A Blog by Jonathan Low

 

Oct 1, 2019

Sprint Cuts Customer Service Calls By 30 Percent Using Chatbots

Whether Sprint customers are as enthusiastic about this development as is management will be interesting to learn. JL

Patrick Kulp reports in Ad Week:

30% of online chats are now conducted solely by machine-learning bots, up from 4% at the beginning of 2019. “Chat use help customers resolve their questions at a good experience rate and allow human reps to handle the more complex questions.” Sprint trains the bots on data from actual conversations between customer service reps and callers, an ongoing process that allows bots to hone their skills. The company has also been using machine learning to adjust the design flow of its website through A/B tests and targeted personalization based on which types of experiences most often leads to sales for a given profile of consumer.
Around a year and a half into Sprint’s efforts to overhaul its call center operations with artificial intelligence, the carrier says the technology has substantially reduced the number of customer inquiries its human representatives need to field.
Sprint chief digital officer Rob Roy said about 30% of online chats are now conducted solely by machine-learning-powered bots, up from about 4% at the beginning of 2019. The chatbots are one prong in a continuing digital transformation the company has undertaken in partnership with Adobe, which has also included tools that transcribe and analyze calls in real-time to suggest next steps for reps and algorithms that inform advertising and in-store sales tactics.
“For us, that’s a massive amount,” Roy said of the chatbot volume. “Chat use cases help customers resolve their questions efficiently, effectively and at a good experience rate and allow our human reps to handle the more complex and harder questions.”
Sprint trains the bots on troves of data from actual conversations between customer service reps and callers, an ongoing process that allows the bots to hone their skills automatically. The company has also been using machine learning to adjust the design flow of its website through A/B tests and targeted personalization based on which types of experiences most often leads to sales for a given profile of consumer.
Roy said this fast-tweaking, data-heavy approach evolved from a turning point that came about 14 months into Sprint’s digital transformation project, which kicked off in 2015. The team decided at that point to move from prioritizing flashy new technology to scaling up data analytics systems that serve all of the company’s various divisions.
“We created a new way of working within the company that focuses less on bringing in shiny tools—those are nice—but really focuses in on what we as a digital team can do to help the rest of the company become more efficient and smarter,” he said. “We wanted to leverage the great things that are about digital in terms of real-time capture of insights and segmenting and personalization and teach the rest of the company how to use those within their their purview.”
Much of this process is powered by Adobe’s AI engine, Sensei, and software included in Adobe’s Experience Cloud bundle of marketing tools, including Adobe Analytics and Adobe Target.

0 comments:

Post a Comment