In the 1980s and 1990s, many large manufacturing companies pursued offshoring strategies. Not necessarily because careful analysis showed a clear link between offshoring and achieving business objectives, but because competitors were offshoring. Within a few years, companies moved large-scale production overseas, often at the expense of supply chain flexibility. The problem was not with offshoring itself, but rather with leaders starting with the wrong questions and not being clear about how offshoring fit into their overall strategy. Federal agencies are making the same mistakes with AI.
The Trump Administration's AI Action Plan, announced in July 2025, creates an urgent need for government agencies to demonstrate progress in artificial intelligence. But urgency without clear direction produces activity rather than results. Leaders across government agencies are asking, “What is our nation's AI strategy?” When the question needs to be asked: “How can AI enable our strategy?” Here's why:
What happens when strategy is replaced by pressure?
The rush to offshoring offers a cautionary tale for federal leaders to reconsider. For many manufacturing companies, offshoring was an entirely after-the-fact decision, driven by intense pressure from Wall Street to demonstrate cost-cutting efforts.
Management announces an offshoring strategy, consultants are hired, and operations are relocated, but the real cost impacts often become apparent over time, including hidden costs in coordination and quality control, and loss of flexibility to withstand disruption. In many cases, operational changes have created strategic vulnerabilities across the supply chain.
The most successful companies with offshoring had strategic goals and considered offshoring as a way to reduce costs and diversify their supply chains. Treating it as a cost-performance tool, rather than a necessity in itself, will make the difference between it becoming a competitive advantage or an expensive distraction.
Similar warning signs are visible in today's AI adoption race. Government agencies are under pressure to demonstrate progress in AI, and the easiest way is typically to launch a pilot, create an AI working group, and report on the number of use cases identified. However, this focus on activities may or may not yield results that are important to the institution's mission.
The hidden costs of deploying AI without a strategy
When AI initiatives are not rooted in organizational strategy, predictable problems arise. First, use cases cluster around process optimization rather than transformation. The team identifies ways to make existing workflows slightly faster or cheaper. While these improvements are real, they are only incremental. The transformative potential of AI to completely rethink current workflows and significantly change the way work is done remains untapped because it is unclear what the transformation should look like to achieve strategic goals.
Second, adoption is fragmented. Different business units pursue different tools to solve different problems, but there is no consistent thread that ties them together. This fragmentation makes it nearly impossible to build organizational capabilities with AI. Each initiative becomes a one-time experiment rather than a building block toward a strategic goal.
The third and most harmful is employee disengagement. When people are told to use AI without understanding how it advances the missions they care about, the mandate feels arbitrary. This could lead to resistance, especially as media coverage of job losses due to AI increases. The goal of AI implementation is to reduce administrative burden and increase productivity. But without a strategic framework, people can spend time on tools they don't understand for unclear purposes, with the opposite effect of decreasing productivity.
What strategic-first AI implementation looks like in practice
Consider two hypothetical federal agencies that both employ the same AI tools.
Agency A starts by asking, “What is our AI strategy?” You may form an AI task force to evaluate vendors, select a platform, and deploy training. Next, track metrics around tool adoption and identified use cases. After one year, report how many of your employees have used AI tools and how many cases have been documented. But when asked how those results connect to the agency's strategic mission, the answer is likely ambiguous.
Agency B starts by asking, “What are our strategic imperatives?” and “Where do we see the barriers to progress and the opportunities to accelerate?” Only then will we consider whether AI can help remove barriers or accelerate opportunities. Create mixed-level teams to test AI tools in a sandbox environment with the potential to fail quickly and share learnings. Success is measured by progress against strategic priorities, not adoption rates. After one year, the percentage of employees using AI tools on a regular basis had decreased, but those employees reported eliminating significant bottlenecks. These case studies and the results achieved are inspiring more people to adopt AI tools.
Which agencies have gained more value from their AI investments? Which agencies are likely to gain even more momentum around AI?
Why does top-down approach alone fail?
Successful adoption of AI across an organization requires both top-down strategic clarity and bottom-up experimentation. Senior leaders need to provide a strategic framework and ask themselves questions such as: Which goals can AI accelerate? Where should we focus our resources? What does success look like?
However, leaders alone cannot identify all valuable AI applications. Tactical employees understand where manual processes create delays, where data exists but is underutilized, and where better information could speed decision-making. Their insights are essential to making AI adoption practical rather than theoretical.
Successful AI integration requires leaders to provide strategic direction, resource allocation, and employees to experiment, learn, and identify opportunities. This can only happen if leaders create a safe space for experimentation and reward employees for doing so, even if the experiment is not successful.
To further stimulate meaningful participation, federal leaders must move beyond simply adopting technology mandates to engaging employees in solving strategic challenges. Inviting people to committees and creating evaluation teams that include diverse perspectives will help clarify the relationship between experimenting with AI and advancing your mission.
Managing the human side of technology change
The success of AI will depend more on human behavior than any previous technology implementation. Two employees with the same purpose and access can produce vastly different results depending on how they approach technology. Success depends on creativity, experimentation, and integration into your daily workflow.
Therefore, implementing AI is fundamentally a behavior change challenge. Employees need to understand how AI will deliver on the strategic goals they care about, not just an attempt to replace their roles.
AI is evolving much faster than traditional management systems were designed to keep up. They are built to produce reliable and repeatable performance rather than rapid changes. Federal leaders may need to go beyond standard practices by using dynamic experimentation teams, engaging more people to find solutions, and leveraging peer-to-peer communication where employees share their findings with each other.
If agencies can avoid the mistakes of previous administrative fads, AI action plans present an opportunity to accelerate mission execution. Agencies that see AI transformation as a people challenge rooted in strategic clarity, rather than just a technology implementation, will realize real value from their investments.
gaurav gupta I am the research and development manager of cotter.
Copyright © 2025 Federal News Network. Unauthorized reproduction is prohibited. This website is not directed to users within the European Economic Area.
