Bad prompts don’t just give worse answers — they cost you extra back-and-forth, which is where most wasted usage actually comes from.
1. Asking for the whole thing at once
“Write me a full marketing plan” produces something generic you’ll rewrite anyway. Break it into stages: audience, positioning, channels, then copy — each one gets a sharper answer because the model isn’t guessing what you actually need.
2. Not saying what to leave out
If you don’t say “no intro, no summary at the end,” you’ll get both, every time. Negative constraints save you an editing pass.
3. Vague length requests
“Keep it short” means something different to every model on every day. “Under 100 words” or “3 bullet points” removes the guesswork entirely.
4. Re-explaining context every message
If you’re repeating the same background info in every follow-up, put it once at the top of the conversation as a block the model can refer back to, instead of restating it each time.
5. Not specifying the output format upfront
Asking for a table after already getting three paragraphs of prose means redoing the whole response. State the format in the first prompt, not as a correction.
The pattern
Every one of these mistakes causes an extra round trip. Fixing them isn’t about being clever — it’s about giving the model everything it needs the first time.
Leave a Reply