Federal agencies are facing significant barriers to scaling artificial intelligence (AI) technologies due to outdated IT systems, workforce shortages, and governance challenges despite widespread interest in implementing the technology, according to a new survey from Ernst & Young LLP (EY).
While federal agency leaders overwhelmingly view AI as essential to modernization and efficiency, the survey finds that many agencies remain stuck in early stages of adoption and are unable to move beyond pilot programs without first addressing underlying technology and workforce limitations.
According to the 2026 EY Government and Public Sector Federal Trends Report, 92% of agency leaders say AI is critical for improving efficiency, and 88% view it as key to modernization.
However, only 50% report having multiple fully deployed AI initiatives, while 38% are still running pilot programs, and 11% are in early exploration stages. Nearly half (46%) are still identifying specific use cases for AI.
The findings underscore a central challenge: agencies cannot effectively scale AI without modernizing legacy systems. Nearly half of respondents (48%) cited difficulty integrating AI with existing IT infrastructure as a major barrier, while 44% pointed to shortages in AI-related skills and training.
Overall, 89% of federal leaders said their agencies face barriers to achieving greater efficiency. Budget constraints (34%), outdated technology infrastructure (32%), and a lack of skilled personnel (31%) were among the most commonly cited obstacles.
“Federal agency leaders are under real pressure to deliver efficiency gains, but technology alone won’t close the gap,” said Paul Donato, EY Americas Government & Public Sector federal leader.
“Modernization requires ensuring legacy environments are AI-ready, strengthening governance frameworks and investing in workforce capabilities,” he said.
Despite these challenges, agencies are pushing forward with modernization efforts. All surveyed leaders reported undertaking efficiency initiatives in fiscal year 2026.
Top priorities for those include improving cybersecurity infrastructure (44%), investing in emerging technologies such as AI and machine learning (43%), and implementing new data systems (40%).
Still, progress remains uneven. While 81% of respondents gave their agencies high marks for modernization efforts, only 22% said a majority of their IT systems are fully post-transformation, and 26% acknowledged they remain largely legacy tech-based.
Time and talent constraints continue to slow implementation. About 48% of leaders said it takes a year or more to move an IT program from pilot to full-scale deployment, with the workforce skills gap (44%) ranking as the top barrier to modernization, ahead of procurement delays (32%) and cybersecurity threats (32%).
“The most significant bottleneck to tech modernization are the three S’s limiting federal government agencies: speed, skills, and scale,” Donato said. “Without these capabilities, the most advanced tech remains underutilized,” he added.
Governance gaps further complicate AI adoption. Only 38% of agencies reported having a comprehensive, unified AI governance strategy – a shortfall that may hinder efforts to move AI initiatives from pilot to production.
The survey, conducted between November and December 2025, included 131 federal government leaders in finance, IT, and human resources roles.