SMART AI: designing intelligent systems for health, finance, and medical decisions for citizens.

Abstract

As global populations age, particularly in low- and middle-income countries (LMICs), there is an urgent need for intelligent systems that address the intertwined challenges of health monitoring, medical diagnostics, and financial management for elderly citizens. Existing AI solutions in these domains are often fragmented, digitally inaccessible, and culturally misaligned, leaving older adults especially those with low digital literacy underserved. This thesis presents SMART AI, an integrated, human-centered framework designed to empower the “smart elderly citizen” through three core subsystems: (1) HishabAI a Bengali-language financial assistant that automates expense tracking, including a dedicated medical finance tracker for healthcare-related costs and budgeting via SMS parsing, OCR, and AI-based forecasting; (2) ElderCare Pulse a wearable-enabled health monitoring system that tracks vital signs, medication adherence, and emergency events, with caregiver and institutional integration; and (3) a CNN based diagnostic engine capable of analyzing medical images to support clinical decision making in resource-constrained environments. The proposed development plan adopts a monolithic Django–PostgreSQL backend, a cross-platform Flutter mobile application, and Vue.js dashboards, with a design philosophy centered on localization, interpretability, privacy, and modular scalability. Field evaluations and prototype testing highlight the system’s feasibility and usability in semi-urban and rural Bangladeshi contexts, demonstrating its potential to deliver inclusive, interoperable AI solutions that connect personal well-being with institutional care for aging populations.