Key points are not available for this paper at this time.
Conversational technologies are increasingly investigated as supportive tools for autistic individuals; however, existing approaches remain largely application-driven and conceptually fragmented, with limited integration between conversational design and the cognitive foundations of daily living skills. To date, no unified framework explicitly conceptualizes chatbot-based interaction as a structured environment supporting functional autonomy in autism. This Perspective article introduces the Neuroinclusive Conversational Framework (NCF), a theoretical model that reframes chatbots as cognitive scaffolds designed to support everyday functioning in addition to social simulation. Drawing on cognitive psychology, executive functioning research, mediated learning theory, and neuroinclusive design principles, the NCF conceptualizes conversational interaction as a structured learning environment capable of supporting planning, sequencing, task initiation, and self-monitoring within daily routines. The framework is articulated along three interrelated dimensions-cognitive-functional, structural-adaptive, and contextual-ecological-each addressing mechanisms relevant to cognitive accessibility, adaptability, and ecological transfer. By synthesizing evidence across diverse conversational technologies and identifying recurring design principles such as predictability, low social pressure, and adaptive structure, the NCF provides a coherent lens for interpreting existing systems and guiding future design. The article further discusses emerging directions, including biofeedback-informed adaptivity, perceptually accessible interface design, and context-sensitive interaction, emphasizing the importance of ecological validity and integration with human guidance. Overall, the NCF offers a theoretical foundation to support systematic research and development of conversational technologies aimed at strengthening daily living skills and functional autonomy in autistic individuals.
Meduri et al. (Tue,) studied this question.