Broadband Customer Complaint Prediction Using Data-centric AI: A Comparative Analysis of Data Synthesis Algorithms

Broadband is central to today’s Internet connectivity and is used by millions of people world-wide. Thus, a high-quality customer experience is crucial for broadband service providers to guarantee customer satisfaction. To improve broadband customer retention, businesses need to predict customer complaints proactively. While the majority of the existing Artificial Intelligence (AI) research focuses on improving prediction models, we take a data-centric approach. We analyze and compare various data synthesis algorithms and demonstrate that using synthetic data is a reliable way to predict customer complaints.