An AI-Based Integrated Model for Business Reputation Verification: Mathematical Formalization and Simulation-Based Assessment

Abstract. Business reputation has become a dynamic, data-intensive and risk-sensitive phenomenon in the digital economy. Traditional approaches to reputation assessment, including stakeholder surveys, media monitoring, expert evaluation and corporate rankings, remain important but are no longer sufficient for real-time reputation verification. Reputational signals are increasingly generated through online reviews, social media reactions, digital news, ESG disclosures, regulatory announcements, customer complaints and algorithmically amplified narratives. Artificial Intelligence creates new opportunities for processing these heterogeneous data sources through sentiment analysis, natural language processing, machine learning, clustering and anomaly detection. However, existing AI-based approaches often remain fragmented because they solve separate analytical tasks without transforming their outputs into an interpretable integral reputation score.

The aim of this article is to develop an authorial AI-based integrated model for business reputation verification and to test its internal logic through simulation-based calculations. The study applies conceptual modelling, mathematical formalization, hypothesis development, and simulation-based analysis. The proposed Business Reputation Verification Index integrates six components: Sentiment Score, Media Visibility Score, Stakeholder Trust Score, ESG Reputation Score, Crisis Probability Score and Anomaly Score. The simulation-based testing of the model is conducted using a synthetic dataset of 300 enterprises.

The findings confirm the internal consistency and analytical logic of the proposed model. Sentiment, ESG reputation and media visibility positively affect stakeholder trust. Stakeholder trust reduces crisis probability, while anomaly intensity increases it. The final Business Reputation Verification Index decreases when crisis probability and anomaly intensity rise. The simulation results demonstrate that AI-based business reputation verification should not rely on sentiment analysis alone. It should combine sentiment, trust, ESG signals, crisis prediction and anomaly detection within a single interpretable framework. The scientific contribution of the article lies in the development of a formalized AI-based reputation verification model that connects machine learning outputs with an integral index suitable for managerial decision-making and early warning analysis.