HP Labs Blog

HP Labs

HP Labs uses Artificial Intelligence to improve HP’s customer service offerings

By Simon Firth, HP Labs Correspondent

January 15, 2020

Members of the Machine Learning Center of Excellence at HP’s R&D Center Bangalore with HP support agents

Members of the Machine Learning Center of Excellence at HP’s R&D Center Bangalore with HP support agents

A pilot project led by Niranjan Damera-Venkata, a Distinguished Technologist in HP’s Artificial Intelligence & Emerging Compute Lab (AIECL) and Head of AI research in India for HP Labs, is using Artificial Intelligence (AI) to help HP customer service agents solve customer issues with greater speed and efficiency, holding out the promise of significantly improving the HP customer experience while lowering support costs.

“The idea is to handle service calls in a new way,” explains Damera-Venkata. “If you think of the existing industry model as giving agents a set of road maps to help them work with a customer towards a solution, we’ve been experimenting with giving them a GPS app that can intelligently and efficiently guide them to where they need to go.”

The research effort teams HP’s Artificial Intelligence & Emerging Compute Lab with its R&D Center Bangalore, the company’s premier R&D unit in India and with HP’s global customer service organization. The core engineering team is part of the machine learning center of excellence at the R&D center, tasked with applying AI to transform customer service.  

The new approach draws on comprehensive data from hundreds of thousands of customer service interactions to diagram possible pathways to a solution for every kind of challenge that customers are likely to face. 

“In the GPS analogy, the program rides alongside the agent, looking at all the roads that they’re going down and, in real time, helps figure out the best route to get to the destination,” Damera-Venkata says.

Call center technologies already exist that use AI to help service agents understand the types of issues that a customer is facing and surface relevant documents. In the HP Labs solution, however, that intelligence extends to helping find the most efficient pathway to a technical resolution as well. That’s a significantly more complex challenge.

“This pilot is one of several that are pointing to the very concrete ways in which AI looks set to positively impact the quality of both the products and services that we are able to offer.”

Niranjan Damera-Venkata, Distinguished Technologist and Head of AI research in India for HP Labs

Testing HP Labs’ new customer service technology in HP’s Bangalore Customer Call Center

Testing HP Labs’ new customer service technology in HP’s Bangalore Customer Call Center

“It’s one thing to create a program that can understand, for example, that a caller wants to return a piece of clothing because it’s too small and then automatically retrieve the relevant sales page to look for a larger replacement,” notes Damera-Venkata. “But troubleshooting a technical issue can go in a myriad of different directions depending on the actual underlying cause of the issue the customer is calling about.”

The joint HP research team has tested their AI-based navigation technology in thousands of live customer interactions at HP’s Bangalore Customer Call Center, which handles calls from customers across much of the world. They have so far established that the system can offer accurate guidance in 75% of troubleshooting calls, allowing the agents using it to address and resolve issues both faster and more efficiently than before.

“There’s also a lot employee turnover in the customer service industry,” Damera-Venkata adds. “We could use this technology to standardize the quality of support and even to train new employees offline in a simulator mode, helping them to work more effectively as soon as they start talking with real people – this lowers the cost of training new hires and reduces agent burnout.”

While the project has already shown promise as a pilot, it stands to gain in both speed and accuracy once it can “learn” from the millions of calls that HP resolves annually. It will also be able to track issues that are slowing agents down in real time, flagging these as problems that need better answers and alerting product teams that there might be a problem that they need to address.

From a broader research perspective, the project furthers HP Labs’ general interest in finding new ways to optimize service delivery. Services can be improved by diagnosing issues faster, for example, or by predicting when and where issues might occur so that spare parts can be ordered in time to keep operations running.  

“Our lab is asking how AI can be used to achieve these kinds of optimization,” says Damera-Venkata. “This pilot is one of several that are pointing to the very concrete ways in which AI looks set to positively impact the quality of both the products and services that we are able to offer.”