Using AI to Improve Risk Management: A Case Study of a Leading Telecommunications Provider

In the dynamic and competitive telecommunications industry, effective risk management is critical. Traditional methods are often manual and labor-intensive, prone to inaccuracies and inefficiencies. This paper addresses these challenges by integrating advanced machine learning algorithms to automate and improve the risk management process at a large telecommunications company, hereafter referred to as TP. The study applies state-of-the-art transformer-based machine learning models - BERT, RoBERTa, DeBERTa, and ERNIE 2.0 - alongside a classical support vector machine (SVM) for comparative analysis. The approach leverages the Cross-Industry Standard Process for Data Mining (CRISP-DM) model, emphasizes text classification algorithms to categorize risk events, and introduces a one-class SVM for novelty detection to identify unprecedented risk events. The results show that transformer-based models achieve superior prediction accuracy over traditional SVMs, with the best-performing transformer model (DeBERTa) achieving an F1 score of 92.5%, significantly improving the risk classification accuracy. In addition, the novelty detection mechanism successfully identifies novel risk events with an accuracy of 77.2% and an AUROC of 79.2%. This integration streamlines risk management processes and provides operational benefits to telecommunications companies. By providing insight into these technologies' economic and operational benefits, this study contributes to the practical application of AI in telecommunications, particularly in risk event assessment and natural language processing.