The job has already changed

"Which model should we be using?"

I get asked that more than anything else. I coach executive general managers at one of Australia's largest banks, and I build production AI systems myself, so it is a fair thing to put to me. For a long time I answered it straight. Which model, how to prompt it, whether to trust what comes back.

I was answering the wrong question, and it took me too long to notice. That question is comfortable, which is why it keeps getting asked. It leaves your job exactly where it was. You are still the expert, still the decision-maker, still the person whose judgement is the scarce thing in the room. The tools only make you faster.

That was true for a while. It is not the phase we are in now.

What has actually changed?

For most of the last century a senior leader's value came from one of two things. Either you knew something the people around you did not, or you could orchestrate people who knew things you did not. Expertise, or coordination. AI is compressing both, faster than most organisations have noticed.

Take the first. When a model can produce a competent first draft of almost any piece of knowledge work in seconds, the scarce thing becomes the judgement about which work is worth doing, what good looks like, and what the organisation is trying to accomplish in the first place. That is a leadership skill, and it sits in your Inner Game. Most leadership development programmes are not training it.

The second shift is harder to see. When a team is part human and part AI, and one senior analyst directs several agents through work that used to take four people, the coordination job changes underneath you. You are no longer managing the person doing the work. You are managing the person who is managing the system. That sounds like a small variation on the same job. It is not. The things that go wrong are different, what you have to keep an eye on is different, and the human skills you draw on are different.

Where does this go wrong?

The gap I run into most is not technical ignorance. Nobody needs their EGM to be an engineer. It is a category error, and I have made it myself: treating AI adoption as a capability problem when it is a judgement and structure problem. We send our people to prompt engineering workshops. We celebrate the early pilots. We do not sit down and ask what the org chart looks like in two years, once the ratio of people to agents has inverted.

The executives I watch handling this well have one thing in common. They have updated their picture of what their own job is. They are not trying to become engineers. They are asking harder questions. Which decisions should never be handed to a system? Where does human accountability sit when AI is in the chain? What does good look like when I can no longer judge the work by doing it myself?

Those questions are not new. Leaders have always had to govern work they could not personally perform. And yet the speed and the scope of this shift mean the old ways of adapting, hire smarter people, stay close to the detail, build institutional knowledge slowly, are no longer enough on their own. The timeline has compressed. The categories have moved.

Treat this as a tools question and you will get incrementally faster at the wrong things. Treat it as a question about the shape of your role and you will be one of the people making the decisions that matter in three years.

This is harder on people than we admit

Someone who has spent twenty years building deep expertise in a domain, and who then watches a model produce in an afternoon what would have taken their team a month, is going through something real. It is disorienting. Some of what they built is worth less than it was. The question of where their value sits now is not rhetorical. I have sat with people while they asked it out loud, and I have asked it of myself.

Some roles will shrink. Some skills will lose their value. The leaders who come through this well will be the ones who can hold that steadily and still make clear decisions for the people who depend on them. Staying steady while the ground moves is inner work, and no model will do it for you.

If you are a senior leader working out what this means for you or your organisation, coaching is where to start, or talk to us about the wider practice.

Attribution

Written by Andrew Ramsden. AI tools assisted research and drafting; all outputs verified.

Accountable

Andrew Ramsden.

Limitations

No external sources.

References

No external sources. Claim labels are in the SOURCED sidecar.