T-Mobile Injects AI Into Customer Service
T-Mobile is accelerating automation of its customer service operations by deploying Tupl's Automated Customer Care Resolution (ACCR) tool, which utilizes artificial intelligence and machine learning capabilities to speed response time on customers' help tickets.
ACCR is one of several AI troubleshooting tools Tupl developed under the TuplOS platform; the ACCR tool expedites customer service cases by providing customer care representatives with detailed cause reports and technical resolutions. According to the announcement, the ACCR tool is "100 times faster and up to 4 times more accurate than legacy resolution methodologies, providing automation levels around 90%." (See Tupl Deploys AI Technology With T-Mobile US.)
"Tupl's tool has enabled us to respond to our customers much faster on technical issues," says Brian King, T-Mobile's SVP of technology service delivery and operations, in the release.
Tupl Founder and CTO Pablo Tapia says that in legacy systems a series of customer service reps, engineers and technicians may use over six different systems to identify the source of a customer issue once a ticket is opened. Tupl conducted several trials to measure ACCR's performance and determined the tool cuts ticket processing time down from several hours to less than an hour.
"[ACCR] offers the ability to create machine learning models and train these models using the engineers' knowledge," says Tapia in an interview with Light Reading. "It does it in a way that's very simple -- the engineers don't need to know about the underlying technology or machine learning algorithms."
During the trials, 80% of help tickets were closed automatically using ACCR. Of the remaining tickets, a resolution was submitted to the engineer detailing how to solve the issue. Tapia adds that the number of unresolved issues was reduced by a factor of four because the ACCR tool analyzes data on a more granular level than in legacy manual processes.
T-Mobile US Inc. is Tupl's first publicly announced wireless provider customer. But the startup's founder says Tupl has a service provider customer in Japan, as well as several customers in Europe, Mexico and Canada. The AI startup is headquartered in the Seattle area of Bellevue, Wash., with offices in Spain and Japan.
While Tupl predominantly works with carriers, Tapia says the company is looking to expand into the utilities market. Before founding Tupl, Tapia worked in strategy and engineering roles at T-Mobile, which is where he got the idea for starting his company.
"I realized there was going to be a big need for automation, and complex-process automation, which is hard to do without using some of the key technologies that we are employing, like big data and machine learning," says Tapia.
Tapia says when it comes to network automation, many service providers are still taking a conservative approach; "Applying automation means you have to take a risk or bet on something different than the status quo," he says.
Many operators have considerable complexity in their networks, typically push back on data access, are operationally siloed and have engineers wary of being rendered obsolete by automation, Tapia explains. In reality, automation tools will create new opportunities for engineers and "change the paradigm of what network maintenance and operations will mean in the future," he says.
"We are confident that [automation] is the only way to go for operators and sooner than later all operators will come to that realization."
— Kelsey Kusterer Ziser, Senior Editor, Light Reading