Agentic AI untuk Otomatisasi dan Personalisasi Layanan Akademik di Perguruan Tinggi

Abstract

The transformation of higher education institutions from conventional Information Technology (IT)-based systems to Smart Universities requires a systemic and adaptive approach based on artificial intelligence. This study proposes and evaluates the design of an Agentic AI architecture to support academic management and services proactively and autonomously. Using a Design Science Research (DSR) approach, this study designs a multi-agent architecture-based system consisting of sub-agents such as Academic Planner, Advising Agent, and Evaluation Agent. The system was tested with a data sample based on academic service simulation using 500 student entries. The test results show an increase in academic service efficiency, characterized by an average response time of 880 ms, a KRS recommendation accuracy of 92.4%, and a user satisfaction level of 4.5 out of 5. A comparison of the baseline and state-of-the-art shows significant improvements in terms of interoperability, personalization, and operational efficiency. This study concludes that the Agentic AI architecture can be a strategic framework in accelerating the digitalization of academic services and supporting the transformation of higher education institutions towards AI-based Smart Universities.