Government leaders are uniform in their view that, while AI is inherently a technological process, it should be approached first and foremost as a business solution. As Federal CIO Suzette Kent notes, "there are so many things that we can do with the spectrum of automated technologies but matching it to the right problem still means you have to understand the citizens you're serving and the mission that you're trying to accomplish."
Media coverage has helped to drive this thinking, with news stories demonstrating the powerful impact this approach is having in driving private-sector business improvements. To that extent, much of the current thinking around AI in government is in fact being driven by business-line leaders. That’s as it should be, but it doesn’t mean that IT can sit on the sidelines.
Even with AI built around specific business needs, there will be technical challenges when it comes to implementation. How do we manage all these AI bots? How do we make sure we're refreshing them when we deploy a new system? IT needs to engage early and often in order to ensure business leaders don’t take it into their own hands to develop exquisite, siloed AI solutions. IT will work hand in glove to ensure these business-centric systems can be managed and maintained effectively.
Improving CX: A Common Priority
Using AI to improve user experience is viewed as one of the most promising aspects of AI. As Kent notes, "what's exciting about many of the newer technologies, whether it's AI or how we use data, is that we can actually leapfrog [current performance] with improvements to customer service and citizens, who set their expectations based on the way they live their everyday lives."
The low-hanging fruit here is frequently the chatbot or automated assistant. While early iterations have been limited in scope, AI promises to make the chatbot a powerful tool for citizen engagement. There’s a multi-faceted benefit here. Chatbots can lighten the workload on federal employees, handling many routine questions. They can also deliver the kind of self-service environment modern consumers have come to expect in the commercial world. Finally, the interactions with chatbots can be fed into the AI application to help agencies better understand citizen needs.
For example, Courtney Winship of U.S. Citizenship and Immigration Services says, "what we're trying to do is analyze intent and try and look at things like comments, common themes that I think will allow us to kind of discern what types of questions or concerns people have. So things like case status or processing times for specific forms: We've been able to do a lot of analysis that allows us to train our backends to respond to the questions in a way that should be relevant to the users on the front end."
Early experiences with chatbots have helped to move the needle in favor of data. Federal leaders increasingly view user data as a driver of good customer experience. It’s not just about having more data, although more data generally does drive better outcomes. Rather, feds recognize AI as a means for making best use of the data already in their care.