Core question
Why does our team not get better results from AI when everyone is already using it?
Short answer
Because almost nobody has a prompting framework. Ask a room of professionals what their repeatable process is for getting good output and the room usually goes quiet. Without a reference point for what good output looks like, a fast plausible answer just feels like magic.
I have run this test in a lot of rooms now, across more than 13,000 executives trained over the years, and the result is remarkably consistent. Everyone has used a language model. Almost no one can describe how they get a good result on purpose.
The root cause is not laziness. It is that people are using a tool without knowing what it does. In a workshop with 45 executives, I asked whether anyone knew how a large language model works, and one replied by asking what an LLM was. He had used ChatGPT. He just did not know that was what it was.
Every participant was a user with no understanding of the tool in their hands. They did not realize the model produces answers by predicting each word based on probability and pattern drawn from its training data. That is math, not comprehension. Without that insight, nobody can tailor a request to get better results, because they do not know what the machine is doing with what they typed.
Once you grasp that these are probability engines, you can improve your odds dramatically by giving a clear task, a role, and context. Details matter.
The honest executive question here is not whether your team is using AI. It is whether your leadership team is truly AI-literate or just AI-curious.