Government agencies face a critical gap: they're required to track and govern artificial intelligence systems they can't actually see. Federal mandates now require agencies to maintain comprehensive inventories of AI use cases, manage associated risks, and demonstrate accountability to the public. But a striking disconnect exists between AI adoption and visibility. While 81% of organizations use generative AI extensively or sparingly, only 31% report having visibility into AI software across their environments. For public sector leaders, this visibility gap isn't just an operational inconvenience; it's becoming a compliance problem.

What's Driving the Push for AI Governance in Government?

The pressure on federal agencies to govern AI responsibly is coming from multiple directions. The White House's Executive Order 14110 called for a coordinated federal approach to safe, secure, and trustworthy AI. More recently, the Office of Management and Budget (OMB) memorandum M-24-10 directed agencies to advance AI governance, innovation, and risk management, including establishing Chief AI Officer positions, expanding reporting requirements, and maintaining AI use case inventories. The National Institute of Standards and Technology (NIST) also released an AI Risk Management Framework that provides a foundation for managing AI risks across organizations. At the state level, legislatures are evaluating how AI should be governed across public services, privacy, hiring, education, and other sensitive areas.

This regulatory momentum reflects a fundamental shift: AI governance is no longer optional. For public sector organizations, governance, transparency, risk management, and accountability have become operating requirements, not emerging best practices. Agencies are exploring AI-enabled citizen services, workforce productivity tools, and data analysis capabilities. Yet the challenge extends far beyond simply deploying new technology. It requires building governance frameworks that can document AI use, understand risk, support transparency, manage costs, and demonstrate value from the start.

Why Can't Organizations See Their Own AI Systems?

The visibility crisis is widespread and growing. Research from Flexera's 2026 State of ITAM Report reveals that 84% of organizations report difficulty tracking and adopting new AI applications, making this the most commonly cited software asset management challenge. Yet despite this challenge, only 31% of organizations have visibility into AI software across their environment. The problem extends beyond AI applications alone. Complete IT visibility across organizational environments has declined to just 36%, reflecting the growing complexity introduced by Software-as-a-Service (SaaS) platforms, cloud services, and AI tools.

This visibility gap creates cascading problems for public sector leaders. Without understanding where AI is being used across formal systems, procured applications, cloud services, embedded software features, and emerging productivity tools, agencies cannot consistently assess risks, answer oversight questions, control costs, or demonstrate how AI supports mission delivery. The challenge is compounded by the rapid proliferation of AI tools. Organizations are moving quickly from experimentation to adoption, with generative AI-enabled public cloud services now used by a majority of surveyed organizations. As AI initiatives mature, fewer organizations report having no plans to use these services, yet governance structures haven't kept pace with deployment speed.

How to Build AI Visibility and Governance in Government

• Establish Comprehensive Inventories: Create and maintain detailed inventories of all AI systems, applications, models, owners, users, and associated costs. OMB M-24-10 specifically requires agencies to inventory AI use cases at least annually, submit inventories to OMB, and post public versions on agency websites. This inventory becomes the foundation for all other governance activities.
• Implement Visibility Infrastructure: Deploy tools and processes that provide real-time visibility into AI usage across the organization. This includes tracking AI adoption across formal systems, cloud services, and emerging productivity tools. Visibility underpins regulatory compliance, responsible AI governance, budget accountability, security oversight, and transparency to stakeholders and citizens.
• Adopt Enterprise Governance Models: Public sector leaders can adapt lessons from enterprise organizations that have spent years building governance disciplines for cloud, SaaS, IT asset management, and financial operations. These proven frameworks can help agencies move from policy intent to repeatable execution.
• Assign Clear Ownership and Accountability: Establish dedicated teams or senior leaders responsible for AI oversight. Large enterprises are investing heavily in governance, with 85% reporting a dedicated team or senior leader responsible for AI oversight. Public sector agencies should follow this model to ensure governance doesn't become a diffuse responsibility.

What Happens When Agencies Wait to Govern AI?

The timing of governance implementation matters significantly. Organizations that wait to govern AI until after adoption accelerates will struggle to regain control. Public sector leaders need governance models that help them identify what AI exists, who owns it, what risks it introduces, and how it supports mission outcomes from the start. This is fundamentally different from how technology governance has traditionally worked. In the past, organizations could deploy technology first and implement oversight later. With AI, that approach creates compounding problems.

The stakes are particularly high for government agencies because they operate under public accountability requirements that private sector organizations don't face. Citizens expect transparency about how AI is being used in government services. Elected officials and oversight bodies demand clear accounting of AI spending and outcomes. Federal agencies must demonstrate that AI investments deliver value while managing risks responsibly. Without visibility into AI systems, agencies cannot fulfill these accountability obligations.

Why Cost Uncertainty Makes AI Governance Even Harder

Beyond visibility challenges, government agencies face another governance problem: AI costs are unpredictable. Unlike traditional technology investments, AI doesn't come with a fixed price tag. Costs fluctuate based on usage, data volumes, cloud consumption, model interactions, licensing terms, and the rapid introduction of new tools and services. As agencies move from pilots to broader deployment, many will discover that AI spending becomes difficult to forecast and even harder to control. This cost uncertainty creates budget planning challenges and makes it harder for agencies to demonstrate fiscal stewardship to taxpayers and oversight bodies.

The convergence of visibility gaps, rapid adoption, regulatory requirements, and cost uncertainty creates a complex governance challenge for public sector leaders. However, the solution is clear: agencies must prioritize visibility and governance infrastructure alongside AI adoption, not after it. By learning from enterprise organizations that have already navigated these challenges, public sector leaders can build the inventories, oversight structures, and financial discipline needed to scale AI responsibly while meeting their unique accountability obligations to the public.