Let the AI advise you before it makes you decide

One of the highest-leverage habits when working with AI costs you a single sentence — and it works even when you have no clue about the topic.

Whenever the AI puts a decision in front of me (“Do you want A or B?”), I don’t answer right away. I hand the question back:

“What would your recommendation be — and why? For both options, explain the implications: on effort, on correctness, and what could come of each in the best and the worst case.”

This flips the roles. Instead of blindly picking between two options you can’t fully judge, the AI now has to justify its recommendation and lay out the consequences. And that justification contains exactly the information you were missing.

Why it works so well

• The recommendation plus the “why” shows you the default direction. In the vast majority of cases the reasoned recommendation is already the right one — you just need to be able to sanity-check it.
• Effort vs. correctness makes the trade-off visible. Often one option is faster, the other cleaner. Only when both sit side by side can you choose deliberately instead of leaving it to chance.
• Best case and worst case surface the risk. This is where you find out whether an option hides a trap you’d never have spotted as a non-expert.

The real trick: you don’t need domain knowledge to make a good decision — you only need the ability to tell a strong justification from a weak one. Almost anyone can do that.

One word of caution

The AI likes to tell you what you want to hear. If both options sound suspiciously equal, ask pointedly: “Where exactly would option B fall apart?” A real trade-off always has a downside — if none is named, the answer isn’t honest enough yet.