Agencies are expanding their use of artificial intelligence as the technology moves from isolated pilots into day-to-day operations across government.
From summarizing documents and researching vendors to supporting core mission functions such as fraud detection, case prioritization and border inspections, AI is being applied to a wide range of agency missions.
The Department of Defense, intelligence community, IRS and NASA were early adopters of AI, but their ranks grew rapidly with the advent of generative AI. Look no further than the federal AI use case inventory, which has grown to more than 2,000 documented applications spanning productivity tools to operating systems.
“This growing interest is unfolding alongside a broader policy push to accelerate AI adoption across government,” says Massimiliano Claps, research director for worldwide national government platforms and technologies at the International Data Corporation.
Executive orders, the White House’s AI Action Plan, and General Services Administration-led procurement efforts have all emphasized experimentation, faster iteration and lowering barriers to entry. All of this helps explain the launch of USAi, GSA’s free, FedRAMP-compliant generative AI evaluation suite, which allows agencies a frictionless way to explore and assess tools before going through with deployment.
GET READY: These are the four biggest AI trends heading into 2026.
What Is USAi?
USAi is a centralized evaluation and access platform designed to help agencies safely experiment with and deploy generative AI tools.
The platform provides access to leading foundational models from multiple providers including Google, Meta, Anthropic and OpenAI. Agencies can test and compare these models without having to manage separate vendor environments.
“The objective of USAi is to provide a set of common, standard tools to experiment with different models,” Claps says.
USAi is aimed primarily at IT leaders, developers and data teams within agencies that are looking for an entry point into AI.
Importantly, USAi doesn’t impose additional governance constraints beyond those built into the underlying models. Agencies can apply their own guardrails or rely on vendor-provided controls, an approach that prioritizes innovation.
Click the banner below to ready data center infrastructure for AI.
Key Features of the USAi Platform for Federal Agencies
USAi consists of three integrated components at its core. The first is a unified chatbot interface that gives federal employees access to leading large language models.
“We wanted to make sure that everyone had access to the latest and the greatest models via a singular chatbot,” says Zach Whitman, chief AI officer and chief data scientist at GSA.
The second component is programmatic application programming interface (API) access, which is intended for operational and automated use cases rather than human-in-the-loop prompting. These APIs support workflows such as large-scale data processing, integration with robotic process automation and embedding AI into production pipelines.
A third key element is the USAi console, which combines access management with model evaluation. Agencies can control who uses which models, configure settings and compare models using standardized evaluation frameworks.
The console measures performance, safety and suitability for specific agency use cases, including the ability to create custom evaluation tests.
USAi also includes tools to assess and track costs, such as token usage and consumption patterns across models.
How USAi Helps Agencies Act on the AI Action Plan
The platform helps agencies move from policy intent to practical execution under the White House’s AI Action Plan by giving them concrete data on how AI is being used inside their organizations, Whitman says.
A central goal of the AI Action Plan is to accelerate agency adoption while making smarter, evidence-based decisions about scale, cost and procurement. USAi supports that by letting agencies observe real usage patterns before committing to long-term contracts or enterprise licenses.
“It gives agencies a chance to try out how their user base will respond to these types of toolings,” Whitman says.
Through the platform’s chatbot and API access, agencies can see how frequently AI tools are used, how intensively they’re consumed and which models are being selected. That data can then be translated directly into cost scenarios.
Click the banner below to keep up with the IT, cyber and AI experts making government efficiency a reality.
The visibility helps agencies determine whether a consumption-based model or a per-seat licensing approach makes more sense as adoption grows.
“There’s a crossover point where it makes sense to move toward a per-seat, unlimited usage model versus per-token consumption,” Whitman says.
USAi’s console also provides insight into where and why AI is being used across an organization. Agencies can see which components or mission areas are using AI tools, what categories of tasks they support and how that aligns with agency priorities.
“It gives a very clear sightline into how it is being used and which mission centers it aligns to,” Whitman says.
Taken together, that combination of experimentation, cost modeling and usage transparency helps agencies align AI adoption with mission goals — turning the AI Action Plan from a policy framework into an operational roadmap.
Click the banner below for the latest federal IT and cybersecurity insights.
Real-World Impact: How Federal Agencies Are Using USAi to Transform Operations
While it’s “too early to tell” what the real-world impact of USAi will be — due in part to the government shutdown that occurred soon after its launch — the platform has been rolled out to more than 20 agencies so far, and they’re starting to report usage patterns back to the GSA, Whitman says.
“There’s been a positive uptake from partners willing to jump on this concept, so we can further shape it to be directly supportive of their mission function,” Whitman says. “USAi is designed to be incorporated into operational and production settings.”
The USAi effort is firmly in a ramp-up phase, with meaningful adoption data gauging impact expected in early 2026.
Success will hinge less on tool availability and more on how GSA structures the initiative. Agencies will see meaningful gains if USAi is delivered as part of a broader program that includes guidance, training and advisory support aligned with existing acquisition vehicles, Claps says.
“It’s all about change management,” Claps says. “And empowering the workforce.”
