In Argentina, competition among mobile communication companies is fierce. Each provider strives to differentiate itself from others. While some telcos choose a low-cost strategy, Telefónica de Argentina stands out from the crowd by providing high-quality services and products. We believe customer experience is one of the most important key performance indicators to consider, even more important than cost. And at Telefónica, we want to continually improve this metric.

One issue that directly affects customer experience is service interruptions due to cell tower failures. We knew that better predicting equipment failure would enable us to take faster action. Additionally, improved maintenance efficiency could help us reduce costs without negatively impacting the quality of our services. To achieve these two ends, we tapped into existing equipment data to better understand why and when our equipment failed.

AI models predict equipment failure

Given our long relationship with IBM, we partnered with the IBM Data Science and AI Elite team to build a machine learning model that would allow us to predict, with at least 20 hours warning, potential tower equipment failure. Starting in late 2019, our joint team developed a proof of concept in six weeks using the IBM Watson Studio solution.

The resulting AI-infused platform gleans insight from existing cell tower data and helps build data models that can predict outages and failures—thereby enabling their prevention. Further, the platform helps us identify how to best allocate maintenance spend.

Best of all, thanks to IBM Watson Studio capabilities, we can continually improve the algorithm and model as they are employed over time.

Learn more about IBM solutions for Telcos

So far, we’re able to predict failures within 24 hours, while our response time to failures has been reduced by a few hours. And we won’t stop there. Our goal is to refine our algorithm and model to help ensure we provide the best telco customer experience in Argentina.

Our quest to improve customer experience

I believe that Telefónica is the first provider in Argentina, and maybe in the region, using AI to gain insight from tower data to predict and prevent equipment failure. We will continue exploring opportunities using Watson Studio and IBM Cloud Pak to more efficiently collect, organize and analyze data so that we can continue to grow our AI capabilities into the business areas.

Due to the flexibility of the IBM Cloud, we have several options: integrating IBM Cloud Pak solutions into our current Hadoop environment with another provider, moving to the IBM Cloud, or shifting to a hybrid cloud environment.

We’ve learned a lot by partnering with IBM, and I believe IBM has also learned from us. It’s been a great partnership, and I look forward to working with IBM in the future to make sure we provide our customers with the best telco experience.