The Team That Got Replaced With Prompts (And What Happened Next)

Two conversations this month have been rattling around in my head, and together they tell you almost everything you need to know about where senior tech hiring is going.

The first was with a cybersecurity leader based in the US, someone who had been heading up security for one of the big tech companies. His entire team was made redundant. The plan, as it was explained to him, was to run the function on prompts. Not a smaller team supported by AI. Prompts instead of people.

Here is the part that should give every leader pause. He told me the approach they trialled before the redundancies did not actually work. They knew. They did it anyway.

We have all seen the public versions of this story. A well-known fintech famously cut its customer service function early in favour of AI, watched satisfaction scores fall off a cliff, and has been quietly rehiring humans ever since. The pattern is becoming familiar: treat AI as a headcount delete button, feel clever for two quarters, then spend the next two years rebuilding capability you already had.

The second conversation was a coffee with a senior executive who has just moved back to Australia after years running large operations in the US. On paper, AI could threaten plenty of what he does. His read on the market was the opposite, and I think he is right.

His observation was that organisations are now hunting harder than ever for genuine operating expertise, because experienced leaders are the people who can inform the models and inform the organisation at the same time. They know what good looks like, so they can tell when the AI output is wrong, and they know the business well enough to decide where AI should and should not be let loose.

And here is the detail I loved. This is a person with decades of experience and every excuse to coast on it. Instead, he had gone and completed an executive program specifically on deploying AI in organisations. Not to become a technologist. To stay credible in the conversations that matter.

Put those two stories side by side and the lesson writes itself.

AI capability at senior level is now non-negotiable. But the capability that matters is not prompt wizardry, and it is certainly not breathless enthusiasm. It is judgment. The ability to look at what AI produces and know whether it is right. The ability to look at a function and know which parts genuinely can be automated and which parts only look automatable from the org chart.

The big tech company in the first story did not fail because AI is useless at security work. It failed because the decision was made by people optimising a cost line rather than people who understood the work. The expertise they deleted was the exact expertise they needed to make the automation succeed.

So what does this mean practically if you are a tech leader or a hiring manager?

First, when you hire senior people, test for AI judgment, not AI cheerleading. Ask candidates where they have deployed AI and what went wrong. The strong ones will have scars and specifics. The weak ones will have slideware. In my screening calls, the difference is obvious inside five minutes.

Second, look at your own bench. The leaders on your team who are quietly upskilling, the way my coffee companion did, are your most valuable people right now. They are also the most poachable, because every board in the country is asking for exactly that profile. If you have one, invest in them before someone like me calls them.

Third, if there is a proposal on your desk to replace a function with AI, ask one question: have the people who actually understand the work signed off that it works? Not the vendor. Not finance. The operators. If the trial did not work and the plan is proceeding anyway, you are not looking at a technology strategy. You are looking at a cost story that will come due with interest.

The leaders who win the next few years will not be the ones who used AI to delete experience. They will be the ones who used experienced people to deploy AI well.

That is a hiring question before it is a technology question. It always was.

See you next week.

James