What Agentic AI Reveals About Leadership Nobody Is Talking About

I was recently shortlisted to speak at a CIO conference on Agentic AI - the kind of AI that doesn't just respond to prompts, but acts autonomously on your behalf, making decisions and taking actions without a human in the loop. My first response, if I am honest, was something close to panic. I am not an AI expert. I have never claimed to be. And the idea of standing in front of a room full of technology leaders and speaking about artificial intelligence felt like a significant overstep.

But then something shifted. I started to think about what I actually know - about alignment, about the conditions that allow genuine agency to emerge - and I realised that every model I have built over the past decade maps directly onto the AI conversation. Not as a metaphor. As a framework. And in some ways, the stakes are higher in AI than they have ever been in human leadership.

That realisation sent me back to a question I have been sitting with for a long time. One that applies as much to leaders as it does to the systems they are beginning to deploy.

Are you genuinely choosing - or are you just moving?

Agency is one of those words that gets used so freely it has almost lost its meaning. We talk about leaders having agency, teams having agency, organisations needing to build a culture of agency. But genuine agency is not simply the capacity to act. Inherent in real agency is choice - the freedom to act, and equally, the freedom not to. That second part is the part most leaders overlook.

Alignment is what makes that choice real.

A leader who is aligned - grounded in a coherent identity, a clear strategy, and execution that consistently reflects their values - acts from a genuine internal source. They can choose their response rather than simply react to circumstance. And crucially, they can choose not to act when something feels wrong. That is agency in its fullest sense.

Misalignment removes that choice. And it does so in one of two ways.

The first is paralysis. The misaligned leader has capacity - in theory - but cannot access it. The constant internal negotiation between who they are and how they are leading consumes the very energy required to act. They appear busy. They attend the meetings, send the emails, make the decisions. But they are not truly choosing. They are managing. And underneath the motion, something is stalled.

The second is compulsion. This is the more dangerous condition - because it looks like performance. The compelled leader acts, but not freely. Driven by external pressure, unexamined beliefs, or inherited expectations, their decisions are reactions dressed as choices. The motion is real. The agency is not. And the outcomes they produce are often ones nobody truly intended.

Every leader I have worked with has lived in both of these places at different points. I have lived in both myself.

There were seasons where I was paralysed - where the gap between who I was and how I was leading consumed more energy than I had left to act from. And there were seasons where I was compelled - moving fast, making decisions, appearing decisive, while quietly producing outcomes that didn't reflect what I actually valued. What I have learned, from the inside, is that neither of these feels like misalignment from within. Paralysis feels like caution. Compulsion feels like commitment. It takes alignment to tell the difference.

What stopped me cold about the AI conversation is this: in humans, misalignment most often produces paralysis. The incoherence is felt. It slows things down. There is at least a built-in brake.

In agentic AI, misalignment produces compulsion - at scale, autonomously, without hesitation. A misaligned AI system has no mechanism to sense that something is wrong and stop. It simply executes - continuously, toward objectives that may not reflect what the organisation actually values. What looks like agency is sophisticated automation pointed in the wrong direction.

The parallel is exact. And the stakes are proportionally higher.

Alignment is not a constraint on agency - for leaders or for the systems they build. It is the precondition for it. Without alignment, what exists is not genuine choice. It is either stagnation or momentum in the wrong direction.

Every autumn, millions of birds migrate thousands of kilometres to destinations they have never seen. They carry no map. They follow no instructions. They move because their inner nature and outer conditions are in alignment - and from that alignment, genuine direction emerges. Disrupt that compass, and the bird doesn't simply fly the wrong way. It loses the capacity to choose at all.

Agency works the same way. Alignment is the inner compass. Without it, capacity and motion exist - but genuine direction does not.

As you deploy AI in your organisation, how confident are you that the systems you are building are aligned to what you actually value - and not simply executing, at scale, toward objectives nobody truly intended?