cd /news/artificial-intelligence/tmf-wants-to-experiment-with-its-own… · home topics artificial-intelligence article
[ARTICLE · art-113581] src=fedscoop.com ↗ pub= topic=artificial-intelligence verified=true sentiment=· neutral

TMF wants to ‘experiment’ with its own AI model

The Technology Modernization Fund (TMF) may experiment with its own artificial intelligence model in fiscal 2027, acting Director Jessie Posilkin said Thursday at GovExec's Infrastructure and Modernization Forum in Reston, Va. Posilkin said the eight-person team within the General Services Administration (GSA) has been using USAi but sees opportunity for experimentation, including building a model based on all TMF proposals received and their outcomes. She also hinted at upcoming announcements, including fiscal 2025 savings figures, four new board members, and new investments at the end of September.

read4 min views1 publishedAug 27, 2026
TMF wants to ‘experiment’ with its own AI model
Image: Fedscoop (auto-discovered)

The Technology Modernization Fund may “experiment” with its own artificial intelligence model in the next fiscal year, acting Director Jessie Posilkin said Thursday.

TMF’s “tiny team” of eight within the General Services Administration is already all-hands-on-deck in managing and supporting the investments and proposals they have, she said. While the team has been using USAi, “there’s a ton of playing around we can do,” she said at GovExec’s Infrastructure and Modernization Forum in Reston, Va.

“I sort of flirted with this idea of … building some sort of model based on all of the TMF proposals that we’ve ever gotten and what’s happened with them,” Posilkin said. “I think, like any other agency team, I see a ton of opportunity for experimentation, and my goal for ‘27 is to get the capacity to start experimenting.”

The fund has been focusing on AI projects lately, like the State Department’s StateChat, which Posilkin said she was initially skeptical about because of the agency’s highly sensitive data.

“But that’s also the exact attitude that I think has led to sometimes government not taking important steps forward in using modern technology,” she said.

Posilkin said State’s approach to mitigate risk with small pilots and extensive testing is the “exact approach you want to see. How can we use the newest and best tools, and also mitigate for the very real risk that when you are working in government, we are talking about actual public services that we don’t want to cause problems with?”

While AI can help the government “move faster on delivering on the really incredible and difficult backlog of legacy technology,” she said, there is still a need for sophisticated engineers to challenge legacy code bases and be program managers who understand the complexity and sensitivity of the data. “AI cannot solve for any of that,” Posilkin said. “You still need really strong humans in the room to help manage those programs.”

Posilkin also hinted at upcoming announcements: savings and value return numbers for fiscal 2025 and news of four new members of the TMF’s board will come in the next month or so, she said. However, she does not know who will replace the board’s chief, outgoing federal Chief Information Officer Greg Barbaccia.

Also coming soon: the announcement of TMF’s new investments at the end of September, she said.

“It’s been pretty remarkable,” Posilkin said. “I can’t share who they’re going to be yet, but we have been really excited both by the investments that the board is making decisions on right now, and what’s in our pipeline. And to me, that’s actually the piece that feels even more exciting.”

While she did not share many details, she said the new projects are “pretty consistent” with previous work, which “reflects the fact that there is a huge unaddressed market for making improvements to federal technology” and shows a “tremendous amount of need.” She signaled that TMF may be bringing in AI coding, and is “excited” that at least one agency is “very seriously” looking at how to use agentic AI and how to safely expand it across government. After a “tumultuous” last year filled with “a lot of instability and uncertainty,” Posilkin said her team has been “really happy” to see “tremendous interest in the TMF, and most importantly pressure from the administration to modernize the services that we have, and we’re seeing agencies respond to that.”

Posilkin has previously said the TMF has $200 million to spend until its new project funding capability expires at the end of the fiscal year next month. Unless the House passes the Senate stopgap that would extend this ability until Dec. 11, Posilkin said Thursday that the fund will have $200 million next year without further appropriations, largely due to agency repayments.

“We are now at the point that we get to say that dollars have come into the TMF through appropriations, gone out to agencies, come back and been reinvested, which was really the dream,” Posilkin said.

The House returns next week and has signaled it will take up the Senate-passed measure. Regardless, Posilkin said, “there are other bills that are potentially being discussed that could extend us to 2032 and beyond. But at this point, my bet is as good as anyone’s about how Congress acts.”

The TMF is expected to save the government $1 billion in the coming years, the Government Accountability Office said last month. In fiscal 2024 alone, TMF projects saved 378 million hours and garnered $12 billion in return on investment, Posilkin said.

“We know that this has been a tremendously successful program, and it really feels quite wild to have the level of support,” she said. “I think one of the stories we tell in government technology is often about failure, and I really want to be sure that part of the story of the TMF is ensuring success and making sure that not only are we getting money back, but we’re also making sure projects succeed.”

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @technology modernization fund 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/tmf-wants-to-experim…] indexed:0 read:4min 2026-08-27 ·