The Most Overlooked Talent Development Opportunity Is Already on Your Team

By D.A. Smith-Hemphill, PhD

There is a growing consensus about AI and talent: Teach people to manage AI. Treat it like a capable new hire. Set the objective, delegate the work, review the output, and own the result. Learning to manage AI this way is the essential first step.

There is a next-level move: Having learned to manage AI, they are also learning to think with it. Managing AI helps people produce more of the work they already know how to do. However, thinking with AI helps people do better thinking than they could do alone. The first is delegation. The second is partnership.

That partnership is a talent development frontier, and the raw material is likely already on the team.

The Gap Is Not Technical

It is tempting to frame the AI skills gap as a tooling problem or training budget problem. The evidence points somewhere more human.

AMA’s 2025 research found that nearly 95% of organizations report some level of AI use, and 58% use it daily across the enterprise. AI training has also scaled quickly, with 78% of organizations offering AI training in 2025, up from 7% two years earlier. Yet 57% of employees still say they feel behind.

Adoption is widespread. Training is widespread. The deeper capability is knowing how to think alongside a thinking machine.

From Managing AI to Thinking with It

Delegation works when the task is already understood. The person knows what “done” looks like, assigns the work, inspects the output and accepts or rejects it. The thinking happened before the handoff.

Partnership is different. It is useful when the problem is not yet fully understood. The person is not outsourcing execution. They are thinking out loud with something that can think back.

That may mean asking AI to argue the opposite position, identify weak reasoning, surface an untested assumption or play the role of a skeptical stakeholder. The output is not the full payoff. The real value is the sharpening of human judgment along the way.

I watched this process evolve with one client, a senior VP preparing to reorganize her team. She first used AI to draft an announcement. The result was serviceable but unsatisfying. When she used AI instead to think through the reorganization, asking it to play the roles of a skeptical employee, a peer VP and her own boss, it surfaced objections and assumptions she had not considered. She rewrote the plan before rewriting the announcement.

The draft was never the point. The pre-mortem was.

The Risk Inside the Easy Version

When AI use becomes reflexive delegation—offload the problem, accept the first plausible answer, and move on—it can quietly make people think less.

Partnership is the antidote. When AI is used to challenge reasoning rather than replace it, every exchange can strengthen judgment rather than outsource it. The same capabilities that help someone draw out another person also help them draw out a machine: clear framing, good questions, humility, judgment and ownership of the result.

The Five Competencies to Build

If AI partnership is a developable capability, it can be taught. Five competencies do most of the work:

Framing is the ability to state a problem clearly before reaching for a solution. It means giving context, intent and constraints rather than a one-line command.

Curiosity and productive skepticism is the discipline to treat a confident answer as a first draft, not a verdict.

Iterative dialogue is the habit of staying in the conversation past the first exchange by pressing for sharper options.

Judgment and accountability is the willingness to remain editor-in-chief, deciding what is wise and what goes out under one’s name.

Intellectual humility is the openness to discover that an assumption may be wrong without experiencing that discovery as a threat.

None of these competencies is purely technical. They are durable human skills—communication, decision making, adaptability and accountability—revealed as the operating system for working with intelligent tools.

What This Means for Training

If AI partnership is a thinking capability dressed up as a technical one, then AI training needs to move beyond features and prompt recipes.

Organizations with strong coaching cultures already have a head start. Where managers give context, ask good questions and invite challenge, the leap to genuine AI partnership is shorter. Where those muscles are weak, more tooling will not close the gap.

The opportunity is to make AI partnership deliberate. Pick one team. Name the five competencies as capabilities to develop and hire for. Practice them until they become habit.

Organizations that take this approach will outthink those that only deploy it.

Frequently Asked Questions

What is the difference between managing AI and thinking with AI?

Managing AI is delegating work the person already understands. Thinking with AI is using it as a partner to challenge assumptions, sharpen reasoning and improve judgment.

Why is the AI skills gap not just technical?

AMA research cited in the article shows AI use and AI training are both widespread, yet many employees still feel behind. That suggests the gap is less about access and more about depth of thinking.

What are the five competencies for AI partnership?

The five competencies are framing, curiosity and productive skepticism, iterative dialogue, judgment and accountability, and intellectual humility.

Why does accepting the first AI answer create risk?

Accepting the first plausible answer can make people think less. Using AI well requires questioning the output, testing assumptions and staying accountable for the final decision.

What is the biggest takeaway?

The most overlooked talent development opportunity is already on the team: building the human capability to think with AI, not just use it.