Comparison of Classical Approaches and Artificial Intelligence Models in Small Business Failure Prediction: A Literature Review

Abstract

Microenterprises, which are essential to the global economy, face significant failure risks due to their structural vulnerability and limited resources. In an increasingly complex and uncertain economic environment, traditional risk management methods have shown their limitations, leading to the exploration of innovative solutions such as artificial intelligence. This paper examines how advanced technologies like neural networks and random forests can help microenterprises more accurately and proactively identify early warning signs of failure. The methodology is based on an in-depth literature review, with data collected from academic databases. After filtering relevant sources, a detailed analysis of methodologies and results allowed for a comparison of AI models and classical models, highlighting the advantages and limitations of each approach. The study revealed that AI offers new opportunities to anticipate risks by analyzing complex data that traditional methods cannot exploit. However, challenges remain, particularly regarding data quality, model interpretability, and access to technologies, thus limiting the full integration of these tools into decision-making processes for microenterprises. Nevertheless, AI represents a potentially transformative lever to enhance the resilience and competitiveness of microenterprises in the face of economic uncertainties.