EVALUATION OF CORETAX IMPLEMENTATION: A TAX CONSULTANT PERSPECTIVE

Abstract

This study examines how tax consultants at ABC Consulting experience Coretax, Indonesia’s Core Tax Administration System, in their professional work. These practitioners were chosen as primary informants because, unlike individual taxpayers who interact with the system occasionally and on their own behalf, they operate it daily on behalf of multiple corporate clients under strict statutory deadlines, making them the most intensive and operationally exposed users of the system. The research follows a descriptive qualitative, interpretivist case-study design and gathers data through in-depth structured interviews. To interpret user experiences in a legally mandated adoption context, the study draws on the Task-Technology Fit (TTF) framework alongside an adapted DeLone and McLean IS Success Model that excludes the intention-to-use dimension. Results show that system quality poses the biggest obstacle, driven by persistent access instability and the lack of a user activity log. Recurring data errors undermine information quality, and the helpdesk is rated as slow and technically underprepared, producing satisfaction scores between 4 and 6.5 on a ten-point scale. A mapping of nine consultant task types reveals that Coretax fully supports only three. In contrast, high-frequency tasks such as mass invoicing and income tax calculation remain poorly served, especially during peak filing periods. The study concludes that Coretax adoption reflects regulatory obligation rather than user satisfaction. Its main contribution is to extend the combined DeLone-McLean and Task-Technology Fit framework to a compulsory, professional-user setting. It shows that system quality functions as a gating condition for task-technology fit when exit is not an option. The study recommends prioritizing server reliability, data accuracy, and functional integration, distinguishing recommendations under the tax authority’s control from those under the consulting firm’s control.