Leading communication services provider uses AI and Hazelcast IMDG to handle over 1 Million support inquires per day

The customer is one of the largest providers of internet, voice, and media products and services to both business and residential users in the United States. This media conglomerate handles over one million customer interactions per day through various support channels, including call centers, websites, and mobile devices. The company aims to provide the best possible customer experience by leveraging advanced technologies to automate support processes and reduce response times. They have a significant focus on digital self-service as the first line of defense to reduce support hold times and offer more personalized customer interactions. The organization is committed to continuous service improvements and prides itself on its ability to handle a large volume of customer interactions efficiently.

As one of the largest providers of internet, voice, and media products and services to business and residential users in the US, this media conglomerate’s employees are responsible for handling over one million customer interactions per day. Its customers can access support services via many different channels - call center, website, or from a mobile device using self-service support; all of these channels roll up into the same customer support organization. The challenge was their ability to handle the large volume of events that were system-generated (from within their own infrastructure) or from users interacting with their hardware or software applications. A major objective for the business was to be able to automate as much of the support process as possible and to reduce support response times for their customers. Having the ability to access up-to-date customer account information such as who the customer is, what services they have, where they live, what’s their history, what’s the current state of the devices in their home, etc. would be key to improving the current event-based support model. By having this data available in near real-time, support can quickly identify and analyze areas where there may be problems with a service or product.

The business use-case was for a large-scale distributed data grid that would have access to real-time device telemetry data from routers, set-top boxes, mobile devices, and applications in addition to all the customer account information from their back-office business systems. This was achieved with the internal development and deployment of a proprietary support platform that leverages artificial intelligence (AI) and machine learning (ML) technologies to augment natural language processing capabilities (NLP). It relies on Hazelcast IMDG (In-Memory Data Grid) to access large amounts of stored, unstructured data, to deliver customers, support agents, and AI chatbots near real-time information to improve the self-service or support experience. The support platform is a hive-mind that uses telemetry data that captures who the customer is, what services they have, where they live, what’s their history, what’s the current state of all the devices in their home, etc. When a customer contacts the support organization with an issue, all relevant information is processed by the support platform to offer the right recommendation to resolve their issue in real-time. The system is expanding steadily and usage and volume have increased significantly across the organization. The forecast is for tens of millions of accounts on the system, with dozens of data sources per account. Presently, they are using Hazelcast IMDG to store transient data in AWS related to their services and products and they are handling about 300K customers a day. As they expand their reach across all of their customer service properties, the platform and Hazelcast IMDG are capable of delivering the scalability to handle the increased transaction volumes.

• Hazelcast IMDG enables all this information to be instantly available, reducing the interaction time with support, which allows them to handle more interactions with a higher problem resolution rate.
• This significantly reduces the instances of having to send a technician on-site and enables the service provider to offer the best customer service experience for its customers across the varied customer-facing service channels.
• The organization relies on Net Promoter Scores (NPS) to track customer sentiment and has reported a dramatic improvement from a negative to a positive score in this important service-centric KPI.
• The support teams have seen huge benefits since using the in-house developed technology for identifying and resolving set-top box errors and high-speed data internet issues.
• Tracking the kinds of errors that come up most often is one of the main sources of data used to train the AI/ML engine to identify major support issues.

• Handling about 300K customers a day.
• Forecast for tens of millions of accounts on the system.