Planning or Piloting AI? Four Things You Need to Know

Date: May 22, 2019

Although it is safe to say that 2018 was the year of Robotic Processing Automation (RPA), we have now set our sights on yet another technology that everyone is talking about: Artificial Intelligence (AI).

AI is computer software with a mechanism to make it possible for machines to learn from experience, adjust to new inputs and perform human-like tasks. If you have an e-mail account you’ve already experienced AI…just check your spam folder. If you are asking Siri the Game of Thrones series finale theory, you’ve experienced AI. If you are checking a bill and get automatic responses on how much you owe, you’ve experienced AI. It is everywhere and not going anywhere anytime soon.

But for government there is significant learning that needs to occur before we start taking advantage of AI-based technologies. That’s what our office set out to do: develop a better understanding of AI and the problems it can solve. Here are the top four lessons we have learned so far in our AI journey:

Lesson #1 – Avoid putting AI at center stage. We started our project looking for opportunities to pilot AI— by integrating AI-based solutions (specifically chatbots) in Fiscal Service’s website and its contact centers. We quickly learned this was the wrong approach and that AI, like RPA, blockchain, or other technologies are a means to an end – not the end itself. By adjusting our lens and focusing on the customer, it became evident where AI would add value and where it wouldn’t. We learned that if we stay focused on the problem we want to solve or the opportunity we want to exploit, the right solutions will show up.

Lesson #2 – Think at an enterprise level. A powerful feature of AI-based applications is being able to identify patterns and make predictions through new connections, but we’ll need to break down the government silos first. For starters, having an AI-based application that is used by more than one work unit increases the agency’s return-on-investment for that technology and makes good financial management sense. More important though, is the value the agency can derive from connecting disparate systems and seeing connections and patterns that would have otherwise been nearly impossible to see.

Lesson #3 – Build a strong foundation (which doesn’t start with technology). Our project taught us something much more basic: For us to use AI to improve the customer experience we would need to create new habits and routines that in the end become part of the organizational culture. This includes creating the right structures and incentives, also coalescing around and promoting shared objectives for improving the customer’s experience. Although there is no doubt that AI and other forms of technology and automation will play a big role in improving the customer journey, working with the technology may be the easiest nut to crack.

Lesson #4 – Managing a “digital workforce” is in its infancy. Beyond the technical knowledge and maturity of emerging technologies, there is a management piece to this puzzle that is still a bit elusive. Like a human workforce, digital workforces will only be as good as we develop them. Decisions around how this new workforce will be managed, organized and by who needs to be a priority for any entity thinking about using AI-based applications. Although we lack a well-tested operator’s manual for how to design such a workforce, giving some serious thought to how this new workforce will be managed will likely prevent many headaches down the road.

For more information on FIT’s innovation journey, read our other blogs or send an e-mail to FIT@fiscal.treasury.gov. To find out more about the Fiscal Service, visit fiscal.treasury.gov or follow us on LinkedIn and Twitter.