Why Justin Fulcher Says Government AI Must Work Within Constraints
The conversation about artificial intelligence in the public sector has a tendency to swing between two extremes. Either AI will transform government overnight, or the bureaucratic constraints facing agencies make meaningful deployment impossible. Technology founder and former Defense Department advisor Justin Fulcher occupies a more useful middle ground: AI can deliver real improvements to government operations, but only when deployed with a clear-eyed understanding of how those institutions actually work.
Institutional Drag Is the Starting Point
Fulcher’s analysis begins not with technology but with diagnosis. Federal agencies face a specific type of dysfunction that he calls institutional drag: the compounding inefficiency that builds up when outdated processes, siloed data systems, and compliance frameworks designed for paper-based workflows remain in place as workloads grow and missions become more complex.
“The issue is not national decline; it’s institutional drag,” Fulcher has written, adding that core systems across government, healthcare, and defense continue to operate as if it were 1975. That framing matters because it identifies where AI can do the most good. If institutional drag is the problem, the highest-value AI applications are those that reduce friction rather than those that add new capabilities on top of broken workflows.
Document processing, data synthesis, routine compliance checking, and administrative coordination are all areas where AI can deliver measurable time savings without requiring structural reorganization or raising significant accountability concerns. Justin Fulcher has argued that those are exactly the applications agencies should prioritize, because they are the ones most likely to earn adoption and deliver durable results.
Constraints as Design Requirements
Justin Fulcher’s credibility on this issue comes from having worked inside the institutions he is analyzing. At the Department of Defense, he focused on acquisition reform and IT modernization, contributing to reforms that cut software procurement timelines from years to months. Before that, at RingMD, he built healthcare technology across highly regulated Asian markets where the gap between what technology could theoretically do and what institutions would allow was a daily operational reality.
The consistent lesson from both environments: successful technology adoption in regulated settings requires treating constraints as design requirements rather than obstacles. Data security requirements, civil service rules, procurement timelines, and accountability standards are not problems to be worked around. They are the parameters within which AI systems must be designed if they are to gain adoption and endure.
Fulcher has noted that this discipline is particularly important for AI, where the potential for opaque decision-making raises specific accountability concerns. Systems deployed in government must be auditable and explainable. They must integrate with legacy infrastructure. And they must earn trust from the workforce that will use them, not just from the officials who approve their purchase. Getting those elements right is what makes modernization last. Visit this page for more information.
Find more information about Justin Fulcher on https://x.com/JustinFulcher