Case Study: a large US telco and media provider operationalizes machine learning with H2O.ai

A large US telco and media provider predicts trending content 24 hours ahead with H2O.ai

A large US telco and media provider faced the challenge of operationalizing machine learning on its complete production-scale data to solve complex business problems. They needed to move beyond analysis and deploy models into live systems to improve customer experience and operational efficiency. To achieve this, they partnered with H2O.ai and utilized the H2O-3 platform.

H2O.ai provided a solution that enabled the telco to train models on full datasets and deploy them as Java code (POJO) or web services. This allowed the company to implement several high-value applications, including a model to significantly reduce avoidable truck rolls, a system to predict content popularity 24 hours in advance, and a customer-experience metric that spans its entire network. The results were a reduction in unnecessary technician visits, improved content forecasting, and the creation of self-healing systems that automatically diagnose and resolve problems.