T-Mobile CEO Mike Sievert said the IntentCX AI service, developed together with OpenAI, will help T-Mobile better understand why some customers have left and will automate engagement with current customers.
T-Mobile and OpenAI are teaming up to take on customer service with a new AI platform dubbed "IntentCX." T-Mobile CEO Mike Sievert said IntentCX will help T-Mobile better understand why some customers have left and will automate engagement with current customers.
Sievert explained that T-Mobile will use AI to "ascertain what went wrong and why and actually prevent those problems" that lead to customer churn. In 2022, T-Mobile received 3 million calls to its customer service centers each week, and "every one of those [calls] is an abject failure on T-Mobile's part to prevent the problem you're calling about and that is the unlock in AI," he added.
"For the first time ever, because we've invested in redefining our data estate and putting our network data, along with our commercial data, into a place where AI can be trained on both simultaneously, we are able to see not just what those network experiences are, but how they subsequently comport to customer satisfaction and performance and success with T-Mobile," said Sievert at T-Mobile Capital Markets Day.
Sievert acknowledged that "the wireless industry is so far behind on digitalization" but said AI can simplify the customer experience and further T-Mobile's goal to "earn every customer's business for life."
"I mean, we do everything else it seems digitally, and yet people spend a Saturday of their lives at a T-Mobile store trying to switch to T-Mobile," said Sievert.
Intent-driven decisioning
T-Mobile's "intent-driven AI-decisioning platform," called IntentCX, will use secure access to T-Mobile data to understand customer intent and sentiment in real time, according to T-Mobile. Scheduled for launch in 2025, IntentCX will then provide options to resolve issues and can proactively take action on behalf of customers.
T-Mobile's team-up with OpenAI will be used to create a "new kind of engagement model for customers, and this isn't with the intention of just slapping on AI to some existing business processes, but actually changing those business processes profoundly so that they take advantage of this transformative technology," said Sievert.
T-Mobile said IntentCX improves on the capabilities of current GenAI and next best action (NBA) services that work from a finite data set and a fixed library of customer resolution options. Because of those limitations, they only provide an educated guess for solutions and have limited ability to take action on resolving issues.
IntentCX will be trained in customer service by T-Mobile's Team of Experts (TEX) business process and supported by OpenAI's research and development team to access billions of data points from customer interactions. Customer data will be culled from their experience with T-Mobile's network as well as their use of T-Mobile's T-Life app, which also provides billing information, updates on Magenta benefits and rewards, and tools to manage T-Mobile subscriptions. IntentCX will provide personalized service and can respond in multiple languages, said the service provider.
In addition, IntentCX will connect with T-Mobile's transaction and care systems to proactively identify customer needs and take action with the customer's permission. The AI service will also analyze T-Mobile's network and service data in real time while working with customers that contact the carrier.
T-Mobile also said IntentCX will use "the highest level of privacy and security measures across every transaction." The AI service will also continue to improve as T-Mobile integrates with OpenAI's newest AI models.
OpenAI's take
OpenAI CEO Sam Altman explained that while T-Mobile and OpenAI will create a custom AI model (IntentCX) using T-Mobile customer data, that data isn't utilized by OpenAI to train its own base AI models. Earlier this month, OpenAI released its latest AI model – OpenAI o1.
"Unlike previous models that are well suited for language tasks like writing and editing, OpenAI o1 is focused on multistep 'reasoning,' the type of process required for advanced mathematics, coding, or other STEM-based questions," reported MIT Technology Review.
On the topic of privacy for T-Mobile's customers, OpenAI's Altman explained that "one thing that is not well understood about OpenAI is that we don't train our models on API customer data. We can do custom models for you with that data. Or we can do, in context learning…but it's not data used for improving the base open AI models ever."
OpenAI has been steadily expanding partnerships with vendors as well. In June, Apple announced it would integrate OpenAI's ChatGPT into future iPhones, reported Light Reading's Iain Morris. But, Apple Intelligence, Apple's AI capabilities, won't be available in the new iPhone 16 until October when the beta version of Apple Intelligence is released.
You May Also Like
