Working With AI Without Wasting Time — Note #3
Most people sit down with an AI tool and type some version of: “Give me the answer.”
That is fine for trivia. It is terrible for decisions.
A more useful approach is: “Show me what I might be missing.”
When you treat AI as an answer machine
- Overconfident text in a single voice
- Hidden assumptions presented as facts
- Outputs you are tempted to copy without questioning
When you treat AI as a thinking partner
- Alternative perspectives
- Edge cases and failure modes
- Challenges to your default assumptions
The difference is in the prompts.
Instead of asking AI to write, ask it to challenge
Instead of: “Write an AI use policy for my company.”
Try questions like:
- “What are the risks I am most likely overlooking if I roll out AI tools to my team?”
- “Challenge my assumptions about AI use in a 20-person company that handles sensitive client data.”
- “List the ways this draft policy could fail in real life, and why.”
Now the model is not pretending to be your lawyer or your compliance department. It is acting more like a sparring partner: poking holes, surfacing blind spots, and giving you raw material to think with.
You still own the judgment. You still own the decision. But instead of starting from a blank page, you are starting from a set of perspectives you can test against reality.
That is where AI is genuinely useful: not as an oracle, but as a mirror and a challenger for your own thinking.