have difficulty using speech communicatively. The technique aims to develop a means of interactive learning by pointing amongst multiple choice options presented at different locations in space, with the aid of sensory prompts which evoke a response without cueing any specific response option. The prompts are meant to draw and to maintain attention to the communicative task – making the communicative and educational content co-incident with the most physically salient, attention-capturing stimulus – and to extinguish the sensory-motor preoccupations with which the prompts compete. Video-recorded RPM sessions with 9 autistic children ages 8 to 14 years who lacked functional communicative speech were coded for behaviours of interest. An analysis controlled for age indicates that exposure to the claimed therapy appears to support a decrease in repetitive behaviours and an increase in the number of multiple-choice response options without any decrease in successful responding. Direct gaze is not related to successful responding, suggesting that direct gaze might not be any advantage for this population and need not in all cases be a precondition to communication therapies.