Best listening experience is on Chrome, Firefox or Safari. Subscribe to Federal Drive’s daily audio interviews on Apple Podcasts or PodcastOne.
As one of the leading agencies in the Trump administration’s campaign to improve customer service, the Veterans Affairs Department sees potential in using artificial intelligence to get a handle on the hundreds of thousands of calls it receives daily from veterans seeking care.
Rosetta Lue, the senior contact center adviser at VA’s Office of Information and Technology, called 2018 the “year of execution” for many of the customer experience priorities the agency has lined up.
Lue, speaking Tuesday at the Digital Government Institute’s 930Gov conference in Washington, said VA has made human-centered design, or trying to solve problems based on feedback from users, the centerpiece of their customer service transformation.
“What we’re finding is, we’re rolling out the technology. We’re rolling out the employee engagement programs, but we’re doing things not the way that fits us, but really the way that they’re telling that they want those problems to be solved,” Lue said.
“When a veteran comes to VA, it is not up to him to employ a cauldron of lawyers to get VA to say yes. It is up to VA to say yes to the veteran,” Wilkie said. “The issues I encountered when I was acting secretary were not with the quality of VA care, but with just getting our veterans through the door to get that care. Those problems are administrative and bureaucratic.”
In order to make sense of all the data that’s coming through VA’s call centers and websites, Lue said the agency has looked at using artificial intelligence tools to handle more routine tasks.
“How do we take that data and begin to maybe automate a lot of the things we’re doing so we can better give an experience to the veterans and their families,” Lue said.
But in order to get the most use out of those AI tools, the VA needs to collect and store data in a more consistent, standardized fashion — basic cyber hygiene techniques that Federal Chief Information Officer Suzette Kent has stressed as a requirement to getting the PMA off the ground.
“At the end of the day, if you cannot collect accurate data — the AI, natural language — all these fancy terms that we’re using, it doesn’t matter. If you put data into a machine is learning things incorrectly, it’s going to impact your analytics. It’s going to impact the answers, the responses that they’re getting,” Lue said. “Besides the technology and websites, I really have to say, the place that we’re finding all those nuggets of information is in that data and that data management approach to analytics and AI, natural language and robotics.”
Keeping up with VA call center demand
The VA runs more than 1,800 call centers, which handle about 140 million calls a year.
A key challenge, Lue said, is keeping up with the volume of calls from veterans who want to speak with a call center operator, and not an automated message.
“There is a vast amount of people who want to speak to somebody — they want to speak to an agent, but we know that in all of your channels, that’s going to be your highest cost,” Lue said. “There are a lot of people who say, ‘I don’t want to talk to anyone. I just want to be able to call and get that self-service.’ So we’re tackling both. We’re saying let’s use data, understand why people are calling, why they’re contacting us online, get that information out there by using the data and automating it and data mining it and get them what they want, but also customize it so those folks, if they want to speak to someone, that we have agents available.”
While automation tools may help call center employees reach veterans more effectively, Lue said the VA still faces some workforce challenges.
“We are looking at, on the veterans experience side, the business side, do we hire more people, do we outsource?” Lue said. “What do we do to meet those demands, to meet those performance expectations?”