Agencies Shift from AI Experimentation to Integration to Boost Efficiency

Agencies are moving beyond the conceptual phase of artificial intelligence and transitioning to mission- and efficiency-focused implementation, officials from the departments of Commerce and State said this week.

“Budgets are constrained … we had to pivot into areas that would provide clearer outcomes … that are not only focused on mission, but really on streamlining operations.” Commerce CIO and Chief AI Officer Brian Epley said during the Equinix Engage: Accelerating Government Efficiency event in Arlington, Virginia.

Epley added that his agency encompasses many diverse missions within Commerce subagencies like the Census Bureau and the National Weather Service, which requires a shared set of principles to guide the agency’s AI evolution. He added that Commerce is moving from siloed experimentation to enterprise integration to best serve each operational unit in the agency.

“It’s not just one mission. I’ve got 12 agencies with state missions. The underpinning is obviously economic, security, growth,” said Epley. “How are we moving from experimental research aspect? I think that’s never ending.”

At the State Department, success hinges on avoiding fragmented, isolated efforts, according to State Department AI and Automation Team Lead Tim Ahrens. The department created “enterprise AI campaign teams” with senior leadership buy-in from across the organization.

“The teams have senior leadership representing a broad rush across the organization and center in on a use case that you can scope focus and deliver in a timely fashion. That is mission critical,” said Ahrens during the Equinix event. “If you do that, everyone’s going to be behind it, and it will be a success.”

A Focus on Data and Security

Epley said that data is the “richest and probably the most powerful” asset, but also the most challenging to manage. Commerce’s components possess vastly different data sets, making curation, standardization and maintenance a continuous effort. To address this, he said the agency emphasizes “security by design” and zero-trust architectures to protect sensitive information, whether it resides in the cloud or on-premise.

“[Data security] forces us to look at the new perimeter and how those things push out, and how it has an impact on the data,” said Epley. “You’ve got to protect the data in so many different ways, not just in transit, not just where it lives, but how you’re going to continuously maintain your data.”

Ahrens said that collaboration is a key tool in State’s AI data security strategy. The agency has data at different levels of classification and security, and State is learning from other agencies how to manage it securely.

“Some of our interagency partners are further ahead than we are,” said Ahrens. “We’re coordinating closely with some of our [War Department] partners to leverage classified cloud environments that they’ve already stood up, modernized and authorized … rather than endeavoring to stand it up from scratch.”

Epley added that agencies throughout government are using AI to further mission. They’re learning from one another for better return on taxpayer dollars, he said, leading to more innovation.

“We are looking for the common values in those frameworks to identify the biggest gain that we can have is actually in the operations,” said Epley. “It increases performance, it increases productivity. It shows an ROI on the dollar to spend, and that allows us to further explore experimentation and mission impact.”