10 Advanced AI Prompting Techniques

1. Role priming: Tell the model who it is before you tell it what to do. A line like "you write landing pages that have sold for years" sets the quality bar it aims for. 2. Tell it what not to do: Constraints sharpen output faster than instructions. "Don't use bullet points, don't hedge, don't add a disclaimer" strips out the filler most models pad answers with. 3. Show it one example: One example teaches more than a paragraph of description. Paste a sample of exactly what good looks like and the model matches the pattern instead of guessing at it. 4. Lock the format first: Describe the shape you want back before it starts writing. Spell out the length, the sections, and the tone, and you stop getting walls of text you have to re-cut. 5. Make it think before it answers: Ask the model to reason first. A line like "work through your logic step by step before the final answer" catches mistakes it would otherwise rush past. 6. Have it critique its own draft: The second pass is where quality shows up. Tell it to find three weaknesses in what it just wrote and fix them, and the rewrite usually beats the first try. 7. Give it the constraints up front: Vague inputs create vague answers. Hand it the audience, the goal, and the word count in the first message so it isn't filling gaps with its own assumptions. 8. Ask for options, not one answer: A single answer gives you nothing to compare against. Request five versions that each take a different angle and you pick the strongest instead of settling for the first. 9. Feed it your own voice: Generic tone comes from generic input. Paste a few things you've written and tell it to match your rhythm and word choice, and the output stops sounding like a press release. 10. Let it ask you questions first: The model often needs context you forgot to give. Tell it to ask you three questions before it starts, and the answer it produces fits what you actually wanted.