AI With a Plan: NRTC Sets Clear Goals as It Implements Call Center Systems

NRTC has been using artificial intelligence for the past three years to train its internet call center employees and improve service for NRTC members and their broadband customers. The effort led to clear improvements in efficiency and lower call center staff attrition. For anyone seeking to replicate that success, Rich King, chief operating officer of NRTC Managed Services, advises them to have clear objectives from the beginning.

“Everything we’re doing is meant to solve a problem. We’re not trying to do it just to be fancy and say, ‘We’re doing AI.’ We’re finding direct applications that are solving a problem or trying to make something better,” he said. “We don’t sit still very easily. We’re constantly trying to improve ourselves, push ourselves to get more efficient,” he said.

NRTC discovered a shortage of qualified IT talent three years ago as it attempted to fill call center jobs. Candidates suddenly were less educated, less skilled, and less experienced. “Covid changed everything,” King said. “We used to require that you have call center experience. I can’t do that today; I’d have no applicants. We used to compete against other call centers for resources. Now I compete against Starbucks and Walmart.”

The clear objective was to train new customer service agents more quickly and efficiently. NRTC turned to AI and automation to narrow the skills gap. The decision paid off by cutting average handle time by 69 seconds and cutting employee attrition by 32 percent over two years. Call center staff saw a 12 percent increase in quality assurance scores over a year.

New systems include:

• Learning Management. An online coaching system that allows techs in training to take practice calls on their own under many different scenarios. Practice sessions once required two people and took longer to complete. “We take previously recorded calls, load them up in the simulator and have techs interact like a live call,” King said. After the practice call, the system immediately provides a transcript, offers feedback, and grades the session.

•  Quality Assurance. AI is enabling us to evaluate and score every call for quality purposes. The learning management system has modules for 11 different call “attributes,” such as the tech’s courtesy, positivity, and empathy during the call. Using “conversation analytics,” the system quickly evaluates the tech’s strengths and weaknesses for each attribute. When techs score low in empathy, for example, the supervisor can assign them empathy training.

• Knowledge Base Search. AI continues to assist agents as they progress to live calls. NRTC maintains a database with detailed information about each internet client. In our customer service support operations, NRTC uses Google AI to scour the knowledge base and the member’s website to direct information to the agent during live calls.

In the future, real time AI could feature even more information during live calls. “If I could have a supervisor assisting a tech on every single call, I would. But that’s not feasible,” King said. “Eventually real-time guidance will get to a point where you have pop ups, or whispers in your headset saying, ‘Hey, look here,’ or ‘Ask this.’”

Future call center interactions could begin with an automated voice helping customers through self-service sessions. That will speed efficiency for routine calls, but what happens when automation has run out of ideas? Will there still be a human being available to answer questions?

King says that questions about the human element and job security are common as he meets with members and participates in industry meetings. “Our intent right now is just to augment, complement and supplement and find efficiencies while improving quality,” he said. “Right now, especially in rural America, you know that human-to-human interface, we believe, is still to be an integral part of customer service.”