As agencies work to regain momentum after the 44-day government shutdown, former Federal Chief Data Officer and generativework.ai founder, Ted Kaouk sees a major opportunity—not just to return to normal, but to reinvent how government innovates. In his conversation with Fed Gov Today with Francis Rose, Kaouk shares a practical, energizing view of how artificial intelligence can help agencies make up lost ground and build a more agile, adaptive workforce.
Kaouk begins with a theme that runs through his entire discussion: speed to innovation. He explains that AI—especially modern large language models—dramatically lowers the cost and complexity of building and testing new ideas. He describes an experiment conducted just before the shutdown: a “30 apps in 30 days” challenge in which his team built an application every single day. The purpose wasn’t to create perfect or production-ready tools, but to demonstrate how fast agencies can prototype, test, and refine ideas. The result, he explains, showed that agencies can now explore a broad range of possibilities quickly and safely.
That speed, Kaouk says, comes from the new ability to work in safe sandboxes using synthetic data, rather than waiting for lengthy approvals to access sensitive information. The combination creates a risk-managed environment where teams can prototype freely without jeopardizing security or privacy. In his view, this is one of the most powerful ways agencies can begin to use AI today. Instead of long planning cycles, agencies can build something, share it with a business partner the next day, and immediately determine whether it’s worth further investment.
Kaouk is clear, however, that AI success isn’t just about tools—it’s about people. He stresses that as AI adoption grows, federal workers must develop a new set of skills centered on guiding AI agents. For decades, government employees have looked to managers and leaders for direction and feedback. Now, he says, everyone will need to learn how to provide that guidance to AI systems as well. He personally practices this every day, working with AI agents and refining their behaviors, and he argues that leaders should do the same so they can teach their teams how to work effectively alongside AI.
This shift means that AI is not merely a technical topic—it is a workforce transformation. Kaouk explains that many agencies feel
As the conversation moves deeper into the nature of AI, Kaouk offers an important clarification. He explains that large language models are not confined to one specific type of task; they can support everything from learning environments to complex decision-making and execution. The key, he says, is to be deliberate with intent. Agencies should understand exactly what behavior they want an AI system to perform and build toward that purpose, rather than adopting AI for the sake of novelty.
Kaouk also touches on the evolving relationship between AI, automation, and traditional coding. He notes that AI is increasingly blending natural language-driven behavior creation with algorithmic automation running in the background. As a result, agencies will need a multidisciplinary mindset, bringing together people from technology, the humanities, social sciences, and business fields. The future of federal AI, he predicts, lies in teams with diverse perspectives who can combine human-centered understanding with technical capability.
This broader shift leads to what Kaouk sees as one of the biggest challenges ahead: culture. Many federal organizations are not naturally structured for cross-disciplinary collaboration or rapid experimentation. He believes Chief Human Capital Officers will play a central role in building the skill sets and organizational structures needed to support AI-enabled work. Agencies may ultimately need to rethink roles, workflows, and the balance between human and agentic contributions, potentially reshaping the workforce more dramatically than at any time in decades.
Kaouk acknowledges that the federal AI landscape is still maturing. Some leaders continue to chase AI “because everyone else is,” while others are just beginning to understand the practical outcomes that are possible. He describes AI progress as a series of breakthroughs—some happening over months, others within a single day—and emphasizes that agencies won’t know what’s possible until they try. Innovation, he says, requires persistence, curiosity, and a willingness to explore beyond the first promising outcome.
Toward the end of the conversation, Kaouk highlights two policy and process changes that could accelerate progress: creating safe experimentation spaces and developing comprehensive training and skills roadmaps. These steps, he believes, would give agencies the foundation they need to innovate responsibly while preparing employees for the transformations ahead.
Kaouk’s message is both realistic and encouraging. Agencies may be emerging from a shutdown, but they’re also entering a moment of extraordinary possibility. With rapid prototyping, deliberate experimentation, and a workforce ready to engage with AI, federal organizations can push beyond old constraints and build a future where innovation is continuous, accessible, and driven by the people closest to the mission.