Task-Level Perceptions of AI Readiness Among Accounting Professionals

Abstract

While artificial intelligence offers significant efficiency gains, accounting professionals evaluate adoption readiness through the specific lens of task-level accountability. This study draws on structured task-perception ratings from 24 practitioners across 24 accounting tasks (576 task-level observations) to examine task-level attributes associated with adoption readiness. The analysis finds that automation potential is positively associated with readiness, suggesting professionals are more open to AI assistance on tasks with higher substitutability. Integration feasibility, defined as the ease with which technology can be embedded into existing workflows, similarly shows a positive association, indicating that governance fit facilitates rather than hinders adoption. Familiarity with large language models shows a negative association, consistent with an experience-skepticism effect whereby greater exposure heightens awareness of technical limitations. The necessity for human judgment is negatively associated but does not reach conventional significance thresholds, suggesting its role may be contingent on the specific nature of the task. Supplementary cluster analysis identifies three distinct task profiles, ranging from judgment-first to automation-oriented, that offer a practical prioritization framework for AI deployment sequencing. Collectively, the findings indicate that task-governance fit, rather than efficiency incentives alone, is associated with adoption readiness among accounting professionals. These results are theory-building and intended to generate hypotheses for larger confirmatory studies.