AI may drive productivity in management, but what about leadership?

A first draft used to require a blank page, a problem to solve and enough time to wrestle with both. Now, an AI tool can produce something usable in seconds. That is a huge shift for business. Employees can move faster, remove repetitive effort and spend less time getting from zero to a respectable first version. For managers focused on productivity, the attraction is obvious.

The harder question is what happens when respectable becomes the standard. Zynga founder Mark Pincus recently captured the issue when he said AI can get people to a “B-plus in seconds”. His warning was that people may struggle to recognize an A if they have never learned how to create one.

That distinction should concern every leader adopting AI at scale. The biggest risk is not that AI produces poor work. It is that it produces work that is good enough so quickly that people stop developing the expertise needed to know when good enough is not enough.

Efficiency and adaptability are not the same

Some work is designed for efficiency. A process is known, the desired outcome is clear, and success means completing the task accurately and consistently. AI is well suited to this kind of work.

If a system can process information faster, summarize routine material or handle a repeatable task more reliably, there is little value in protecting manual effort for its own sake.

But leadership becomes more difficult when there is no clear process to follow, or even more challenging, when what worked before is now obsolete.

Markets shift. Customers behave differently. A strategy stops working. A competitor changes the rules. In those moments, the organization does not need people who are simply good at following instructions. It needs people who can make sense of ambiguity, apply judgement and decide what to do next. People who can leverage a deep understanding of what is and combine it with a radical vision of what could be. That is adaptive performance, and it is much harder to automate.

What elite performers understand about technology

Sport offers a useful comparison. Top athletes now have access to extraordinary amounts of information. Their training can be measured in minute detail. Video can expose technical weaknesses. Sensors can reveal patterns that would once have been invisible.

Yet the data alone does not create an elite performer. Competitors can often access similar tools.

The difference comes from how the athlete interprets the information and what they do next. They test adjustments, learn from feedback and build judgement through repeated practice. Over time, they understand not only what works, but why it works. That matters because competition does not stand still. The athlete must be able to respond when conditions change.

Business is no different. If employees repeatedly hand over the difficult part of thinking to a machine, they may still produce competent work. What they lose is the experience of making choices, testing assumptions and learning why one answer is stronger than another. Moreover, they lose the ability to innovate and adapt when the situation requires a new type of solution. AI is great at aggregating the “standard” solution in the market, but it can’t see beyond what already exists. Those who rely on AI for the critical thinking part of a process, are at high risk of being infected with the same limitation.

The inefficient moments often build the skill

Many of the experiences leaders are tempted to remove are the same experiences that develop capability. Writing the first draft yourself forces you to organize an argument. Making a decision with incomplete information builds accountability. Getting something wrong, then understanding why, creates a reference point that can improve the next decision.

These moments are not always efficient, but they are essential for high levels of skill. That is why they are valuable. If AI removes too much of the struggle, people can arrive at an answer without understanding the route that produced it. That is manageable while the task remains predictable. It becomes a weakness when the context changes and there is no obvious precedent. At that point, speed matters less than expertise.

Use AI to make people stronger

The answer is not to slow adoption or protect work that technology can do better. Leaders should be more precise about what they automate. A useful question is whether the task is primarily about execution or innovation.

If the value lies in speed, consistency or removing repetitive work, AI should do more. If the value lies in thinking out of the box, the human still needs enough involvement to build skill. Leaders also need to keep raising expectations. As AI makes competent output easier to produce, competent output will become less distinctive. Advantage will come from people who can spot what is missing, ask better questions and keep improving when others accept the first plausible answer.

That requires exposure to excellent work, useful feedback and space to make decisions before turning to the machine. This has practical implications for how teams work. Before asking AI for an answer, employees can form an initial view of the problem. After AI responds, they should explain what they agree with, what they would change and why. That keeps the tool in the role of sparring partner rather than substitute, while giving managers something more useful than faster output: evidence of how their people think.

The strongest organizations will not be those that automate the greatest number of tasks. They will be those that are clear about what should be automated and what must still be led by human expertise. AI can help people reach a solid answer faster. Leadership is making sure they still know how to reach a better one when the situation demands it.