At twenty-nine I delivered 241 per cent sales growth at Electrolux. I was dismissed within the year.
I want to walk through that year as one decision, step by step. Then I want to say what I would decide differently if I had the same year in front of me again. I'm writing it now because the mistake I made then is easier to make today, and AI is the reason.
The situation
At twenty-nine, I delivered that growth. Looking back, I would ask what the result cost the team.
What matters is the shape of it. There was a team doing the work. There was a number that would show whether the year had gone well. A number alone cannot show what the year cost the people doing the work.
What the number said
The number was 241 per cent sales growth.
A number like that is easy to accept. It looked like success, and I read it as success.
Today an AI answer does the same job. It arrives finished and confident, and it sounds right. When something sounds that right, it is easy not to ask how it was made. I didn't.
The check I did not make
The check I needed was a simple one. What was this result costing the team that produced it?
I didn't make that check. I looked at the number, and I didn't look closely enough at the people behind it.
The point I can make from my own account is that the result was not enough.
A sales figure tells you what was sold. It doesn't tell you what it took to sell it, or whether the same people can do it again next year.
The decision, and the one I would make now
The decision I made was to treat the number as proof that everything else was fine. If the growth was there, I took it that the way we were working must be right. I moved fast, and I decided on the outcome.
I'd like to say I have left that habit behind. I haven't. I still move too fast, and I still default to outcomes. The work, for me, is noticing earlier and listening longer.
So here is what I would do with the same year.
I would ask the people doing the work what the pace was costing them. I would ask while the number was still rising, and not wait until it had arrived.
If they told me the pace was breaking them, I would change the pace before I reported the year as a success.
I would ask them one at a time, and not only in a meeting. That would give each person a chance to speak without colleagues present.
I would treat a quiet team as a reason to ask again. When a team under pressure goes quiet, that tells you something too.
And before I presented the number as a success, I would put what it had cost beside it. The people reading the report would then see both.
These changes would take time, and I would need to make room for them.
What happened after
I was dismissed within the year. I take responsibility for how I worked.
The lesson I took from it is one line: results are not the same as progress. I had a result. I had mistaken it for progress.
Why this matters more now
In my book The Human Layer, not yet published, I wrote: "You can be right about the technology and wrong about the humans. And the humans will always win."
The lesson in this account starts with a sales number, not an AI answer. I was right about the number and wrong about the humans, and the humans won.
Leaders still need to address weaknesses in how the organisation works. AI speeds up what is already there.
With AI, good numbers arrive faster than they used to. A forecast or a cost saving can look finished before anyone has asked what it took to produce. If the way a team works is already under strain, AI can speed that up too, and the number can still look good.
So when an AI figure looks good, the question I would put to any leader, and to myself first, has two parts. What did it cost the people whose work it describes, or who must act on it? And who did you ask?
Those answers help you decide whether the number represents progress. You still need to check the number itself.
I give keynotes, briefings and workshops on this.