There is a conversation I have been having a lot lately — in coaching sessions, in classrooms, in boardrooms. It goes roughly like this.

A senior leader describes a problem they are trying to solve. The problem is real, the stakes are high, and they have been working on it for months. They have data. They have consultants. They may have an AI tool giving them analysis and recommendations. And yet the problem persists.

When I ask what has been tried, the answers are almost always variations on the same theme: they have been addressing the visible symptoms — the events at the top of the iceberg — without ever getting to the structures and beliefs that generate those symptoms.

This is not a failure of intelligence or effort. It is a failure of thinking method. And it is the gap I believe is most important for leaders to close right now.

What AI Does Well — and Where It Stops

Let me start with an honest assessment of what AI is genuinely good at, because it is quite a lot.

AI excels at the analytical layer of leadership work: processing large volumes of information, identifying statistical patterns, generating options, summarising complexity, flagging anomalies. Tasks that once took teams of analysts weeks can now be completed in minutes. This is a genuine and significant capability.

But there is a layer of thinking that AI cannot reach — not because of current limitations that future models will overcome, but because of something more fundamental about the nature of systemic understanding.

AI learns from patterns in data. It extrapolates from what has happened. What it cannot do is question the assumptions embedded in the data itself, challenge the frame through which a problem has been defined, or synthesise across domains in a way that produces genuinely novel insight. It cannot look at a recurring organisational problem and ask: what would have to be true about this organisation’s beliefs for this problem to keep happening?

That question — and the thinking required to answer it — is irreducibly human.

The Iceberg Model: A Framework for Deeper Thinking

One of the most useful frameworks I have found for developing systemic thinking in leaders is the Iceberg Model, drawn from the systems thinking tradition.

The model proposes that every organisational problem has four levels:

Level 1: Events. What just happened. The missed target, the failed launch, the team conflict. This is the visible tip of the iceberg — what everyone can see and what most responses address.

Level 2: Patterns and Trends. What keeps happening. When you look across time, the same types of problems recur. The same teams underperform. The same initiatives stall. Patterns suggest that something structural is at work.

Level 3: Systemic Structures. Why it keeps happening. The processes, incentives, role designs, and relationship dynamics that produce the patterns. A sales team that consistently over-promises is operating inside a system where over-promising is rewarded — or where not over-promising is penalised. Changing the structure changes the behaviour.

Level 4: Mental Models. What beliefs make this possible. The deepest layer — the assumptions, values, and worldviews that the people in the system hold, often without being aware that they hold them. These mental models shape the structures, which produce the patterns, which create the events. Lasting change almost always requires work at this level.

Most leadership interventions — and most AI-assisted analysis — operates at Levels 1 and 2. The most transformative work happens at Levels 3 and 4.

Why This Matters More Now, Not Less

There is a tempting conclusion that as AI takes over the analytical work, the role of the human leader becomes smaller. I think this is exactly wrong.

As the analytical layer gets commoditised, the distinctively human layer becomes more valuable. Anyone with access to the right tools can get a good analysis of what happened and what the data suggests. The question of why it happened, whether the right question is being asked, and what would need to change at the deepest level — these questions require human judgment, and human judgment of a specific kind.

The skills that matter most in this environment are:

Systems thinking — the ability to see wholes rather than parts, to understand feedback loops, to trace problems back to their systemic roots rather than applying surface-level fixes.

Critical thinking — the ability to question assumptions, distinguish between evidence and inference, identify when the frame of a problem is itself the problem, and reason from first principles rather than from convention.

Connecting the dots — the ability to synthesise across domains, draw analogies, recognise when a pattern from one context illuminates something in another. This is perhaps the most distinctive human cognitive advantage: the ability to make meaning across different kinds of knowledge.

None of these skills are taught systematically in business schools. None of them are developed by years of functional expertise. And all of them can be deliberately developed — through practice, through coaching, through the kind of structured reflection that most leaders never make time for.

What This Looks Like in Practice

In coaching sessions, I often begin by asking a leader to describe a problem they are currently stuck on. Then I ask: have you seen this pattern before? Almost always, the answer is yes.

The next question is: what would have to be true about your organisation — its systems, its incentives, its culture — for this pattern to keep producing itself? This moves us from Level 1 to Level 3.

Then the hardest question: what would have to be true about what you believe — about leadership, about your role, about how change works — for you to have kept approaching this problem the same way? This is Level 4. This is where the real work begins.

The answers are rarely comfortable. But they are almost always illuminating. And the leaders who are willing to go there — to examine the mental models that have been governing their decisions without their awareness — consistently make the most significant breakthroughs.

A Provocation to Close With

If you are a senior leader, here is the question I would invite you to sit with.

The problems you are currently wrestling with — the ones that feel stuck, that have resisted the usual solutions, that keep coming back in slightly different forms — what would it mean if those problems were not problems to be solved but symptoms to be understood?

What is the iceberg beneath the visible tip?

And what mental model would you need to question before the iceberg could melt?

These are not analytical questions. They cannot be answered by a dashboard or a language model. They are thinking questions — and they are the most important ones on your agenda.

I work on these themes in my coaching practice, in my executive education programmes, and in the EGN Leadership peer groups I chair in Singapore. If this resonated — or if you disagree — I’d genuinely like to hear from you.