Federal civilian CIOs and IT leaders overwhelmingly view artificial intelligence (AI) as essential to mission success, but most remain stuck in pilot mode, according to new research from MeriTalk, Dell Technologies, and NVIDIA.
The study, From Sandbox to Scale: The People, Processes, and Platforms Needed to Accelerate AI in Federal Civilian Agencies, surveyed senior IT decision-makers in federal civilian agencies.
Ninety-three percent say getting AI right is mission-critical, and 95% already credit generative AI (GenAI) with measurable productivity gains. Yet only 33% have embedded AI agency-wide.
The majority, nearly three-quarters of all agencies, are piloting AI projects and are seeing wins in areas including HR, finance, procurement, cybersecurity, and citizen services.
But scaling remains elusive. Eighty-one percent of Federal leaders admit they are more likely to launch a new pilot than expand an existing one. Projects that stall most often face governance and compliance blocks (45%), poor data quality (44%), and integration hurdles (41%). Only half of agencies have a formal AI governance framework, and just 41% have a dedicated AI budget – gaps that continue to slow enterprise adoption.
Agencies that are successfully scaling pilot programs enterprise-wide say they are investing across three areas. On the people side, CIOs cite change management (56%), expanded AI training (55%), and leadership teams or centers of excellence (53%) as critical enablers.
On the process side, governance frameworks (52%), mission-tied use cases (48%), and performance metrics (43%) help drive progress.
And, on the platform side, modernized data infrastructure (57%), scalable cloud or hybrid environments (49%), and centralized governance tools (48%) are proving essential.
The research explores:
- The biggest factors in AI success
- Where CIOs anticipate the most significant payoffs
- The AI-related priorities that are on the rise
For insights, download the full report: From Sandbox to Scale: The people, processes, and platforms needed to accelerate AI in federal civilian agencies.