AI UX PlaygroundNewsletterJoin 2K+ AI designers and PMs on Substack. New teardowns, patterns, and prompts as they drop.

Product

Prompt Engineering Best Practices

Learn effective prompt engineering techniques for better AI interactions.

Prompt

Explain prompt engineering best practices for [AI model/tool name]. Cover:

1. Core Principles
   - Clarity and specificity
   - Context setting
   - Role definition
   - Output formatting

2. Prompt Structure
   - Effective prompt templates
   - Order of information
   - Use of examples (few-shot learning)
   - Chain-of-thought prompting

3. Advanced Techniques
   - Prompt chaining
   - Iterative refinement
   - Constraint setting
   - Output validation

4. Common Pitfalls
   - Vague instructions
   - Overly complex prompts
   - Missing context
   - Ambiguous requirements

5. Optimization Strategies
   - Token efficiency
   - Response quality improvement
   - Cost optimization
   - Speed optimization

6. Examples
   - Before/after prompt comparisons
   - Domain-specific examples
   - Use case demonstrations

Provide practical, actionable guidance with examples.

How to use

  1. 01

    Replace [AI model/tool name] with your specific tool (e.g., "ChatGPT", "Claude", "Google Gemini", "Perplexity")

  2. 02

    Optional: Add your use case context. Example: "I use [AI tool name] for [writing code/creating designs/analyzing data]. I want to improve my prompts for [specific task]."

  3. 03

    Paste the prompt into the AI tool you want to learn about (yes, use ChatGPT to learn ChatGPT prompting!)

  4. 04

    Read through the response and save it as a reference document - you'll refer back to it often

  5. 05

    Pick ONE technique to practice immediately. For example, if it mentions "role-setting", try: "You are an expert UX researcher. Analyze this user interview..."

  6. 06

    After trying, ask follow-up: "Evaluate my prompt: [your prompt]. How can I improve it using the techniques you taught me?"

Tips

  • This is a "meta-prompt" - it teaches you to write better prompts. Use it as your first step to mastering any AI tool
  • Create a "prompt library" document: Save this output + examples of your best prompts for future reference
  • Practice with real work: Don't just read - immediately apply one technique to a task you need to do today
  • Before/after comparison: Take one of your old prompts, apply these techniques, compare the results
  • Ask "Show me 3 examples of badly written prompts vs. well-written prompts for [your use case]" for concrete learning