Statement of the Problem: Hypertension and diabetes are prevalent chronic diseases that heavily burden health systems worldwide.

Effective management relies on regular self-monitoring, yet poor adherence and limited patient engagement often reduce treatment efficacy. Al-based conversational agents, such as ChatGPT, offer a potential solution to bridge this gap by facilitating real-time tracking, personalized feedback, and motivational support.

Methodology & Theoretical Orientation: We conducted a 4-week pilot study involving 50 patients with hypertension and/or type 2 diabetes managed in primary care. Participants received an Al-powered mobile assistant designed to log daily blood pressure (BP) and blood glucose (BG) values and deliver adaptive guidance. ChatGPT was customized to:

• Remind users of measurements when due.

• Interpret readings with actionable suggestions.

• Provide educational tips and motivational messages.

Usage data (log frequency, user interactions) and clinical metrics (BP, BG averages) were extracted.

Additionally, semi-structured interviews (n= 15) assessed usability, perceived trust, and self-efficacy.

Findings: Users logged measurements on 85% of scheduled days, with 92% engagement in Al conversations. Mean systolic BP decreased by -6 mmHg and mean fasting BG by -0.4 mmol/L over 4 weeks (p<0.05). Interviews revealed high satisfaction (88%), improved understanding of their condition, and enhanced feeling of support between clinical visits.

Conclusion & Significance: Al conversational agents like ChatGPT can effectively support chronic disease self-management, improving adherence and clinical metrics in only one month. This approach aligns with patient-centered care, health promotion, and digital transformation strategies in cardiology. Further randomized studies should explore lond-tem benents, cost-effectiveness, and integration into clinical workflows.