Background: Individuals with disabilities and type 2 diabetes require self-management programs that are accessible, sustainable, inclusive, and adaptable. Health coaching has been shown to be an effective approach for improving behavioral changes in self-management. Health coaching combined with telehealth technology has the potential to improve the overall quality of and access to health services. Objective: This protocol outlines the study design for implementing the Artificial Intelligence for Diabetes Management (AI4DM) intervention. The protocol will assess the feasibility, acceptability, and preliminary efficacy of the AI4DM telehealth platform for people with disabilities. Methods: The AI4DM study is a 2-arm randomized controlled trial for evaluating the delivery of a 12-month intervention, which will involve telecoaching, diabetes educational content, and technology access, to 90 individuals with diabetes and physical disabilities. The hypothesis is that this pilot project is feasible and acceptable for adults with permanently impaired mobility and type 2 diabetes. We also hypothesize that adults in the AI4DM intervention groups will have significantly better glycemic control (glycated hemoglobin) and psychosocial and psychological measures than the attention control group at the 3-, 6-, and 12-month follow-ups. Results: The AI4DM study was approved by the university's institutional review board, and recruitment and enrollment will begin in October 2021. Conclusions: The AI4DM study will improve our understanding of the feasibility and efficacy of a web-based diabetes self-management program for people with disabilities. The AI4DM intervention has the potential to become a scalable and novel method for successfully managing type 2 diabetes in people with disabilities.