AI and RPA Coming for Understaffing, Not Staff

Machine learning AI modernization

A panel of Federal leaders agreed today they aren’t worried that artificial intelligence (AI) and robotic process automation (RPA) technologies will leave Federal employees out of work, rather, they see the technologies as an opportunity to help an overloaded workforce further agency missions.

“I’m in a place right now where I’m so understaffed to do the job that I need to do, that I’ve got people leaving stuff undone at the end of the day,” said Bill Pratt, director of strategic technology management at the Department of Homeland Security at an event organized by Government Executive. “It’s really not, to me, an issue of us fearing AI taking over. We desperately need it, because we’re already behind the 8-ball.”

“In our case, it’s not really a situation where the staff are concerned and don’t see themselves in the picture. They’re actually the ones driving us, from a shared service perspective, because it allows us, as an administrative shared services provider, to take on additional customers, hire additional people, helps us lower our costs, and increase our efficiencies,” said Marisa Schmader, assistant commissioner for fiscal accounting in the Office of Financial Innovation and Transformation within the Bureau of Fiscal Service at the Treasury Department.

Acquisition emerged as a common theme among Federal RPA and AI pilots, as agencies try to free up professionals to provide more service to industry partners and agency customers.

“GSA is pretty forward leaning in this area. They have an emerging technology division, they’re using RPA for things like compliance of solicitations,” said Joanie Newhart, associate administrator for acquisition workforce programs at the Office of Management and Budget. “The Defense Logistics Agency is doing a pilot around an automated ordering tool. How cool would that be? If you need something, you don’t have to go interface with an actual person, you can go online like you do at home. Imagine that,” she joked.

Newhart also pointed to the Department of Health and Human Services and its recent efforts to use emerging technology, and said the agency has a demo scheduled for next week.

Pratt described how an acquisition reform project highlighted the need for AI and RPA. After conducting a pilot to loosen acquisition rules around some major procurements, the department’s recommendations for reforming acquisition included adopting more emerging technology. “We would love to see some great pilots from our vendor partners” to flesh out the right use case and technology for a pilot, he said.

Schmader offered some pieces of advice for prospective projects.

“Within Treasury … we went through a series of pilots to get our arms around it, and I want to share a few tips for anybody who wants to start down that path. One, start small. Carve out something you feel you have well documented–it doesn’t have to be perfect–that you think you can see some value-add from,” she said.

“The next thing I would suggest is to focus on your infrastructure. Make sure to engage your IT and your security staff very early on. We did that, and it was helpful, but I think we could have pushed it more, to make sure they were even more engaged, especially from a security perspective. Make sure you start with a solid foundation because once you start doing this and people can see what you’re able to do–there has not been a crawl-walk-run for us. It’s been crawl-run-run faster,” Schmader said.

One of the main concerns for agency pilots is to demonstrate the return on investment (ROI) to take on further projects.

“I want to show the value of all the things that we did … but as of right now, we can’t show the ROI. I want to get some really good data analytics so we can start building this story to show the benefit,” said Pratt. He expressed frustration at not being able to get funding to run the pilots, but needed proof of ROI from the pilot to get started.

“It’s a tricky wicket,” added Newhart.

Recent