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

Artificial Intelligence (AI) is a key driver of digital transformation in tax administration, enabling governments to streamline administrative processes and improve taxpayer compliance. However, In Indonesia, the shift towards integrated digital platforms continues to encounter operational difficulties, necessitating a clearer understanding of the impact of AI-enabled capabilities on system usability and taxpayer satisfaction. This study uses Nielsen’s usability framework alongside Perceived AI Usefulness (PA) to evaluate Indonesia's digital tax system. A quantitative method was employed to collect data online from 238 active taxpayers between April and May 2026. Sample size adequacy was validated using Hair et al.’s PLS-SEM guidelines and G*Power analysis. The PLS-SEM results indicate that Perceived AI Usefulness significantly improves all four core usability dimensions: Efficiency, Learning Usability, Minimized Errors, and Memorability. Furthermore, immediate functional usability, particularly in terms of Efficiency, Learning Usability, and Minimized Errors, has a substantial positive impact on User Satisfaction, whereas Memorability exhibits no statistically significant effect.