IBM’s Watson supercomputer made its debut in February on quiz show Jeopardy in rapid-fire fashion, but part of its post-game show future could be changing the telecom call center landscape.
The artificial intelligence system, touted for its ability to understand and constantly improve upon its comprehension of natural language, could alter the way operators interact with their customers, IBM’s CTO of telecommunications research, Paul Bloom, told Telecom Asia.
“The call center landscape could experience a changing paradigm due to Watson’s ability to understand the nuances of natural language,” said Bloom. “Watson improves with experience and learns from previous answers. I can see Watson helping service representatives do their jobs by providing answers more quickly than the manual process takes.”
Unlike most call center applications that require specific forms of input in order to provide answers, Watson will be able to comprehend the same question asked in various formats, said Bloom. “The customer’s question need not be precise, and Watson will still be confident and capable of asking a follow-up question. Operators can easily develop sets of questions that may not even require a service rep.”
Watson’s abilities combined with analytics could benefit any business that works with unstructured data, said Bloom. “Telcos probably have more data than any other enterprise; predictive analytics could analyze a customer’s social networks and call patterns so operators can perform micro segmentation of their customers down to the individual. How they monetize that info internally could make a difference.”
Bloom recalled a case where IBM’s analytics was implemented in South Africa’s MTN, to help the operator examine the characteristics of an individual’s social network. “Analytics was able to identify individuals who were leaders in their social networks, and subsequently predict who the operator had to target in a marketing campaign,” said Bloom. “The leader has a big impact on followers and if he or she should leave a telco, there’s a 2.5 times higher probability followers would churn to another company.”
The same set of analytics was subsequently replicated in several telcos across Asia, Europe and North America. “Harnessing the power of analytics and having telcos treat micro segments of customers differently could help operators with churn, a significant issue for many firms.”