Measuring AI-Driven Usability and User Satisfaction in Indonesia's Digital Tax Governance System: A PLS-SEM Analysis

Artificial Intelligence (AI) has become an integral component of digital transformation in tax administration, enabling governments to improve operational efficiency, service quality, and taxpayer compliance. Despite these benefits, the implementation of AI-powered tax systems still faces challenges related to usability and user acceptance, particularly in developing countries like Indonesia. This study aims to examine the factors influencing users' acceptance of AI-enabled tax systems by integrating usability dimensions into the research model. A quantitative approach was employed by collecting data through an online questionnaire from 215 respondents in June 2026, and the data were analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM). The findings reveal that AI and Machine Learning significantly improve system efficiency, learning usability, minimized errors, and memorability. In addition, efficiency, learning usability, and minimized errors significantly contribute to user satisfaction, whereas memorability does not have a significant effect on user satisfaction. This study contributes to the existing body of knowledge by extending usability research in digital tax governance through the integration of Artificial Intelligence and Machine Learning. It provides empirical evidence that AI-enabled usability, particularly efficiency, learning usability, and minimized errors, plays a crucial role in enhancing user satisfaction. The findings offer both theoretical and practical implications for tax authorities, system developers, and policymakers by emphasizing the importance of prioritizing operational usability over memorability in the design of AI-enabled digital tax systems.