Prompt Pattern Reference
Instruction-Based Prompting
Give the AI clear, structured instructions about what to do, how to format the output, or what constraints to follow. This produces focused, predictable responses.
Write a 150-word product description for a noise-cancelling headphone. Use short sentences. Include price ($249), battery life (30 hrs), and one customer quote.
Summarise photosynthesis in exactly 3 bullet points suitable for a 10-year-old. Use simple language and no jargon.
Role-Based Prompting
Assign the AI a specific professional identity — such as a financial advisor, travel guide, or teacher — then frame your question within that expert's perspective. This focuses the response through a particular lens of expertise.
You are a certified financial planner. A 28-year-old client earns $80,000/year and has no savings. What are their first three steps to build wealth?
Act as a local travel guide specialising in Kuching, Malaysia. A family of four has 3 days. Create a cultural and food-focused itinerary.
Chain-of-Thought
Ask the AI to think step-by-step, show its reasoning, or break a complex problem into parts before giving a final answer. This produces more thorough and logical responses.
Think through this step by step: A company's revenue grew 15% but profits fell 5%. What are the most likely explanations? Show your reasoning before concluding.
Walk me step by step through how you would explain compound interest to a 12-year-old, then show an example calculation.
Few-Shot Prompting
Include one or more example input-output pairs in your prompt to show the AI the exact format and style you want. The AI will follow the pattern you demonstrate.
Classify the feedback as Positive, Negative, or Neutral. Example: Clean layout → Positive. Too bright → Negative. Now classify: The onboarding took me 5 minutes.