The CEO does not use it. So the signal is clear: neither does anyone else.
Except everyone is using it. They just have ways of looking like they are not.
I have seen this before. Large organisations with significant red tape. The workers are not busy. The project has not arrived yet. The approval is pending. So they fill the time with something useful. They build the skill. They run the experiment. They find out what the tool can actually do. And when the project arrives, they are ready, and nobody asks too many questions about how.
Shadow AI is not a technology problem. It is a culture problem. The signal from the top does not stop the usage. It just drives it underground. The log shows nothing. The output is better than it should be. The timeline is shorter than anyone expected. And the organisation celebrates the result without understanding what produced it.
The CEO who does not use it is not just missing a tool. They are creating a two-tier organisation. The people who are building the capability quietly, and the people who are not. The first group will continue to outperform. The second group will continue to wonder why.
The more interesting question is not whether your people are using AI. They are. The question is whether you have created the conditions where they can use it openly, improve together, and build something the organisation actually owns.
Do you have a culture where that is possible? Or are you managing the signal while the shadow grows?