Related Experiment Videos
Predicting corporate management performance using AI: Incorporating CEO strategy insights from sustainable management
Xiao Wang1, Feng Sun2, Yong Ki Kim3
1School of Business Administration, Binzhou Polytechnic, Binzhou, Shandong, China.
Plos One
|May 6, 2026
Summary
This study uses AI to predict corporate performance by analyzing financial data and CEO messages. Combining both data types significantly improves prediction accuracy, with Transformer models showing the best results.
Area of Science:
- Accounting
- Artificial Intelligence
- Corporate Finance
Background:
- Predicting corporate management performance is crucial for investors and stakeholders.
- Traditional methods often rely solely on financial data, potentially overlooking strategic insights.
- Sustainability reports offer a rich source of strategic information through CEO communications.
Purpose of the Study:
- To develop and evaluate an AI-based model for predicting corporate management performance.
- To assess the impact of integrating strategic information from CEO messages with financial data.
- To compare the effectiveness of various machine learning and deep learning models in this prediction task.
Main Methods:
- Utilized a dataset of 1,271 South Korean listed companies (2016-2023).
- Employed text mining on CEO messages within sustainability reports, categorized using the Sustainable Balanced Scorecard (SBSC) framework.
- Applied eight classifiers: KNN, SVM, GBM, CatBoost, GAN, RNN, LSTM, and Transformer, evaluating hybrid models combining financial and SBSC strategy variables.
Main Results:
- Hybrid models integrating financial and strategy variables outperformed financial-only models.
- The Transformer model demonstrated the highest predictive accuracy (Accuracy=0.8467, AUC=0.8481, F1=0.8572).
- Incorporating SBSC strategy indicators enhanced mean model performance across all tested classifiers.
Conclusions:
- AI models combining financial data and strategic insights from CEO messages offer superior corporate performance prediction.
- Deep learning models, particularly Transformers, show significant promise in analyzing complex strategic information.
- Findings support the value of integrating qualitative strategic data with quantitative financial metrics for enhanced corporate analysis and investment decisions.
