The Commerce Department is encouraging employees across the agency to develop their own artificial intelligence (AI) solutions, an approach that a top technology official said is already producing enterprise-wide benefits and accelerating AI adoption.
Speaking July 14 at MeriTalk’s Shift Happens event in Washington, D.C., Matheus Passos, acting chief technology officer and deputy chief information officer (CIO) at the Commerce Department, said the agency has rolled out agentic AI across mission support functions.
At the enterprise level, Passos said Commerce is initially focusing on common business functions such as human resources and accounting, while working with early adopters to expand AI use over time.
To help coordinate those efforts, Passos said Commerce established a department-wide AI integrated product team (IPT). Initially, the AI IPT was meant to facilitate compliance with Office of Management and Budget mandates, but he said it has since grown into a broader AI community.
“That became a strong community,” Passos said. “We have a champion, basically, of AI [at] each bureau, and that person has the authority of the CIO to deal with AI things, and they’re meant to push. So people are going to start doing it on their own.”
“What we did is we require all AI to be documented and logged at the enterprise level, and we make that very well known,” he added. “What happens as soon as people start doing it, is they reach out. So, by process, they reach out, and then we prioritize our resources to facilitate for them.”
Passos said the approach has helped surface AI projects developed by employees outside of dedicated AI roles.
“We had some of our folks in our AI community [who] started using AI to help deploy AI across the enterprise … using AI to test AI, to track cost, to develop costing models, to develop performance metrics for it,” he said. “And it was surprising because these are folks that have [other] daytime jobs.”
For example, Passos pointed to one Census Bureau employee who used AI to build a tool for tracking and deploying AI across Commerce.
“Using AI, [he] created a tool to help track and deploy AI across the whole enterprise, and he shared with us within two weeks,” Passos said. “We tested it, and we had it available to the community, and every bureau was using it almost immediately. It cost us nothing, except maybe his Friday night, you know, with a glass of wine. I don’t know how he put it together, but it was great.”
Additionally, Passos said employees are using the technology to quickly create resources for the broader workforce.
“One of our detailees in 45 minutes just put together a new AI video that’s going to be the basis for our whole enterprise training,” he said. “I’m thinking, wow, what can you do in 45 minutes? I can’t check my emails in a day in 45 minutes, right? Unless I use AI.”
Egon Rinderer, senior vice president of federal and enterprise growth at NinjaOne, added that agencies should focus AI deployments where success can be objectively measured rather than simply tracking AI usage.
“I think where you can safely start implementing [AI] is anywhere we have previous units of measure that we can then compare outcomes against,” Rinderer said. “In other words, I can see that this outcome has changed for the better.”
While AI continues to reshape government work, Passos said he does not expect the technology to make employees obsolete. Instead, he said agencies should focus on helping workers learn how to use AI effectively and empowering them to improve the processes they know best.
“I’m less worried about the retraining piece and more about reshaping the expectations and letting them understand that they, with better tools – which AI is just a tool – have a better voice on their workload, on their processes,” Passos said. “That’s what I would ask them to do.”