Streamlining Customer Contact with Uniphore
Uniphore is a conversational AI and automation platform that aggregates and analyzes call data to help businesses automate their customer support process.
The company, which was founded in 2008, provides three main services for call centers: Call analysis, real-time agent guidance, and conversational self-service. In particular, the platform is a pioneer in providing call analytics in real time—not just after the conversation has ended.
“We are able to actually impact the call while it’s still happening, while the call center agent is on the phone with a customer,” explained Christopher Daniels, Vice President of Sales at Uniphore. “Maybe they’re getting notifications giving them advice on what the next best action would be.”
Combining AI and automation across the entire customer journey, prioritizing agent performance, and enabling data-driven decisions are huge differentiators for Uniphore.
As a large utility, National Grid has a robust customer contact center. For this specific proof of concept, though, Uniphore focused on National Grid’s internal Employee Service Center, where employees call with questions regarding payroll, benefits, retirement plans, and other HR topics.
A common pain point for call center agents is the post-call process, which can be tedious and slow. At NG’s Employee Service Center, for instance, each call must be documented and categorized in the company’s case management system, complete with notes. However, call times vary significantly, and not every agent does their best work when they have to converse and take notes simultaneously.
“It really depends on the individual, because we have some that are really great at multitasking,” said Hilary Snyder, an HR business services manager at National Grid. “They can have the conversation with the person and type notes at the same time. Others might be handwriting stuff or taking quick notes – so someone might go into an after-call state for five minutes just to wrap up all their notes.”
For this small-scale proof of concept, Uniphore analyzed almost 1,000 historical audio calls specific to IVR menu options of payroll, scanning for particular types of language. Using this data, its software was able to identify eight call categories, demonstrating that it could theoretically take post-call analysis off an agent’s plate (and reduce the average call duration) if deployed live.
Daniels said that, based on the trial run, “The current average after-call work could be brought down by 30, maybe 40 seconds.” That saved agent time could be reallocated to other value-added activities.
“The agents are the face of your company. They’re the ones interacting with your customers, who are providing the revenue for the business every single day. How do we make their jobs as enjoyable as possible and make it as low-risk as possible?”
National Grid remains interested in automating aspects of its customer support—not only streamlining the post-call process, but also using tools like chat bots to answer simple questions fast.
“You don’t want to lose the human touch, but you want to make it more convenient for people, too,” Snyder said.
Despite its small scale, this proof of concept also has larger implications for the future of automation in call centers, especially those in the energy sector. Daniels posits that reducing agents’ workloads through automation will improve morale, as well as the image of the customer support job at large.
“The agents are the face of your company,” Daniels said. “They’re the ones interacting with your customers, who are providing the revenue for the business every single day. How do we make their jobs as enjoyable as possible and make it as low-risk as possible?”
For utilities like National Grid, AI resources may also help agents identify—in real time—customers who might be interested in additional money-saving services.
“By picking up on customer cues that maybe a traditional agent would miss, our technology can identify that this hypothetical customer, based on their profile, based on the conversational cues they just made, would be a perfect candidate for a home energy assessment,” Daniels said. “Let’s make that offer.”
All in all, the goal is to create a system that works more efficiently for both agents and callers—and ensures National Grid can continue to deliver excellent service to its over 20 million customers.
