I think it is because AI takes away the digging and the noise. Not because it thinks for me. Because there is less in the way of my own thinking.
Where it comes from.
I was asked to run a working session for a management team. About strategy, and what I thought made a good strategy.
Not long before, I had asked Claude to challenge my own approach to strategy. So when the request came, I was clear on what I thought was good and what was not. Within two or three minutes, I had sketched out what I wanted to say and what I wanted to do.
I think that would have taken me much longer before. I would have wanted to test it for longer, too. Now I could ask Claude to roast the plan against my own framework. Where did it fall apart? Where did it connect with other models or ideas I had? It worked really well.
What it actually does.
Research is the part I hand over. An agent reads what I would have had to read, and comes back with the shape of the problem instead of twenty open tabs.
The difference is not just speed. It is less noise. Less to sift through before I can see what matters. Sometimes that leaves me with a clarity that feels almost absurd.
Why the briefing sharpens you.
Briefing it properly helps, too. The right question, the right context, the right structure. Then I have to judge what comes back. Prompting well and judging well are both thinking. They sharpen my own view.
But the clarity is not only in the briefing. I think it also comes from having the digging and the noise taken out of the way.
Before you ask AI for an answer, ask it for the core of the problem in five lines. Write your own version next to it and compare.
Source: my own notes, October 2026.

