Vadim Vaks is an Solutions Architect at Hortonworks, helping organization realize the power of Big Data. Vadim has a background in Enterprise Architecture, Service/Data Integration, Distributed and In-memory computing. Vadim is Drexel University School of Information Technology Alum.
Transforming the customer experience through real time predictive maintenance and workforce management
Customer service is a critical point of focus for any service provider. In the case of Telecom, much of the existing customer experience depends on reliable and consistent operation of devices on the customer premises (CPE). When a piece of customer premises equipment does break down, the customer looks to front line employees like customer service representatives and front line technicians to restore service. The efficiency and efficacy of these customer interactions will largely determine whether the customer has a positive experience using the service and in the long run, whether they will remain a customer.
This presentation explores the implementation of a “Connected Platform” that enables customer premise device data collection in real time and integration with existing Workforce Management platforms. The platform leverages large volumes of data collected from CPE devices to create ML models that can predict whether and when a particular piece of CPE equipment may fail.
The platform can then leverage integration with Workforce Management platforms to determine the nearest available technician that can be assigned to proactively address the potential CPE malfunction. Such a platform could enable a customer service representative or automated systems to proactively contact the customer, attempt to resolve the potential malfunction over the phone, and if necessary, schedule an immediate service call.
This approach would greatly reduce the number of unplanned CPE outages, greatly reduce time to resolution of ongoing CPE outages (MTTR), and improve workforce utilization efficiency (reduced unplanned truck rolls, reduced technician time un-utilized). The end result is a more satisfied customer base for a lower operational cost.
The presentation will address topics of CPE data collection, application of Data Science practice to create ML models using collected data, as well as implementation and operation of streaming applications that apply the resulting ML models in the live stream of CPE and workforce data.