Tax digital twin: A reference architecture for continuous enterprise tax compliance

Abstract

Enterprise tax compliance remains predominantly reactive despite significant advances in enterprise information systems. Existing tax technologies primarily identify errors after transactions have been executed, limiting organizations' ability to anticipate how strategic business decisions—such as market expansion, supply-chain restructuring, or regulatory change—will affect future tax obligations. This paper addresses this limitation by developing a Tax Digital Twin (TDT) reference architecture that enables continuous simulation of enterprise tax compliance before operational decisions are implemented. Following a Design Science Research (DSR) methodology, the study introduces a reusable architectural artifact that extends digital twin theory beyond its traditional application to physical assets into the domain of enterprise tax management. Building upon Grieves' physical–virtual twin paradigm and control-theoretic synchronization principles, the proposed architecture represents enterprise tax posture as a continuously synchronized virtual model capable of predictive analysis, scenario simulation, divergence detection, and decision support. The architecture is formalized through a state representation model, synchronization mechanisms, and divergence-driven governance processes, and is illustrated using four representative enterprise scenarios involving multi-jurisdiction expansion, exemption certificate expiration, tax engine migration, and post-merger integration. Rather than replacing existing AI-based tax detection systems, the proposed Tax Digital Twin complements them by introducing proactive compliance simulation and enterprise decision intelligence. Based on a structured review of the digital twin and enterprise tax compliance literatures, no prior study was identified that proposes a reference architecture for continuous enterprise tax compliance simulation using digital twin principles; this paper addresses that gap and establishes a theoretical foundation for future empirical research, AI-enabled tax systems, and next-generation enterprise information systems.