An Explainable Artificial Intelligence Framework for Intelligent Corporate Financial Decision-Making and Predictive Risk Assessment

Abstract

Corporate financial instability can affect business continuity, liquidity management, investment decisions, and long-term organizational performance, creating a need for reliable and interpretable risk-assessment tools. This study examined the financial and macroeconomic characteristics of bankrupt and healthy firms and developed a simple explainable artificial intelligence framework for corporate financial risk assessment and decision support. A quantitative secondary-data approach was applied using the Hybrid Bankruptcy Dataset for Indian Firms covering 2015–2019, comprising 100 firm-year observations from 20 firms. Descriptive statistics, Mann–Whitney U tests, and Pearson correlation analysis were used to examine the data, while Logistic Regression and Random Forest models were applied for bankruptcy classification. SHAP-based explainability was used to identify the variables contributing most strongly to model predictions. The findings showed significant differences between bankrupt and healthy firms for Debt-to-Equity Ratio and Current Ratio, whereas Return on Assets and Net Profit Margin did not differ significantly. Random Forest produced better overall predictive performance than Logistic Regression, although both models showed only moderate classification capability. Explainability analysis identified Debt-to-Equity Ratio as the most influential predictor, followed by Current Ratio, Net Profit Margin, and Return on Assets, while macroeconomic variables contributed less strongly. The study demonstrates that combining statistical analysis, machine-learning prediction, and explainable AI can provide a practical and interpretable approach to corporate financial risk assessment and support more informed financial decision-making.