DEVELOPMENT OF AN OFFLINE BILINGUAL ACADEMIC ASSISTANT. CASE STUDY: ATBU STUDENTS & GUARDIANS

This study presents the design and prototype of an offline, bilingual Artificial Intelligence (AI) system aimed at providing academic support to students and guardians of Abubakar Tafawa Balewa University (ATBU), Bauchi, Nigeria. Addressing significant challenges such as inconsistent internet access, language barriers, and complex digital interfaces prevalent in the region, the system is designed to operate entirely offline on Android devices. It integrates both English and Hausa language support to facilitate access to essential academic information, including CGPA calculation, fee structures, academic calendar guidance, and registration walkthroughs. The methodology employed focuses on a user-centered design approach, incorporating lightweight natural language understanding (NLU) through predefined phrase matching and keyword triggers, supported by a conceptual integration of MobileBERT for scalable language processing. Data is stored locally using SQLite and JSON formats, ensuring rapid retrieval and responsiveness without relying on server communication. The user interface mimics popular chat applications to enhance familiarity and ease of use for individuals with varying levels of digital literacy. Through a modular architecture, the project prioritizes core functionalities and demonstrates their feasibility via mockups, simulated interactions, and structured user interviews. While the current phase emphasizes system design and prototyping rather than full deployment, the evaluation plan includes task accuracy reviews, language responsiveness tests, and offline performance verification. This research contributes a practical, localized AI solution that champions accessibility and inclusivity in educational technology for underserved communities, offering a template for future developments in similar contexts across Nigeria and West Africa.