AI Is Reshaping Leadership Development – But Not How You Think

Angelina Eng recently posed a question in MarTech that should concern anyone responsible for developing leaders: as AI absorbs the analytical and operational work that junior professionals used to cut their teeth on, who is building the judgment those professionals will need later? The question lands differently when you consider that most leadership pipelines assume people arrive with years of messy, hands-on problem-solving behind them.

Organizations are racing to embed AI across every function. The efficiency gains are measurable. But underneath the optimization runs a quieter loss: the ambiguous, unglamorous work that served as the training ground for leadership judgment is disappearing. We are engineering away the experiences that build leaders - and few organizations have noticed the gap.

The Flattening Paradox: Fewer Layers, Fewer Learning Grounds

McKinsey senior partner Alexis Krivkovich recently described AI as giving leaders “a superhuman capacity to manage across bigger scopes.” The implication, reported in Business Insider, is that companies can now flatten their structures and strip out the management layers accumulated over the past decade. The Times of India amplified the same message: it is time to cut.

The logic is straightforward. If AI can summarize, analyze, and route information that used to require a human layer, those layers become expensive overhead. Wider spans of control. Faster decisions. Leaner structures. From a cost and speed perspective, the math works.

From a leadership development perspective, it creates a problem nobody is solving. Each management layer that gets removed was also a development stage. Mid-level roles were where emerging leaders practiced judgment under real conditions - handling ambiguity, navigating conflict, making decisions with incomplete information, and learning to fail safely with limited blast radius. Those were not management layers. They were proving grounds.

The irony is hard to miss. The same organizations investing millions in leadership development programs are simultaneously dismantling the structures where leadership actually developed. They are replacing organic development with programmatic development and assuming the two produce equivalent results.

They do not. A leadership development program teaches frameworks. A middle management role teaches you what happens when the framework meets a real situation and breaks down. The program gives you tools. The role gives you judgment. AI can accelerate the first. It has no way to substitute for the second.

Fujitsu’s research on the AI multiplier effect underscores the stakes: leadership choices about how AI is deployed shape long-term organizational impact. If those choices are made by leaders who never had the developmental experiences that build sound judgment, the multiplier works in reverse.

From Skills to Capacity: What AI Cannot Replicate

Jennifer Britton, writing for the Coaching Tools Company, draws a distinction that cuts to the center of this challenge. The gap is not whether leaders should learn coaching skills. The question is whether organizations are building coaching capacity - across leaders and peers - as a fundamental part of how they operate.

The distinction matters more than it might appear at first. Coaching skills are teachable. You can learn to ask open questions, practice active listening, develop the discipline of withholding advice. These are techniques, and they have value. A half-day workshop can introduce them. A good training program can build competence.

Coaching capacity is something different. It is the underlying ability to operate in a certain way consistently, especially under pressure. Capacity is what remains when conditions are uncertain, when the playbook does not apply, when the person across from you needs something that no template can provide.

Skills give you language and structure. Capacity gives you the steadiness to use them when it counts.

This distinction becomes critical as AI reshapes work. AI handles structured, predictable tasks well - summarizing, drafting, analyzing, generating options. The work that remains for humans is unstructured: navigating tension, holding paradox, making decisions in flight while the situation is still forming. That work requires capacity, not skills.

Most leadership development investments still target skills. They produce people who know the right frameworks but freeze when conditions get messy. The coaching world has a name for this gap. Coaches who can follow a model but cannot hold the space when a client brings genuine uncertainty have skills without capacity. The same dynamic plays out in organizations. Leaders trained on frameworks who cannot adapt when the framework fails. Managers who know the right questions but cannot sit with the discomfort of not having the answer. The skills are present. The capacity is missing. And AI is not going to build it for them.

Leadership Development as a Design Problem

London Business School research examining AI implementations at UltraTech Cement and Crisil reached a conclusion that extends well beyond technology strategy: AI advantage comes not from acquiring better models but from designing how AI is used in real decisions and under what conditions it can be trusted. The same principle applies directly to leadership development.

Most organizations treat leadership development as a procurement problem. Select a program. Send people through it. Measure completion. Repeat. The approach mirrors how many of those same organizations first approached AI: which vendor, which model, how fast can we deploy?

The more productive question - for AI strategy and leadership development alike - is a design question. What conditions need to exist for the outcome we want? For AI, that means designing decision-making processes where humans and algorithms complement each other. For leadership, it means designing environments where people actually develop the judgment, resilience, and relational capacity the organization needs.

CIO.com recently reported on how analytics and AI are reshaping the boundaries of IT leadership - the recognition that where leadership begins and ends is shifting. That observation applies far beyond IT. As AI absorbs operational and analytical work across functions, the boundary of what leadership means is changing everywhere. Organizations that treat this as a training problem will keep buying programs. Organizations that treat it as a design problem will redesign how leaders grow.

The design approach asks different questions. Instead of “what program should we send leaders through?” it asks “what experiences develop the judgment we need?” Instead of “how do we train people on AI tools?” it asks “how do we create the conditions where leaders learn to make good decisions alongside AI?” The answers rarely look like a curriculum. They look like deliberate assignments, coaching relationships, real authority over real problems, developing executive presence, and structured reflection on what happened.

Where Leaders Actually Develop Now

Venture capitalist John Doerr, speaking at Rice University’s Doerr Institute for New Leaders, returned to a principle that sounds simple but carries significant weight: leadership is about meaning, and meaning is built through relationship. Citing Jim Collins, he described leadership as “the art of getting people to want to do what must be done” - influence over control.

If traditional management layers are shrinking and programmatic training cannot fill the gap on its own, where does leadership development actually happen? The answer is not new, but it requires renewed attention: leaders develop through relationships.

Coaching relationships. Mentoring relationships. Peer learning. Deliberate practice on real decisions with someone alongside who can help them process what happened and why. Doerr credited much of his own development to influential mentors - Andy Grove, Bill Campbell - people who shaped his judgment not through programs but through sustained relational investment.

The temptation in an AI-optimized environment is to reach for AI-delivered solutions. itel, a BPO company in Jamaica, recently launched AI-powered personalized training podcasts as part of its development strategy. There is nothing wrong with AI as a complement - their approach addresses knowledge transfer at scale, and that matters. But knowledge transfer is not the same as developing the internal capacity that makes a leader trustworthy under pressure.

What develops that capacity is harder to scale and impossible to automate: a coaching conversation where someone is genuinely challenged. A mentor who tells you the truth about what they observed. A peer group where leaders bring real problems and receive honest feedback. Structured reflection after a high-stakes decision, when the temptation is to move on to the next one.

These development mechanisms share a common feature. They require another human being who is paying attention, who has their own capacity for presence and challenge, and who can hold the space for someone else to grow. No AI system replicates that. The question is whether organizations are investing in it.

Building Coaching Capacity Across the Organization

Britton’s work points toward a practical response: shared coaching capacity distributed across leaders and peers, not concentrated in a handful of executive coaches serving the top of the organization. Fujitsu’s research reinforces this through its framing of leadership choices as multipliers - the decisions leaders make about how AI is used compound over time, for better or worse.

The traditional model places coaching at the top. A CEO gets an executive coach. A few senior VPs might receive coaching as part of a leadership transition. The rest of the organization gets training programs, performance reviews, and maybe a mentor if they are lucky. That model was already insufficient. In an AI-restructured organization where leadership judgment needs to be distributed more widely, it is completely inadequate.

Building coaching capacity across an organization means something specific. It means developing leaders at every level who can hold a coaching conversation, not just leaders at the top who receive one. It means peer coaching relationships become a normal part of how teams work. It means managers learn not just to direct and evaluate but to help the people around them think through ambiguity and develop their own judgment.

The BCG finding that 68% of communications leaders describe their function as an AI laggard points to a wider pattern. The gap in most organizations is not technical capability. It is human capability - specifically, the ability to develop people in the ways that matter most when AI handles everything else.

Microsoft’s observation that AI has shifted from tool to business strategy applies to coaching capacity as well. When coaching capacity becomes organizational infrastructure rather than an individual benefit, it stops being a perk and becomes part of how the organization sustains itself. Leaders who can coach develop other leaders who can coach. The capacity compounds. In an environment where AI handles the structured work, this human multiplier may be the most important investment an organization makes.

The Question That Matters

The question most organizations are asking about AI and leadership development is: “How do we use AI to develop leaders faster?” It is the wrong question.

The right question is whether we are building the conditions where real leadership still develops - the relational depth, the tolerance for ambiguity, the capacity to sit with not-knowing long enough for genuine insight to emerge. AI can deliver content, simulate scenarios, and personalize learning paths. None of that builds the internal architecture that allows a leader to be steady when everything around them is uncertain.

Every organization flattening its structure, automating its workflows, and optimizing its processes is also deciding - whether intentionally or not - what kind of leaders it will have ten years from now. The organizations that thrive will be the ones that recognize coaching capacity as essential infrastructure, not optional enrichment.

The efficiency gains from AI are real. The development gap they create is also real. Closing that gap is not a technology problem. It is a leadership design problem. And it starts with the decision to invest in the human capacity that no algorithm can build.