Ask what it's missing

The move When AI's first answer lands, paste back a prompt that makes it list what it's missing and mark the fact that would move the answer most. Read the gaps as a map of your own blind spot; which one matters stays your call.

The short version: AI sounds just as sure on half your brief as on the whole thing, and never flags which. So a confident answer to a thin brief is a map of the holes in your own thinking. Ask what it's missing and those holes come back as a list of what you never said. It can point at the gaps; deciding which one matters, and going to fill it, stays yours.

AI is never unsure. Whether you feed it the full picture or half of it, the answer comes back equally certain, equally polished, equally ready to send. Everyone worries the thing will make something up. The quieter danger is the confident, fluent answer that's just missing what mattered.

A polished answer, missing what mattered

Picture a reply it drafts to a client: clean, reasonable, done. What it doesn't know is that the client threatened to walk last week, because that lived in your head, not the prompt. The draft reads the same either way. So its confidence tells you nothing about whether it actually has what it needs.

The machine has it worse

Daniel Kahneman named the human version: WYSIATI, what you see is all there is. We judge on what's in front of us and forget to weigh what isn't. A chatbot has it worse. The only thing in front of it is the box you typed into.

Ask what it's missing

So a confident answer to a thin brief isn't the model being smart. It's a map of the holes in your own thinking. When that first answer lands, ask what's missing before you use it. Paste this back as your next message and those holes come back as a list of what you never thought to say:

You answered using only what I gave you.
What are you missing that could change this?
List the three facts or constraints you'd most want, and mark the one that would move your answer the most.

It still can't know the client threatened to walk. That never reached the box. What it can do is hand you the questions that would've surfaced it: the history here, the stakes, the one number that changes the call.

Which gap matters is your call

The model can point at the gaps. It can't tell you which one matters and it can't go fill it. That call is the job and it stays yours.

Confidence isn't competence. The gap between them is where you're still needed.

Reply and tell me the one AI-at-work move you wish you'd mastered. I read every one and they shape what lands here next.