Why workforce planning should shift from jobs to human-AI work systems

Most workforce planning conversations start with familiar questions: How many people do we need? Where are we understaffed? Which roles should we prioritize? These questions have defined workforce planning for generations. But in an environment increasingly shaped by AI, they can lead leaders toward the wrong decisions.

The biggest workforce risk today is not a talent gap or even a skills gap. It’s a design gap.

AI is changing the nature of work in ways that are easy to underestimate. It’s not replacing entire jobs overnight, but it is rapidly reshaping the tasks individuals perform within those jobs. Some work can be done faster. Some work is going away. And some work is becoming more dependent on human judgment than ever before.

This creates a different kind of decision than workforce planning has traditionally required. The question is no longer simply, What workforce do we need? Instead, it becomes: What is the best way for this work to get done—and how should humans and AI work together to make it happen?

That shift may sound subtle, but it isn’t. It moves workforce planning out of staffing exercises and into work design—how work flows through the organization, what gets done by whom, and where judgment sits. For leaders, this reframing turns workforce planning into a governance and design decision, where accountability is maintained, and outcomes matter more than headcount.

We can call it a human-AI work system, an intelligent orchestration of technology and humans working together to accomplish the mission.

Consider a familiar scenario in which workload is rising in an area like program integrity. The default response is often to add capacity. More investigators and analysts will catch more instances of fraud. While true, federal leaders don’t often have the luxury to fluidly add and reduce manpower, nor is that always the most effective move.

A more useful starting point is to look closely at the work itself. What you typically find is a mix of routine elements that follow patterns; analytical components that benefit from speed and scale; and moments that require judgment, context, and accountability. AI performs well on the first two, and humans remain essential to the third.

The leadership challenge is to rebalance the work, letting technology take on what it does well so that people can focus where their judgment actually matters. This is where many organizations get stuck. They introduce AI into existing workflows without changing how the work is structured. The result is more tools layered onto unchanged processes, with limited improvement in outcomes or decision quality.

If that sounds familiar, it’s a sign of the design gap.

You can see it in the questions an organization asks. If the conversation is still centered on questions like How many people do we need? or How do we fill this role faster?, then the workforce planning model hasn’t really changed.

A different set of questions signals a different approach:

• Where do we need more judgment, not more activity?
• Which decisions deserve more attention, not just faster processing?
• How do we focus human effort on the work that actually drives outcomes?

Those questions lead to different decisions. They shift focus away from simply changing the size of the workforce and toward reshaping how work is done. For senior leaders, this reframes workforce planning from predicting headcount to deciding where judgment is essential and how work should be governed as conditions change.

Roles will shift, work will be redistributed, and some things will compress while others expand. It’s almost impossible to fit neatly into traditional workforce planning models. Workforce planning becomes less about predicting how many people are needed and more about designing systems that adapt as the work itself evolves.

Putting this shift into practice

For senior leaders, this shift comes down to a few practical decision checks.

• First, ask your teams to break down critical work into its core components. Not at an academic level, but enough to distinguish between routine activity, scalable analysis, and judgment-heavy decisions.
• Second, challenge whether AI is being used to redesign work or simply accelerate it. Faster processes are not always better processes. The goal is not speed for its own sake, but better outcomes.
• Third, consider the supporting infrastructure—governance, training, and decision support—that needs to be put in place to ensure the workforce is prepared to apply good judgment where it is most needed. Ironically, that may mean some upskilling in AI but an even greater focus on imparting real-world experience.

None of these require a wholesale transformation. But together, they begin to tackle what will inevitably be an ongoing design challenge as AI advances, new tools emerge, and work continues to evolve.

The agencies that move forward most effectively will not be the ones that simply adopt AI faster. They will be the ones that are more deliberate about how work gets done. They will move beyond planning jobs and toward designing human-AI work systems that adapt as missions evolve—and that make accountability and judgment clearer, not harder, as technology advances.