Nokia beefs up software offering with machine learning-powered bots

Resolving service issues before they even hit a subscriber’s phone or tablet sounds like a dream come true, and it’s exactly what Nokia is promising with its new Autonomous Care capabilities.

The Finnish vendor announced that beginning in the third quarter, it will offer the Nokia Autonomous Care software, providing deep machine learning capabilities to help resolve service-impacting issues before they happen.

If that sounds like a tall order, Nokia Bell Labs has some numbers to back it up. Working with an unnamed Tier 1 service provider in the Americas, Nokia found that machine-learning capabilities could predict and resolve up to 70% of residential issues before the subscriber is even aware of a problem. Other Nokia data found that the software could handle up to 80% of care issues without customer support agent intervention.

“Our thinking is, if you can have a self-driving car these days, why can’t you have self-driving customer care?,” said Rich Crowe, head of OSS Marketing at Nokia.

As telecom service providers move to become digital service providers, their customers’ expectations are on the rise because they’re used to services from webscale companies that just work. “Communication service providers need to move into that space where things just work,” he told FierceWirelessTech.

“That’s really what we’re enabling for service providers with our autonomous care solution that we’re offering,” he said. “What we’re introducing is machine learning-powered bots” with machine learning algorithms provided by Bell Labs that are able to predict a service outage before it happens and resolve it before the customer is impacted.

“We believe we’re the first telco vendor to announce this kind of autonomous proactive self-driving care solution,” he said. “We believe that in the future, most issues are going to be handled in this way.” 

What about the 20% or 30% of the time when actions are not predicted? In those situations, he said, the bots are also available to help agents and customers. Nokia offers augmented care in which the bot supports the agent and subscriber by doing some basic tasks for them, such as looking up device settings.

Nokia's software works with any commercially available assistant, including Apple Siri, Amazon Alexa, Microsoft Cortana and Facebook Messenger, to name a few. Nokia has its own MIKA—or Multi-purpose Intuitive Knowledge—targeting telecom network engineers, but that's not tied to this deal.

Nokia also announced Cognitive Analytics for Crowd Insight, a new software application using Nokia Bell Labs’ machine-learning algorithms to track and analyze the aggregate movement of subscribers using real-time network data instead of GPS or application data. It's supposed to open additional revenue streams by allowing service providers to operationalize their data, such as helping retailers identify the best high-traffic areas for new stores or allowing a city to identify the optimal location for bus stops, according to Shelley Schlueter, head of Analytics Marketing at Nokia.

“What we’re offering to operators is a way for them to create new revenue streams to vertical markets using their existing telecom data,” she said. For instance, “we are starting the process by gathering information from the telco network, but we are respecting privacy, we are anonymizing that data before anything else gets done to it. So while the privacy restrictions in the U.S. have been lifted, we are a global company and we sell worldwide so we have to make sure that privacy protection is in place.”

Nokia doesn't intend to remove the customer service agent entirely, though. “There’s always going to be room for the agent,” Crowe said, but the human touch is reserved for the more challenging tasks. “There’s always a role for the person but we want to offload the basic troubleshooting and make that as quick and painless for the operator and for the customer as possible.”