Designing a Model of the Impact of Artificial Intelligence Capabilities in Accounting Systems on Reducing Errors in Financial Reports

Abstract

The present study aimed to design and validate a model explaining the impact of artificial intelligence capabilities in accounting systems on reducing financial reporting errors in organizations and financial institutions. This study was conducted using a mixed-methods exploratory sequential design. In the qualitative phase, semi-structured interviews were conducted with 18 university professors, financial managers, auditors, and accounting information system specialists in Tehran using purposeful and snowball sampling methods until theoretical saturation was achieved. The qualitative data were analyzed through thematic analysis and coding procedures using MAXQDA software. In the quantitative phase, the statistical population consisted of accountants, financial managers, auditors, and accounting system specialists employed in financial institutions and private companies in Tehran during 2026. Based on Cochran’s formula, 384 participants were selected through stratified random sampling. Data were collected using a researcher-developed questionnaire designed according to the qualitative findings and existing theoretical literature. The validity of the instrument was confirmed through expert evaluation and confirmatory factor analysis, while reliability was assessed using Cronbach’s alpha and composite reliability coefficients. Quantitative data were analyzed using SPSS version 27 and SmartPLS version 4 through descriptive statistics, confirmatory factor analysis, correlation analysis, and structural equation modeling. The findings demonstrated that all dimensions of artificial intelligence capabilities had significant positive effects on reducing financial reporting errors. Real-time financial monitoring had the strongest effect on reducing financial reporting errors (β = 0.39, p < 0.001), followed by automated anomaly detection (β = 0.34, p < 0.001), decision-support functionality (β = 0.31, p < 0.001), intelligent data processing (β = 0.28, p < 0.001), predictive analytics capability (β = 0.22, p < 0.001), machine learning integration (β = 0.19, p < 0.001), and system adaptability (β = 0.17, p < 0.001). The structural model exhibited satisfactory fit indices, and the coefficient of determination indicated that artificial intelligence capabilities explained 74% of the variance in reduction of financial reporting errors. The findings indicate that artificial intelligence capabilities significantly improve the accuracy, transparency, and reliability of financial reporting systems by enhancing monitoring processes, anomaly detection, predictive analysis, and intelligent decision-making. The study confirms that intelligent accounting technologies can substantially reduce financial reporting errors and strengthen organizational financial governance. Therefore, organizations are encouraged to invest in intelligent accounting infrastructures and develop technological readiness to improve financial reporting quality and operational efficiency.