Designing AI Productivity Assistants for Neurodiverse Professionals: A Hybrid Low-Code Approach to Cognitive Co-Regulation

Traditional productivity tools mostly cater to the needs of neurotypical users, which unfortunately excludes neurodivergent individuals with conditions such as Autism Spectrum Disorder (ASD), Attention Deficit Hyperactivity Disorder (ADHD), dyslexia, and dyspraxia. This work proposes a neuro-pluralistic method with AI-powered cognitive co-regulators. The neuro-pluralistic method with AI-powered cognitive coregulators involves a hybrid architecture that combines Low-Code/No-Code (LCNC) frontend prototyping with high-code Python and LLM-based backend development, using privacy-preserving Retrieval-Augmented Generation (RAG). A prototype was tested alongside a controlled withinsubjects study with 24 neurodivergent professionals (12 ADHD, 6 ASD, 4 dyslexia, 2 dyspraxia; mean age 32.6 years, standard deviation 6.8). The prototype produced drastically lowered cognitive load results (a 48.3% reduction, NASA-TLX: 78.4→40.5, t(23)=8.94, p<0.001, d=2.15), higher goal completion rate (45%→75%, t=7.21, p<0.001, d=1.82), and less time on task completion (30.8%, 65.4→45.2 minutes, t=6.45, p<0.001, d=1.62). The hybrid design cut down the time-to-prototype to only three weeks as opposed to the months in traditional development. Performing an ablation study (n=12) confirmed that DopBoost (18% load reduction), Task Re-entry support (22%), and personalization (15%) have unique contributions. The paper offers neuro-inclusive AI validated design principles, a dual-memory personalization framework (PRIME), and a replicable blueprint for rapid R&D in accessibility-focused AI systems.