Objectives: Chatbot technology improves access to and engagement with diabetes education. However, few studies have evaluated the feasibility and rigorously assessed the impact of chatbots among individuals with type 2 diabetes (T2DM) using theory-based approaches. This pilot study assessed the feasibility and preliminary impact of a chatbot on glycemic targets, loneliness, and perceived health beliefs among adults with T2DM. Methods: An explanatory mixed-methods approach, comprising a one-group experimental design and qualitative interviews, was used. The chatbot simulation, named “TakonGendhis,” was developed based on conceptual models derived from the technology acceptance model and the health belief model. Feasibility included usefulness, ease of use, and intention to use. Preliminary impact was evaluated based on changes in glycemic targets, loneliness, and health beliefs from baseline to 12 weeks post-intervention. Qualitative data were gathered through individual interviews and focus group discussions and were analyzed thematically. Narrative synthesis was employed to integrate findings from the quantitative and qualitative phases of the study. Results: The scores for usefulness, ease of use, and intention to use were 26.55, 27.32, and 34.03, respectively. Quantitative analysis revealed reduced loneliness, improved health beliefs, and lower glycemic scores after the 12-week intervention. The qualitative study identified 4 themes: feasibility, beliefs, emotional support, and areas for improvement. Conclusions The intervention was feasible and had beneficial preliminary impacts on glycemic targets, loneliness, and health beliefs. Addressing feasibility, beliefs, emotional support, and identified areas for improvement may increase patients’ willingness to use the chatbot.