MultiChoice South Africa taps Watson to support customer service staff

Share this post:

MultiChoice South Africa WatsonGreat customer service is essential for any business to prosper and grow.

In a fiercely competitive marketplace, such as the South African media broadcast sector, it becomes even more important to provide your customers with speedy, efficient service whenever they seek advice or need to solve issues with their subscriptions.

At MultiChoice South Africa, customer service is one key element of our strategy to thrive in a challenging new era. The call center is at the heart of those plans. We have more than 7 million customers and rely on more than 1,500 agents to answer enquiries. It is an intense operation that requires months of training before a new agent is ready to take the first call. And even once an agent is fully skilled and ready to assist customers, the ever-changing landscape of video entertainment means that training and upskilling are constant requirements.

MultiChoice South Africa must stay as agile as possible and be ready to rethink schedules, develop new content services and, most of all, ensure existing subscribers enjoy a great quality service when interacting with us. Always on the lookout for ways to better support our call center agents, we started to explore how new technologies could help the process.

Deploying an AI-based chatbot

IBM Watson Assistant in the IBM Cloud was the perfect solution. MultiChoice South Africa worked with the IBM team to create and train a smart virtual assistant named TUMI (The Ultimate Master of Information). The assistant enables call center agents to enter questions during customer enquiries in natural spoken language, and TUMI pulls up the relevant information and answers in real time.

During the initial proof of concept, we set the Watson chatbot a simple challenge: answer queries faster than our best call center agents could. The IBM Services team trained an early version of TUMI with 30 service requests, and the results blew us away.

Since then, MultiChoice South Africa has continued to train the chatbot, placing it in the hands of our 20 most experienced agents to answer questions day in, day out for a 10-day period, providing key insights into areas for improvement. We’ve now trained the virtual assistant to recognize and respond to 2,000 separate customer service enquiries.

Faster, better customer service

Since its launch in the call center, the virtual assistant has been a big success. TUMI now provides call center agents with answers to 200,000 enquiries every month. It now offers correct responses to over 70 percent of all queries, an exceptional result considering we haven’t told agents what can be asked of TUMI. So questions like “Hey TUMI, who is the president?” are commonplace!

Our agents can now access the critical information they need during calls through TUMI, so we are completing calls 20 percent faster than previously, a key advantage that contributes to greater subscriber satisfaction, which is essential as the media market grows ever more competitive in South Africa.

The virtual assistant has also helped us to significantly shorten training times for our call center agents by over 60 percent. Previously, it would take six to eight weeks before a new agent would have the knowledge and skills needed to field calls. Today, thanks to the presence of TUMI, this takes just two weeks, a major win that also enables us to reduce our operational costs significantly.

Read the case study for more details.

More Cognitive stories

Chatmantics improves customer experiences by replacing call center IVR with AI on IBM Cloud

We’ve all called a company’s customer service number only to be greeted by an interactive voice response (IVR) system. The IVR will say to press one for this and two for that and so on. You must wait for the correct number to enter your account information “so that we may better serve you” but […]

Continue reading

Comprehensive strategy for enterprise cloud: Where power meets flexibility

Enterprises have embraced the cloud: according to the 2018 IDG Cloud Computing study, 73 percent of enterprises surveyed now host at least a portion of their enterprise computing infrastructure on the cloud, and many have lifted and shifted their simplest workloads to the cloud. But recent survey data also reveals multiple concerns about security, vendor […]

Continue reading

How an AI application is helping improve quality control in the egg farming industry

People typically open a carton of eggs before buying it to be sure none of the eggs are cracked. No one wants to deal with the mess of cracked eggs. For egg farmers, one “bad egg” in a carton means the retail store must be credited for the entire carton. That could potentially waste as […]

Continue reading