Think about the last time someone in your organisation had a genuinely good idea for doing something better. Not you. Someone three levels down, who does the work every day and can see exactly where it is broken.
What happened to the idea?
In most places I have worked, the honest answer is one of three things. It was never said out loud, because there was no obvious person to say it to. Or it was said, and it joined a queue for the attention of someone senior, where it waited behind forty other things until the person who had it stopped caring. Or, the best case, a supportive manager took it on, ran it as a side project for a year, and then left, and the project quietly died with their departure.
I have spent time inside large hierarchies and I now run small companies, and I do not think any of this is anyone’s fault. It is the shape of the thing. Authority sits at the top. Everything new has to climb up to be approved and funded. And the people at the top, however capable, can only think about so many decisions in a week.
The gate made sense when trying things was expensive
It is worth being fair to the old model. When testing an idea meant a business case, a budget line, a vendor, a project team and six months, of course you wanted a senior person deciding which ideas deserved it. The gate was there because each attempt was costly. Scarce attempts need rationing, and rationing needs a rationer.
What has changed is the price of an attempt.
Someone in accounts can now build a working version of the thing they have been wishing for, in an afternoon, with tools the organisation already pays for. Someone in operations can test whether a report could assemble itself before anyone has written a proposal about it. The attempt costs almost nothing. It needs no budget, no vendor, no team.
So the organisation is left with a gate designed to ration something that is no longer scarce. Ideas queue for permission to do what could have been tried and judged in less time than the queueing takes.
Two ways to use AI at the top
The first way is the obvious one, and it is fine as far as it goes. Give the executives better tools. Summaries of the papers, analysis on demand, a sharper brief before every decision. The people at the gate make better calls, faster. I help leadership teams do exactly this.
But notice what it preserves. The gate is still the gate. You have widened it a little. Every idea still has to climb, and the number of things a handful of people can properly consider in a week is still the ceiling on how fast the organisation can change.
The second way is harder and much more valuable. Move the trying down to where the work is, and change what the top is for.
What the inversion looks like
In the inverted version, the default flips. People at every level are expected to explore how AI could improve their own corner, without asking, inside some clear limits. Most of what they try will help only them, and that is fine. Some of it will turn out to matter to a whole team. A little of it will be worth changing how the organisation works.
The job of the structure is then to notice the promising ones and carry them upward, rather than to make every idea argue its way in from the bottom. Evidence travels up. Permission does not have to travel down first.
And the role of the top changes from deciding what gets tried to deciding what gets adopted. That is a smaller, better job. It is a sense check and an approval gate on things that have already shown they work, arriving with the evidence attached. A lot of the preparation for that decision can itself be automated: a standard one-page brief on what was tried, what it saved, what data it touches, what could go wrong and how easily it can be undone. The executive reads a page and says yes, no, or not yet.
I find it a rather appealing picture, and not only for the organisation. Take the grunt work away from people and give them room to improve their own work, and most of them turn out to be far more inventive than their job description ever asked them to be.
Why it usually does not happen
Because the difficult part is not process. It is letting go.
If you have spent a career becoming the person who decides, being asked to decide less feels like being asked to matter less. It is not. But it takes a secure leader to see that their value now lies in setting the limits, asking the good question at the sense check, and clearing the path for what works, rather than in being consulted on everything.
It also takes new rules, because “everyone experiment freely” with no boundaries is how you get customer data pasted somewhere it should not be, or six teams building the same thing, or two versions of the monthly number. The freedom has to come with a few firm lines: which tools, what data may go in, what no machine does without a person’s say-so, and whose number is the official one. Inverting the hierarchy is not the absence of governance. It is a different kind: fewer permissions, clearer limits.
And it takes patience with failure. Most experiments will not go anywhere. That is the point of making them cheap. An organisation that punishes the ones that fizzle will quickly have none.
One small step
This is a large shift and I would not try to make it in one move. Here is the smallest version I know.
Pick one team. Tell them, in your own words and in front of each other, that for the next ninety days they are allowed to try anything that improves their own work using the tools the company provides, within three or four stated limits, without asking. Ask only that they note down what they tried. At the end, sit down for an hour and look at the list together. Choose the one or two that deserve to become how the team works, and put your weight behind those.
Then notice what the hour felt like. You did not generate the ideas. You did not manage the experiments. You looked at evidence and made two decisions. That is the job, in the inverted version. It is a good job.
I have written about how a personal routine becomes a business process, which is the mechanics of the upward path, and about why people need to feel safe before they will show you how they work, which is the precondition for all of it.
If you lead an organisation and suspect you have become the gate, tell me about it. It is a more common and more fixable problem than it feels from the inside.