As the administration doubles down on artificial intelligence (AI), federal civilian IT teams are racing to move beyond pilot programs and scale adoption with confidence.

OMB Memorandum M-25-21 called on agencies to “adopt a forward-leaning and pro-innovation approach,” and the new America’s AI Action Plan outlines near-term priorities to expand access to frontier models, grow the AI workforce, and accelerate adoption across government.

Yet, many agencies remain stuck in pilot. As administrative pressure builds and generative and agentic capabilities expand, what do public sector leaders need to shift from sandbox to scale?

In this new study, MeriTalk, Dell Technologies, and NVIDIA surveyed 75 federal civilian IT leaders to pinpoint where AI pilots are gaining traction and what differentiates projects that successfully scale.