by Meredith Delaware
There is tremendous pressure on federal agencies to quickly adopt artificial intelligence (AI)-enabled tools, from top-down mandates to prioritize AI integration to bottom-up pressure to address efficiency. The direction is clear. Arriving at a point of readiness, however, remains a challenge. As the Government Accountability Office summarized in a 2025 report on Generative AI Use and Management at Federal Agencies, many agencies still struggle with basics such as establishing data readiness and building necessary infrastructure, much less adopting sophisticated AI workflows by employees at all levels.
The same need for efficiency that is pushing agencies to push for AI adoption is the reason it’s difficult to increase AI adoption rates. Employees are rarely provided uninterrupted time and space to learn how to navigate AI systems, experiment with these technologies, explore data through AI, and rework existing processes. However, this work is essential. Without prioritizing experimentation as a central focus for AI implementations, people will continue to follow existing processes and patterns when daily pressures continue, measurably slowing AI adoption.
AI experimentation can help government agencies uncover powerful insights. By encouraging teams to experiment in areas that require minimal upfront investment, government agencies can strategically identify potential AI projects, foster user adoption, and deliver greater successes.
Experimentation Enables an AI-Ready Workforce
As of Dec. 2024, 61% of state, local, and federal government employees had never used generative artificial intelligence on the job, according to a survey by Eagle Hill Consulting. Only 8% of U.S. government employees were using the technology for work daily. Encouraging your workforce to experiment with existing AI tools can help overcome barriers to AI use. Barriers may include a lack of AI-related skill sets, as well as resistance to using new technologies and processes.
A 2025 report by the U.S. General Accountability Office found that government agencies eager to expand the use of generative AI have faced difficulty developing an AI-literate workforce. That may be because many agencies overlook the importance of securing buy-in from the humans who will use the technology. Engaging the data owner, engineer, analyst, and user is essential for advancing mission-critical AI initiatives. When agencies don’t consider how people are impacted by AI readiness initiatives, their professionals are prone to return to old patterns.
User adoption and change management techniques are essential to help agencies realize the benefits of their AI investments faster. Agencies will have to invest in—and actively support—the importance of change management.
Practical Steps for Fostering AI Adoption
In the private sector, 62% of respondents to a 2025 McKinsey global survey on the state of AI reported that their organizations were, at least, experimenting with AI agents. Respondents also reported use-case-level cost and revenue benefits at this stage, with 64% agreeing that AI helped enable innovation even through experimentation.
Government agencies ready to realize this same value might consider nurturing small-step experimentation to help teams get comfortable with AI tools and drive their adoption. The Interior Department took this approach when it launched a generative AI project to improve internal searches. The agency tailored a prompt script for each business line to encourage employees to evaluate how they could use the tool to streamline routine work within their division. The prompt encouraged experimentation and allowed their teams to get familiar with AI tools.
Below are several additional steps leaders can take to foster AI adoption through AI experimentation.
- Set aside dedicated time for AI exploration. According to a 2025 AI Adoption Reportby the Wharton School of the University of Pennsylvania, 43% of surveyed business leaders warn that a lack of time to practice can increase the risk of skill atrophy. Yet nearly a quarter (23%) of state, local, and federal government employees surveyed by Eagle Hill Consulting noted that they lacked the time to experiment with AI. Organizations can help employees build familiarity with and skills in using AI tools by dedicating time to AI experimentation.
- Start experimenting with the tools you have available. Tools such as Microsoft Copilot and the recently launched USAi suite are already available to government organizations. Experimenting with these tools can help teams enhance their AI skill sets.
Another low-risk way to introduce teams to AI is to encourage experimentation with unutilized data. Many organizations preparing for an AI future set out to collect as much data as possible. There’s an expectation that “eventually” this data will have an AI use case attached to it. Teams can even use AI tools to explore potential use cases for this data. - Launch short experimentation sprints. By thinking smaller, agencies can often unlock bigger rewards. Low-investment methods of experimentation can help employees become familiar with AI and ease barriers to future adoption. Moreover, spending even a brief amount of time on exploration can uncover powerful insights. By launching 6-, 8-, or 10-week experimentation sprints, agencies can encourage teams to interact with AI while limiting the time they invest in this phase.
- Identify AI champions or build buddy networks. Idea sharing through mentorships or peer networks can encourage team members to discuss tools, use cases, and lessons learned through experimentation. AI advocates can take the lead by openly sharing their experiences using AI and guiding team members toward appropriate tools.
- Tie AI experimentation to career growth. Agencies can incentivize AI use by connecting it to mission objectives – and to individual performance goals. Scott A. Snyder, a senior fellow at the Wharton School of the University of Pennsylvania, recommends incentives such as AI innovation prizes, bonuses based on AI impact, and compensation boosts based on demonstrating proficiency in AI-related skill sets.
It’s Time to Explore AI’s Full Potential
To meet demands to implement AI in federal operations, agencies must take steps to identify actionable AI use cases. An experimentation-based approach is a proven way to develop an AI-literate workforce and deliver tangible value to the organization.
PCI-GS can help organizations get more from this experimentation. We meet organizations where they are, recommending the AI tools and providing the expertise you need to achieve your AI goals, whatever your project constraints. Whether you’re beginning to assess your existing ecosystem or are ready to address interoperability gaps to help data flow across systems, we can help.
