This prompt fixed 80% of my ChatGPT problems.
I was getting so frustrated with AI-generated word salad. I spent an hour trying to get it to write a simple email to my kid’s school, and it came back with something that sounded like a TED Talk written by a toaster. Sound familiar?
Then I tried something totally different: I told ChatGPT to stop writing and build a one-page brief first. It was a complete game-changer. Instead of guessing, it asked a few smart questions, filled out a clear plan, and then nailed the email on the very first try.
I turned this trick into a supercharged meta-prompt called BriefBox. It’s a wrapper that forces any AI to get the key details down before it starts writing. No more endless revisions!
⚙️ Here’s the breakdown:
📌 It makes the AI act like a research agent, creating a structured BRIEF before drafting.
🚀 It has two modes: a fast “Lightning” mode and a “Deep Dive” mode that asks up to 3 targeted questions for complex tasks.
📜 The brief captures everything that matters: your goal, audience, key message, tone, and format.
✅ The final draft sticks strictly to the brief, killing all the generic filler and “as an AI” nonsense.
This simple shift from drafting to briefing first makes your outputs so much sharper and more effective. It’s an awesome way to engineer better context every single time.
Want to try it yourself? The full prompt with all the rules is in the Reddit post!
I turned ChatGPT into a one-page intake form. Quality went up, word salad went down.
byu/ThePromptIndex in