10 Prompt Engineering Techniques That Actually Work
Skip the theory. Here are 10 proven prompt engineering techniques with real examples you can copy and use today.
10 Prompt Engineering Techniques That Actually Work
I've tested hundreds of prompting techniques. Most are overhyped. These 10 actually make a measurable difference in output quality.
1. The Role Technique
Assign a specific expert role with years of experience.
Before: "How do I improve my website?" After: "Act as a UX designer with 10 years of experience in e-commerce. Review my website [URL] and give me 5 specific improvements to increase conversion rate."
2. Chain of Thought
Ask the AI to think step by step.
Before: "Should I raise prices?" After: "Think step by step: Given that my costs increased 15% and competitor prices rose 10%, should I raise my prices? Consider: customer price sensitivity, margin impact, competitive positioning, and timing."
3. Few-Shot Examples
Show the AI what you want by example.
Format: "Here are examples of what I want: Input: X → Output: Y Input: A → Output: B Now do: Input: C → Output: ?"
4. The Constraint Technique
Limitations often improve output.
"Explain quantum computing in exactly 3 sentences, using no jargon, for a 12-year-old audience."
5. Output Format Specification
Tell the AI exactly how to structure the response.
"Respond as a markdown table with columns: Feature, Pros, Cons, Rating (1-5)"
6. The Iteration Technique
Don't accept the first output. Refine.
Prompt 1: Generate initial draft Prompt 2: "Make it more concise and add specific data points" Prompt 3: "Adjust the tone to be more conversational"
7. The Negative Prompt
Tell the AI what NOT to do.
"Write a product description. Do NOT use buzzwords like 'revolutionary', 'game-changing', or 'best-in-class'. Do NOT use exclamation points. Do NOT start with a question."
8. The Evaluation Prompt
Ask the AI to critique its own work.
"Now review your response. Rate it 1-10 on: accuracy, completeness, clarity, and actionability. What would you improve?"
9. XML Tags (for Claude)
Use XML to organize complex prompts.
<context>Background info here</context>
<task>What to do</task>
<format>How to structure output</format>
<constraints>Rules to follow</constraints>
10. The Mega-Prompt
One comprehensive prompt that produces a complete deliverable.
Combine: Role + Context + Task + Format + Examples + Constraints into a single, detailed prompt. More upfront effort, but the output is dramatically better.
The Meta-Lesson
The best prompt engineers aren't people who memorize techniques. They're people who think clearly about what they want and communicate it precisely. Every technique above is really just a way of being more specific about your expectations.
Start with technique #1 (Role) and #5 (Format). These two alone will improve 80% of your prompts.
Tags
Pick by task, not just by model
See which AI model wins for your specific job — resume writing, coding, logos, video ads, and 28 more.
Browse all tasks