An AI-Based Menu-Driven Healthcare Chatbot for Underserved Communities

Abstract

Language barriers, low levels of digital literacy, and inadequate medical resources limit people's access to quality healthcare information, particularly in rural and semi-urban areas. In this research study, we present SehatBot, an AI-driven healthcare chatbot that utilizes the WhatsApp Business platform as its delivery vehicle. The goal of SehatBot is to provide multilingual, structured healthcare services in developing regions. Users can ask questions about symptoms they are experiencing, view vaccination schedules, and access health alerts issued by local governments through a user-friendly messaging interface. SehatBot was developed using Flask as the backend framework with a modular architecture. Additionally, it uses webhook technology to communicate with the Meta WhatsApp Business Cloud API, eliminating the need for users to install any additional applications. To provide reliable and context-aware responses, a hybrid NLP approach has been implemented by combining rule-based methods, a menu-driven finite-state mechanism, and optional machine learning-based intent classification. The healthcare knowledge base is stored in JSON format and managed using PostgreSQL in the production environment and SQLite in the development environment. Preliminary evaluations were conducted with 10-15 participants, who reported that SehatBot operated reliably, effectively supported multiple languages, and responded within 0.5 to 1 second. These findings suggest that SehatBot provides a scalable solution for improving health awareness, reducing misinformation, and increasing access to trusted healthcare information in underserved communities.

Keywords

Graphical Abstract

Novelty Statement

This study presents a hybrid AI healthcare chatbot combining rule-based reasoning, finite-state workflows, and NLP to deliver multilingual healthcare support efficiently.