Nokia is using machine learning algorithms developed by its Nokia Bell Labs to give computers the ability to learn without being explicitly programmed. It’s using the new capabilities in its customer experience portfolio for its Motive Service Management Platform (SMP) 7.0 and Motive Care Analytics (CAL) 2.0.

Earlier this week, Nokia’s CEO Rajeev Suri announced the company was establishing a separate software business unit. Asked if the new machine learning software technology announced today will be part of this new business group, Rich Crowe, Nokia’s head of marketing for customer and network operations, says, “It already is. Motive SMP 7.0 and CAL 2.0 come from Nokia’s Applications & Analytics business group, which was established at the beginning of the year.”

Nokia is gunning to compete with major standalone software companies, whose margins it envies. Today’s announcement “is in line with Nokia’s software portfolio strategy to deliver programmable operating systems and augmented cognition systems,” adds Crowe. “With each element delivered, Nokia wants to underscore what makes us different from pure-play software vendors: our understanding of the network.”

The goal behind today's announcements is to reduce help desk calls due to outages by 85 percent and shorten the average call-handling times by 5 to 15 percent. Nokia also aims for the technology to eliminate inappropriate truck rolls — where a service technician is dispatched to a customer location, but the problem actually stems from a network outage not at the edge.

A new self-optimizing system within SMP 7.0 determines the ideal sequence of tasks that deliver the highest probability of resolving billing, subscription, and network service issues in the shortest amount of time. Machine learning software determines this ideal sequence by analyzing data from previous similar tasks, the network, customer premises equipment, and trouble tickets.

Nokia’s CAL 2.0 correlates customer help desk calls and self-care actions with network and application topologies to identify call anomalies. For example, an unusual pattern in help desk calls could indicate the location of network problems impacting customers. Once anomalies are identified, CAL 2.0 initiates actions to resolve service disruptions and other issues.

Crowe says Nokia’s machine learning software can be deployed either in the operator's data center or in a Nokia hosting center. GPUs are not required. Customers can scale the number of servers up or down as needed to handle the workload.

Machine learning was already in the news earlier this week when Google announced its new Google Cloud Machine Learning Group, which will focus on delivering cloud-based machine learning software to businesses.