Trase, a Northern Virginia agentic AI startup that only charges companies if its platform actually makes them more efficient, emerged from stealth this week with a $10.5 million pre-seed round.

Headquartered in McLean, Trase is focused on sectors where AI could be a big help, but complexities make it difficult to implement, like healthcare and national security, according to president Baskar Sridharan, who recently joined the firm after a career at AWS and Microsoft.

“These are the most complex industries,” Sridharan told Technical.ly, “and they are really overburdened with a lot of administrative tasks.”

The startup enters a booming field. Bethesda’s E:cue emerged from stealth in March and houses an agentic artificial intelligence platform designed to help go-to-market teams to better comprehend consumer data. Palantir, a national security-focused government contractor, also houses an AI agent platform for users to build assistants.

“[Trase] removes a barrier for enterprises which want to try to adopt AI but they don’t know if it’ll succeed or fail.”

Sridharan said Trase is unique for several reasons, including its business model. The startup doesn’t charge for services up front, and instead bills companies based on how operations become more efficient to make revenue. This is measured in a few ways, including hours saved, he said.

“That basically removes a barrier for enterprises which want to try to adopt AI but they don’t know if it’ll succeed or fail,” Sridharan said.

One customer, Duke Health in North Carolina, has used Trase’s tech to streamline prescription refills. The turnaround time for it is three times faster, per Sridharan, which results in $193,000 cost savings per year and 5,490 hours saved.

Trase, with a current staff of 30 employees, started developing its platform in the fall of 2024. It began by deploying its own engineers within other companies to have a full understanding of what was needed to improve operations, per Sridharan.

“We didn’t start with the platform,” he said. “We started with … understanding how humans interact with the systems in these industries, and then walk backwards from there.”