The Thinking Partner: How Leaders Can Use AI for Their Own Development

By Hubert Saint-Onge

A Different Kind of Value

Imagine a conversation in which an idea you have been carrying for years suddenly connects with three others you had not seen as related — and in that moment, a framework emerges that is sharper, more elegant, and more useful than anything you had arrived at alone. You did not receive an answer. You did not delegate a task. You thought your way to something genuinely new, through dialogue, with a partner who pushed back at exactly the right moments.

This is what AI makes possible — but very few leaders have found their way to it. Most are using AI for something considerably more modest: drafting communications faster, synthesizing research, preparing briefings, and generating options. These are real and valuable uses. But they represent a fraction of what is available to leaders who approach AI differently.

The argument of this article is straightforward: the most significant value AI offers leaders is not productivity — it is development. Used as a genuine thinking partner, AI can accelerate the growth of the very capabilities most in demand in today’s turbulent environment: wisdom, judgment, metacognitive awareness, and the relational intelligence needed to navigate genuine interdependence. But this requires a fundamentally different relationship with AI than the one most leaders are currently cultivating.

The Context: Why This Matters Now

In a series of articles on leadership in turbulent environments, I have argued that the business environment has entered a qualitatively different era. Volatility is no longer episodic. Ambiguity is structural. Technological acceleration, geopolitical instability, and cumulative complexity now interact in ways that obscure cause-and-effect relationships and invalidate many of the assumptions on which traditional leadership relied.

In this context, adaptability has become the primary strategic capability for organizations and for leaders individually. The demands on leaders are not simply greater in volume; they are greater in kind. The sheer level of interdependence that must be managed — across teams, functions, organizations, and, increasingly, the boundary between human and artificial intelligence — taxes even the most capable leaders. Sensemaking has become a continuous, collaborative discipline rather than a periodic exercise. The quality of judgment under uncertainty has never mattered more.

What this means for leadership development is significant: the pace at which leaders need to grow has outrun what traditional approaches — executive education, coaching, structured experience — can adequately address on their own. Leaders need to develop faster, not just work faster. The question is how?

The Dependency Trap: How Most Leaders Are Using AI

Most leaders, given access to a powerful AI system, gravitate instinctively toward the same relationship: they ask questions and receive answers. They describe a problem and receive recommendations. They request a draft and receive one. Over time, this pattern becomes habitual — and its consequences are easy to miss precisely because the outputs are often impressive.

The problem is not the quality of AI’s responses. It is what this pattern does to the leader over time. When thinking is consistently delegated rather than exercised, it atrophies. When assumptions are never interrogated, they calcify. When a leader reaches for an answer before they have fully sat with a question, the question never does its developmental work.

There is a compounding factor that makes this particularly insidious for leaders: AI’s default posture is sycophantic. It is designed to be engaging, affirming, and responsive. Leaders who are already surrounded by people who tell them what they want to hear now have a sophisticated new source of validation — one that expresses agreement in articulate, well-organized prose. The result is the illusion of rigorous thinking without its substance.

This is not a technology problem. It is a relationship problem. The dependent relationship most leaders bring to AI undermines the very capabilities they most need to develop. A different relationship is possible — and its effects are categorically different.

The Missing Capability: Metacognition

Among the capabilities that distinguish the most effective leaders, one stands out as both critically important and systematically underdeveloped: metacognition. Defined simply, metacognition is thinking about one’s own thinking — the ability to observe one’s reasoning as it unfolds, recognize the assumptions embedded in it, notice the emotional forces shaping it, and calibrate judgment in real time.

Most leaders have developed this capability least of all. This is not accidental. The entire professional environment of a leader rewards decisive forward movement. Meetings, deliverables, decisions, results — the rhythm of organizational life leaves little space for the reflective pause that metacognitive awareness requires. Leaders who pause to examine their own thinking are often perceived, at least initially, as less decisive than those who do not.

Yet it is precisely this capacity that separates experience which produces wisdom from experience which accumulates. A leader can navigate two decades of complexity and emerge with either deeply refined judgment or deeply entrenched habits — the difference lies in whether they examined their own thinking along the way. Wise leaders embody the principle of going slower to go faster. Their reflective practice is not a detour from effectiveness; it is its foundation.

For action-prone leaders, this reframe is essential. Metacognition is not reflection for its own sake. It is the mechanism through which experience is converted into judgment, and judgment into wisdom. Without it, the leadership development loop — experience, reflection, insight, refined action — breaks at its most critical point.

What the interdependent AI partnership uniquely offers is a structured opportunity for metacognition that action-oriented leaders would otherwise bypass entirely. The act of externalizing your reasoning to a genuine thinking partner — articulating what you believe and why, having it reflected back, being asked what you are assuming, noticing where you resist the challenge — is metacognition made practical and accessible. The AI does not slow the leader down. It gives the leader’s thinking somewhere to go, generating insight rather than simply output.

The Foundation: Knowing Who You Are

Before a leader can engage AI as a genuine thinking partner, a prior condition must be met: they need to know what they are bringing to the partnership. This means knowing their values — not as a list of abstract ideals, but as a living understanding of what matters most to them and why.

In my work with more than five hundred leaders over several decades, I have been consistently struck by how many senior leaders — people of real capability and accomplishment — cannot name or articulate their own values when directly asked. This is not a failure of character. It reflects the fact that values operate largely below the surface of conscious awareness. As Brian Hall’s pioneering work on human development demonstrated, values are best understood as prioritized ideals — and their order of priority shapes every significant choice a leader makes, whether or not the leader can see the connection.

The implication for the AI partnership is direct. A leader who has not done the work of values clarification enters this relationship without a reliable anchor. They cannot consistently recognize when AI output resonates with what they genuinely stand for, and when it pulls them in a direction that is subtly wrong for who they are. Without that anchor, the partnership cannot be fully trusted — and the leader cannot bring their full self to it.

A useful touchstone here is the Amish community’s approach to technology. The Amish are widely misread as technophobes, as people who reject modernity. In reality, their relationship with technology is far more sophisticated. They engage each new technology deliberately, asking not whether it is efficient or powerful, but whether adopting it strengthens or weakens what they care about most as a community. This values-first framework allows them to engage technology from a position of grounded confidence rather than reactive fear.

Leaders who have done the work of values clarification can engage AI from a similar position. They know what they stand for. They can explicitly bring those values into their AI partnerships, establishing them as the context in which the dialogue unfolds. And they can recognize dissonance when it arises — the moment when something AI offers, however well-reasoned, does not fit who they are. The clarity of one’s values transforms AI engagement from a potential source of vulnerability into a relationship of purposeful confidence.

The Interdependent Partnership

With metacognitive awareness developing and values providing the foundation, a genuinely different relationship with AI becomes possible — one defined not by dependence but by interdependence.

The distinction matters precisely because interdependence is the leadership capability most urgently needed in the turbulent environment described earlier. The most effective leaders are not those who impose their will or defer to others, but those who build shared understanding through genuine exchange — holding their perspective firmly enough to contribute it, and openly enough to be changed by what they encounter. This is what interdependence requires between people. And it is exactly what a genuine AI partnership requires.

In a dependent relationship with AI, the leader asks and receives. In an interdependent one, the leader thinks alongside. They bring a perspective and expect it to be challenged. They articulate an assumption and invite its interrogation. They resist the first good answer in favour of the better thinking that friction produces. They remain, throughout, the author of their own conclusions — but conclusions that are richer, more tested, and more trustworthy for having been genuinely contested.

What makes this particularly powerful as a development practice is the transfer effect. The relational moves that genuine AI partnership requires — holding your ground while remaining genuinely open, building shared understanding across difference, tolerating the productive discomfort of having your thinking challenged — are exactly the moves that high-stakes human interdependence demands. Leaders who practice these moves with AI are rehearsing for the most demanding relational challenges in their professional lives. The practice is not merely an analogy for effective leadership. It is its direct cultivation.

At its best, an interdependent dialogue with AI produces thinking that neither party could have reached alone — new connections, unexpected reframes, insights that emerge from the friction of genuine exchange. This is the experience that leaders who have discovered it describe as exhilarating. It is qualitatively different from anything a productivity tool can offer. And it is available to every leader willing to bring the right orientation to the relationship.

What This Looks Like in Practice

Building an interdependent AI partnership is not primarily about technique. It is about orientation: a set of commitments that the leader consistently brings to every exchange. That said, specific practices help make the orientation concrete, especially for leaders who are building new habits against the grain of action-prone instincts.

Structured reflection after high-stakes moments. Rather than moving immediately from one challenge to the next, the leader takes a deliberate pause with AI after significant moments — a difficult conversation, a decision under pressure, a moment of reactive leadership. Not to receive an assessment, but to think out loud: what happened, what they were assuming, how they might have read the situation differently, what they would do again and what they would change. The AI’s role is to hold the space and ask the questions that deepen the reflection.

Assumption interrogation. Before committing to a significant course of action, the leader surfaces and examines the underlying assumptions. What would have to be true for this to be the right choice? Where am I most confident — and why? Where might I be wrong? AI is genuinely effective at identifying the load-bearing assumptions in a line of reasoning and pressing on them. The leader’s discipline is to invite that pressure rather than resist it.

Sensemaking dialogues. In conditions of ambiguity, the leader uses AI as a thinking partner for sensemaking — not to receive a definitive interpretation, but to work through competing readings of a complex situation, test them against each other, and arrive at a more grounded perspective. This is the practice that most directly builds the sensemaking capability that a turbulent environment demands.

Pre-mortem thinking. Before a critical decision or initiative, the leader imagines it has gone badly and works backward with AI to understand how. What would most likely cause this to fail? What are we not seeing? What assumptions are we least willing to question? This practice systematically surfaces the blind spots that confidence and commitment tend to obscure.

Values-grounded check-ins. Periodically, the leader pauses to ask whether the thinking they are developing is authentically theirs — whether it reflects their values, their hard-won judgment, their genuine perspective. This is the metacognitive practice that keeps the partnership from becoming a subtle form of dependence. The anchor is always: " Is this mine?

In each of these practices, the discipline is the same: resist the pull toward the first satisfying answer, stay in the productive discomfort of genuine inquiry, and treat the exchange as a thinking process rather than a production process.

The Developmental Payoff

Leaders who cultivate interdependent AI partnerships do not simply become more effective in those conversations. They develop capabilities that transfer — and that compound over time.

Metacognitive awareness deepens. The habit of examining one’s own thinking, practiced consistently in AI dialogue, becomes more available in the moments that matter most: under pressure, in conflict, when certainty is most tempting and most dangerous. Leaders become more reliably able to notice what they are assuming, recognize when emotion is distorting judgment, and distinguish between conviction that has been tested and conviction that has merely been held.

Judgment sharpens. The consistent practice of interrogating assumptions, running pre-mortems, and thinking through complexity without rushing to resolution produces a quality of judgment that experience alone rarely generates. The leader becomes, over time, someone whose conclusions are worth more — because they have been arrived at more honestly.

Sensemaking capacity grows. The ability to navigate genuine ambiguity — to hold multiple interpretations, test them against evidence, and act without premature closure — is directly developed through the practices described above. In a turbulent environment, this is not a peripheral capability. It is at the core of strategic leadership.

Relational intelligence deepens. The leader who has learned to engage AI as a genuine thinking partner — holding their perspective while remaining open, building shared understanding through dialogue, welcoming rather than deflecting challenge — is practicing the exact relational moves that high-stakes human interdependence requires. The transfer to leadership with people is direct and meaningful. In an environment defined by the complexity of organizational interdependence, this development may be the most valuable of all.

An Invitation

The leaders who will develop fastest in the coming decade are not necessarily those with access to the best information, the largest networks, or the most sophisticated tools. They will be those who have learned to think rigorously and reflectively, and to work in genuine partnership with others, human and artificial.

AI offers every leader something previously available only to those fortunate enough to have an exceptional thinking partner: a capable, available, and endlessly patient interlocutor who can hold the other side of a genuine intellectual exchange. Whether that offer becomes a developmental gift or a sophisticated shortcut depends entirely on the leader.

The starting point is not a new technique. It is a prior question: do you know what you stand for clearly enough to bring it into a genuine partnership? From that foundation — values clarity, the willingness to reflect, and the discipline to engage with friction rather than avoid it — the practices described in this article become available. And their effects, sustained over time, are profound.

The invitation is to begin. Not to use AI better, but to use it differently — as a thinking partner rather than an answer machine, as a space for genuine reflection rather than efficient production, as a relationship that develops you rather than one that merely serves you.

In a turbulent world that places extraordinary demands on leadership, the ability to learn faster, think more clearly, and develop more deliberately is no longer optional. AI, approached with the right orientation, is the most accessible and powerful development instrument leaders have ever had.

The question is not whether to use it. It is who you will be when you do.

Hubert Saint-Onge is a consultant specializing in leadership and organization development. His work on leadership in turbulent environments, values-based leadership, and the relationship between human wisdom and artificial intelligence is available at saintongealliance.com.