Level up your prompting with 19 prompts as a cheat code to unlock the hidden, genius level of AI.
Most people use ChatGPT for basic things like summarising an article or writing a quick text and they get a completely average, boring answer because that is the AI's default setting. These prompts completely break that default. They force the AI to slow down, stop guessing, and actually think deeply, act like a world-class expert, and double-check its own facts before it gives you an answer.
AI defaults to average because average is safe. These prompts remove safe as an option.
Use this when you know the AI is about to give you the same answer it gives everyone else.
Example: You're replying to a customer complaint and don't want to send the same "sorry for the inconvenience" every other company sends.
Pretend I paid you $50,000 to get this exactly right. Not a rough draft, not the safe, generic version. I want the answer only real expertise and real effort would produce.
Here's what I need: [describe your task]
Before you write anything, think about what actually separates a $50,000 answer from a free one. Then write that.
Use this before you write anything long. It forces the AI to define quality before it tries to hit it.
Example: You're writing your weekly update to your manager and want it to actually get read, not skimmed and forgotten.
Before you write [the article / email / plan], give me a rubric for what a 10 out of 10 version looks like. List the actual criteria, not vague ones.
Then write it so it scores a 10 against your own rubric. Check it against that list before you show me.
Use this when you need rigor more than confidence. Good for anything with numbers, claims, or arguments in it.
Example: You're comparing two broadband deals before you switch providers and don't want to act on numbers that turn out to be wrong.
Only give me an answer that someone like Einstein would consider genuinely rigorous, not just impressive-sounding.
If a point wouldn't survive real scrutiny, cut it. I'd rather have three claims that hold up than ten that don't.
Use this at the start of a long working session so the tone doesn't slide back to average halfway through.
Example: You're an hour into planning next week's meals in one long chat and can feel the suggestions getting repetitive.