Treasury Department contact centers are spending 80,000 hours each year on calls that should not have required human interaction, according to Management and Program Analyst Jennifer Hill, but now the agency is using employee and citizen feedback to find emerging tech that can fix that.
Hill shared the story of Treasury’s innovation team at ACT-IAC’s Jan. 22 Artificial Intelligence (AI)/Intelligent Automation (IA) forum. She said that discovering solutions to the call center conundrum taught the team the “right way” to apply technology.
After successes with RPA and blockchain, the agency wanted to implement an AI-based chatbot to streamline its contact centers. The innovation team noticed that 75 percent of the four million calls and emails that get sent to one of Treasury’s contact centers each year are misdirected. Of the other 25 percent, the majority dealt with tasks that the caller could have done on their own, like filling out forms.
“For us, mapping out that customer journey and identifying where those pain points were … really helped us look at where [technology] makes sense,” Hill said. “This has really been a big culture change for us.”
Now, Treasury is now building out the knowledge base for conversational AI to tackle some of these contact center frustrations with a human-centered approach. As the innovation team begins this process, they’ve found that preliminary AI has opened the door to new AI use cases.
“AI actually enabled AI,” Hill said. “We can actually use natural language processing to pull calls out and actually do some analysis on them to figure out what are our use cases, like where does it make sense to apply technology? It also helped us identify how many things are coming in that we could actually make it where calls are never needed.”
Human-centered design at Treasury is also starting to go beyond the contact centers. Updates to its website design, for example, are also taking into consideration the customer’s perspective and knowledge.