For years, organisations rewarded people for producing outputs — reports, decks, analyses — and we quietly conflated the production of work with the thinking behind it.

Organisations that treat AI as a magic answer machine are exposing a leadership gap before a technical one. The real question is not what tasks AI can take over, but what uniquely human capabilities organisations have under-invested in for a decade. That is where the real gap lives.

AI now generates the output in seconds, and what is left exposed is whether the person actually understands the problem, can judge whether the answer is sound, and can decide what to do with it. The gaps I see most clearly are in critical evaluation, contextual judgment, and the ability to ask a sharp question. Many capable professionals can no longer point to the moment where their distinct value sits, because the visible part of their work has been automated.


There is also a widening gap in what I would call human discernment: the capacity to sit with ambiguity, to weigh competing interests, to read the politics and emotion in a room, and to make an ethical call when the data is silent. As a people leader — and through EmpowerThriveHub, the advisory practice I am building for founders and business leaders — I see this play out constantly.

The leaders who thrive are not the ones with the most sophisticated tools; they are the ones who can still think clearly under pressure, hold a difficult conversation, and earn trust — and those are precisely the capabilities AI cannot replicate and cannot hand to them. The third gap is data and AI literacy itself: not coding, but knowing what a tool can and cannot be trusted to do

First, I would gently challenge the word "train." Most companies default to a one-off AI tools workshop, tick a box, and wonder why nothing changes. Managers do not need to be trained on prompts; they need to be developed as leaders of work that is now part human, part machine.

That is a different and more demanding task. The starting point should be self-awareness, not tools. A manager who does not understand their own decision-making style, biases, and behaviour under pressure will simply automate their blind spots faster. Employees also do not trust AI because of the technology itself. They trust the leaders introducing it.

Practically, manager development should focus on three areas. The first is fluency: hands-on exposure so managers actually use AI on real work and lose their fear of it. The second is judgment: learning when to trust AI output, when to challenge it, and how to remain accountable for decisions it informs. The third is human leadership: coaching, giving meaningful feedback, having difficult conversations, and protecting psychological safety as roles evolve.

Those capabilities are also what build trust. Managers need to communicate openly about why AI is being introduced, what it will and will not do, and how it affects people's roles. Employees should help shape how AI is used, understand how important AI-assisted decisions are made, and know there is always human accountability.

AI can scale existing bias as easily as it scales efficiency, making transparency and explainability essential. AI changes daily work, but it also increases the need for managers who can lead people through uncertainty with honesty, consistency, and sound judgment. This development should be continuous rather than a single event, using short, applied learning and peer discussions based on real situations. People remember what they practise, not what they are told, and they trust what leaders consistently demonstrate rather than what they announce.

This question keeps many HR leaders awake, and rightly so.

When AI dramatically increases what a person can produce, output volume becomes a poor measure of contribution. If everyone can generate ten reports in the time it once took to write one, counting reports tells you very little. HR has to shift its focus from output to value, judgment, and outcomes. The questions become: did this person solve the right problem, did they use AI with sound judgment, and did their work move the business or people forward in a meaningful way?

In practice, performance frameworks should be rebuilt around capabilities AI cannot deliver: quality of decision-making, originality of thinking, collaboration, ethical conduct, and human leadership. The contribution of the person also has to be separated from the contribution of the tool by recognising how effectively someone directs, challenges, and improves AI rather than rewarding the machine's raw output. I am pragmatic about AI-enabled monitoring. Used transparently and proportionately, it can help teams improve, but it should never replace performance evaluation because measuring activity is not the same as measuring value, and excessive monitoring quickly erodes trust. Any AI input into performance or disciplinary decisions must remain explainable, contestable, and subject to human review.

Strong leadership in this moment is profoundly human, perhaps more than at any other point in time. The temptation is to believe that AI calls for more technical leaders.

I would argue the opposite: as machines take on more technical work, leadership becomes increasingly about meaning, trust, and direction. A strong leader is first a sense-maker. People are anxious about what AI means for their relevance and livelihoods. Leaders who acknowledge those concerns honestly, provide a clear view of where things are going, and help people see their place in that future offer something no algorithm can: purpose and belonging.

Second, strong leadership is ethically grounded. The pace of AI adoption creates pressure to move fast and ask questions later. Leaders who uphold fairness, transparency, and the dignity of their people build organisations that endure. My work in Rotary, particularly in governance, ethics, and inclusion, has reinforced my belief that values are not a constraint on performance but its foundation. Third, strong leaders remain learners. They do not pretend to have mastered a technology that changes every week; they model curiosity, admit uncertainty, and create the psychological safety for people to experiment and learn.

Above all, they keep people at the centre, using AI not to extract more from fewer people but to create space for more meaningful, more human work. That, to me, is what strong leadership looks like today.

Conversations with the executives, authors, and operators who run complex organizations.

Explore our current leadership spotlight.