EduVa: Prototyping and testing AI-powered interactive LMS with adaptive modules and assessments
TL;DR
The findings demonstrate that the EduVa LMS is valid, usable, and effective for implementing adaptive learning, and shows a strong correlation between AI assessments and learning objectives.
Abstract
The rapid integration of artificial intelligence (AI) in higher education has increased the demand for adaptive Learning Management Systems (LMS) capable of delivering personalized learning at scale. However, empirical studies covering the full lifecycle of designing, prototyping, and testing AI-powered LMS remain limited, particularly in developing regions. This study aimed to design, prototype, and test EduVa LMS (Education Virtual Assistant), an AI-powered interactive LMS with adaptive modules and AI-driven assessments, implemented across ten private universities in East Nusa Tenggara, Indonesia. A Design Science Research (DSR) approach integrated with the ADDIE model guided the development process. A total of 360 participants were involved, including experts, students, and lecturers. Validation results showed a Content Validity Index (CVI) of 0.80, confirming content validity. Usability testing using the System Usability Scale (SUS) yielded mean scores of 82.4 (students) and 85.8 (lecturers). Adaptive performance analysis indicated 82.1% assessment accuracy, 3.7 content adjustments per session, and an 87.9% completion rate. A strong correlation (r = 0.81, p < .01) was found between AI assessments and learning objectives. The findings demonstrate that the EduVa LMS is valid, usable, and effective for implementing adaptive learning.