A new technology direction is taking shape inside the General Services Administration, according to sources cited by WIRED. In a Monday morning meeting, Thomas Shedd, the recently appointed Technology Transformation Services director and Elon Musk ally, told GSA workers that the agency’s new administrator is pursuing an “AI-first strategy.”
The message was broad: Shedd described a GSA that would operate more like a “startup software company,” with more automation, more centralized data, and a larger role for AI across federal work.
What Shedd told GSA staff
The meeting was held in person and on Google Meet. Sources tell WIRED that Shedd used it to outline projects the agency could pursue in the coming weeks and months, with a particular emphasis on artificial intelligence.
One example was the development of “AI coding agents” that would be made available for all agencies. Shedd also said he believed much work at TTS and across the broader government could be automated, especially finance tasks.
That framing matters because TTS is housed within the GSA and has functioned as a federal technology task force. Shedd’s comments suggest a push to make AI a central tool not only for one office, but for government-wide work.
Shedd also told employees to think of TTS as a software startup that had become financially unstable. That comparison appeared to connect the agency’s technology goals with questions about efficiency, staffing, and the future structure of work inside the office.
Why the shift has raised questions
The meeting followed WIRED reporting that many of Musk’s associates had moved into high-level roles at the GSA and the Office of Personnel Management. Before joining TTS, Shedd was a software engineer at Tesla, one of Musk’s companies.
According to the source article, the transition has already caused confusion among GSA staffers. Employees had been placed in surprise one-on-one meetings, asked to present their code, and in some cases met with young engineers who did not identify themselves.
That uncertainty shaped the Monday discussion. Staff wanted to understand who was involved, what projects were coming, and what the future of the agency’s technology work would look like.
Shedd addressed the young engineers during the call, sources say. He said one of them felt comfortable enough to introduce himself in meetings on Monday, while also saying he was nervous about their names being publicly revealed and their lives upended.
AI, data, and privacy concerns
One of the most consequential ideas raised in the meeting was a centralized data repository for the federal government. Shedd said he was actively working with others on a strategy to create one, sources say.
But the details were not clear. The source article says it was uncertain where such a repository would be based or whether the projects would comply with privacy laws. Shedd described those concerns as a “roadblock” and said the agency should keep pushing forward to see what was possible.
That is where the debate becomes larger than software development. Centralized government data, automated finance tasks, and AI coding agents all suggest a more integrated technical environment. They also raise practical questions about oversight, legal compliance, security, and domain expertise.
“This does raise red flags,” a cybersecurity expert who was granted anonymity due to concerns of retaliation told WIRED on Monday, who noted that automating the government isn’t the same as automating other things, like self-driving cars. “People, especially people who aren’t experts in the subject domain, coming into projects often think ‘this is dumb’ and then find out how hard the thing really is.”
The warning is not that automation cannot be useful. The point is that government work can include constraints that are not obvious from the outside. A process that appears inefficient may be shaped by legal, security, privacy, or operational requirements.
How TTS and DOGE fit together
Sources say Shedd closely connected TTS and the United States Digital Services, which was rebranded as the United States DOGE Service, or DOGE, under Trump. He described them as “pillars” of a new technological strategy.
Later in the meeting, Shedd said there was no plan to merge the two groups. Instead, projects would move through both depending on available staff and expertise, while he continued to emphasize collaboration between TTS and DOGE.
That distinction leaves the structure unresolved. The groups may remain separate, but the source article indicates that their work could become more closely linked as the new strategy develops.
Staffing and workload remain unclear
The meeting did not resolve several employee concerns. Sources said Shedd could not answer many questions about deferred resignations, the return to office mandate, or whether the agency’s staff would face substantial cuts.
At one point, he indicated that workforce cuts were likely for TTS, but did not provide more detail. Similar questions had also been asked of Department of Government Efficiency leadership in a Friday meeting first reported by WIRED.
Near the end of the call, a TTS worker asked whether staff would be expected to work more than 40 hours per week to handle upcoming work and the possibility of laid-off workers. Shedd responded that it was “unclear.”
For now, the picture is a strategy with a clear direction but many unresolved details. GSA leaders are signaling a move toward AI-first government technology, automation, coding agents, and centralized data. Staff are still waiting for answers on how those ambitions will affect privacy, team structure, job security, and daily work.