January 13, 2025
Presented by Maximus
Leaders from the General Services Administration describe a government-wide effort to modernize acquisition and accelerate responsible adoption of artificial intelligence by aligning policy, technology, and workforce practices. During a discussion at ACT-IAC’s Imagination ELC, GSA officials explain how changes to the Federal Acquisition Regulation, new AI platforms, and shared services are coming together to help agencies buy and use emerging technology more effectively.
Laura Stanton, Deputy Commissioner of GSA’s Federal Acquisition Service, says the ongoing FAR overhaul is both critical and far-reaching. While structural changes—such as updates to Multiple Award Schedules and the introduction of new contract types under OMB’s category management framework—are important, she emphasizes that cultural change is essential. Contracting officers have been trained for years to follow established methods, and adopting the new FAR requires leadership, training, templates, and strong support from institutions like the Federal Acquisition Institute and DAU. Without that foundation, Stanton says agencies will struggle to realize the benefits of reform.
To help the workforce adapt, Stanton points to practitioner handbooks and the development of more targeted buying guides, particularly for generative AI. She explains that agencies need clear decision frameworks to understand when and how to procure emerging technologies. GSA is already collaborating internally to develop guidance that walks acquisition professionals through these complex choices.
David Shive, GSA’s Chief Information Officer, explains that GSA has been using artificial intelligence for years, but the rise of generative AI prompts a new approach. About 20 months ago, GSA begins rapidly prototyping and piloting tools to understand what industry offerings can—and cannot—do. Through that process, GSA identifies the need for a strong control environment to ensure AI behaves in ways that support government missions.
That work leads to the development of GSAi, a platform that looks like a chatbot on the surface but is designed to provide deep observability beneath it. Shive says the real value lies in capturing telemetry about how users interact with AI systems and how the underlying components interact with one another. Because other agencies express interest, GSA creates USAi, a shared version of the platform that allows agencies to explore multiple large language models while collecting the same detailed usage data.
Zach Whitman, GSA’s Chief AI Officer and Data Scientist, says USAi quickly becomes a collaborative, multi-agency effort. Agencies use the platform to learn how AI is actually adopted across mission areas, whether through chat, APIs, developer tools, or integrated workflows. Whitman explains that telemetry data reveals not just how often AI is used, but what kinds of mission functions it supports, allowing leaders to align AI usage with strategic goals.
The speakers also describe how this data informs acquisition decisions. By collecting expert feedback on AI responses, GSA can compare model performance across cost, speed, energy use, and quality. Shive says this allows the government to become a more informed buyer, rather than relying on vendor claims.
Stanton adds that GSA’s approach gives agencies options: they can use shared services like USAi or procure solutions directly through OneGov deals. Deeply discounted AI pricing allows agencies to experiment, test use cases, and decide how to scale in the future. Across the discussion, the speakers stress that collaboration, transparency, and learning are key to bringing AI into government responsibly and effectively.