Generative AI Streamlines Accessible Math Visualization Creation for Educators.

This paper introduces a structured, six-phase workflow that empowers educators to design accessible interactive mathematical visualizations using generative AI, eliminating the need for programming skills. The workflow, developed through a design-based research approach, ensures that the generated tools comply with WCAG 2.2 Level AA accessibility standards. The six phases—Foundation, Customization, Mathematical Depth, Application, Accessibility, and Pedagogical Control—guide the prompting of AI models, scaffold code generation, and define critical points for human verification. The research evaluated this workflow across four deployed tools covering various calculus topics. Findings indicate that the structure reliably produces complete first-pass tools. However, human oversight remains essential at every phase to verify mathematical accuracy, accessibility through screen-reader testing, and overall pedagogical suitability before proceeding. This platform-independent workflow serves both instructors and students, significantly lowering the barrier to creating high-quality educational resources.