<aside>

💡 AI outputs are rarely perfect on the first try — and that’s normal. This page shows you how to improve results quickly using small, targeted prompt edits (instead of rewriting everything). Think of it like debugging: you diagnose what’s wrong, apply a fix, and rerun.

</aside>

1) The Iteration Loop (simple & effective)

Use this loop anytime an output isn’t good:

1. Diagnose what’s wrong (too vague? wrong tone? missing facts?)
2. Pick one fix (don’t change 10 things at once)
3. Re-run with a short “repair prompt”
4. Repeat until the output is usable

💡 Rule: One problem → one repair instruction.

2) Diagnose the problem (fast checklist)

• [ ] Most issues fit here:
• [ ] Too generic / fluffy
• [ ] Too long / too short
• [ ] Wrong tone (too casual, too marketing, too harsh)
• [ ] Poor structure (hard to scan)
• [ ] Missing key information
• [ ] Not actionable (no next steps / no owners)
• [ ] Hallucinations (invented facts)
• [ ] Not adapted to audience
• [ ] Risky (sensitive/confidential)