The present societal framework heavily relies on human communication, often using language, expressions and non-verbal cues. Autism Spectrum Disorder (ASD), a neurodivergent condition affecting approximately 1–2% of the population, is characterized by impediments in the development of communication or unusual patterns in language acquisition. Teaching pragmatics is essential to help ASD children navigate the intricacies of social interactions, which are essential to building relationships. This paper proposes a framework for both evaluation and acquisition of language in children that can be used to intervene into autistic communication styles, using the powerful AI language models of today. The project features the design of a smartphone application to use a CNN-based language model assisted by emotion and voice tonality recognition, and trained on the forementioned frameworks to bridge the gap between autistic and allistic communication. On necessary implementation, this app may provide a cost-effective, accessible and sustainable way to implement communication therapy for ASD children.