This content is sponsored by Verizon.

When COVID hit, 40 million people were out of work, and the government needed to respond. But many federal agencies were not prepared for the workload that relief efforts were about to demand. One agency in particular did not have the staff to process the calls that were about to be coming in. So it turned to a technical solution to increase its capacity to respond to the needs of the public during this crisis.

That agency turned to Verizon. It needed to drastically scale its ability to field incoming calls from the public, and it needed to do it almost immediately. Verizon responded with Natural Language Interactive Voice Recognition (IVR).

“Natural Language IVR allows you to have conversations, so you can actually ask the question, where’s my check? When will I receive my check? Why didn’t I get my check? You can ask a barrage of questions. The agency already had web pages out there with questions and answers, but people like to talk to someone,” said Stephen Sopko, managing client partner at Verizon. “The artificial intelligence would actually go out to the web page, based on the question that was asked, and then it would parse it and go, ‘This is what we believe the best answer is.’”

And the AI learns, via a tuning process that helps it learn new ways of asking the same question. So it not only learns that “Where is my check ” and “When will my check be mailed” most likely have the same answer, but it also accounts for accents and language barriers as well. It does this by placing confidence scores on the answers, and then verifying whether it’s answered correctly. If not, it will flag the question for review, and prompt the caller to ask the question again, searching for variances.

That also helps the AI itself become more natural in its conversation flow. It can ask follow-up questions to narrow down the answer to a question. For example, if checks were mailed on different dates based on the taxpayer’s last name, it might verify the caller’s last initial in order to provide the correct answer. And all of this occurs in a free flow conversation, not a series of rigidly defined input options.

And Verizon was able to do all this extremely quickly. Normally, Sopko said, a buying cycle for a product like this takes six to nine months.

“We knew that this wasn’t just for this agency. This was for the people of the United States. People weren’t going to be able to eat, they weren’t going to be able to pay their bills. We had to be on top of this. For Verizon, this became all hands on deck. And I know people say that; we really meant it and did it and executed upon it,” Sopko said. “Our solution partner Nuance came together with us and our Verizon team, and they just did monumental things to get this up and running. We turned it up and live within two months. And millions upon millions of calls came in that the agency just wouldn’t have been able to answer.”

And they did that with a result of 70% containment, meaning 70% of the calls ended with the caller receiving answers they were satisfied with. Sopko said that’s a fairly high rate of success; government agencies are usually happy with 30-40% containment.

One thing that helped is during the process, Microsoft bought Verizon’s partner company, Nuance. That helped Verizon align with several key resources, such as putting the IVR in Azure, and using Microsoft’s QnA Maker product. This also allowed Verizon to piggyback on Microsoft’s FedRAMP authorization through the Joint Authorization Board process, allowing them to reduce the timeline to around nine months. Sopko said Verizon’s Natural Language IVR will be FedRAMP High certified within the next two months.

That’s important, because Sopko said the agency has more work for Verizon and its Natural Language IVR.

“They were so impressed with what we’ve done and the adoption process — we made sure we had folks that would help with the adoption of these technologies, because that’s a big thing to get the constituents to use it — that next year, this will go up and be live for standard agency operations,” Sopko said. “On a good day, during the agency’s busy season, it is able to handle about 30% of its calls. So the hope is the adoption comes onto that, the tuning process comes into play, that we start seeing a similar increase in percentages of calls being handled through natural language.”

Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.