Chronic illness, particularly, Type 2 Diabetes Mellitus (T2DM), is a critical health issue in Sri Lanka with the prevalence being attributed to lifestyle and dietary choices and socioeconomic aspects. Although the global research has already applied Artificial Intelligence in predicting diabetes, no research has been conducted in Sri Lanka since no country-relevant datasets or culturally sensitive preventive interventions reflecting unique dietary and behavioral patterns have yet been available. To overcome this, the proposed study comes up with an Agentic Artificial Intelligence model that automatically interprets medical history, lifestyle, and self-reported health data to grade an individual into low, moderate, and high as regards hyperglycemia risks.