Think of a conversation you had with a friend or colleague, and they didn’t quite understand what you were trying to say. Maybe you had to rephrase a question or explain something differently — how you ask or explain something matters a lot! Working with prompts in AI is very similar.
Why Does the Prompt Approach Matter?
For instance, you ask a classmate or colleague for help on a particular topic. You will get better results if you were to say, “Could you help me understand how this [topic] works,” rather than just saying, “What is this [topic]?” The same principle applies to AI — the more precise and structured your request, the better the response.
Prompt Patterns
Think of “prompt patterns” as recipes for asking questions. Just as a recipe helps you make a dish consistently good (hopefully!) each time, prompt patterns help you get reliable, high-quality responses from AI. Generally speaking, effective prompts tend to have the following characteristics:
- They set the context – instead of jumping straight to the question, set the scene:
- “I am a first-year computer science student trying to understand …”
- “I need to explain this concept to my 10-year-old brother/sister/cousin …”
- They are specific and detailed–rather than asking, “Tell me about the First World War,” try:
- “Can you explain the three main causes of World War I?”
- “What were the key turning points of World War I? Focus on the alliances and imperialism.”
- They provide examples – using examples in a prompt is like giving the AI a roadmap, like so:
- “Use the following structure …” (in the case of a document or an email)
- “The tone should be professional and technical …” (in the case of summarization)
- They specify the format desired – if you need the information structured in a particular way, say so:
- “Can you explain this in a series of bullet points?”
- “Can you break this down into three main sections?”
- “Please provide an example after each main point.”
- They use follow-up questions – chatting with AI is a dialogue, so don’t hesitate to ask for more details or clarification:
- "Can you explain that last point in simpler terms?”
- “Can you give me a real-world example of that concept?”
Prompt Before and After
Basic Prompt | Effective Prompt |
|---|---|
“Tell me about climate change.” | “I’m preparing a 5-minute presentation for my environmental science class on climate change. Could you help me outline the three most important points that should be covered, with concrete examples for each point? Please provide a brief, attention-grabbing introduction for this presentation to my classmates.” |
Can you see the difference? The “Effective Prompt” provides the AI with a clear context (a class presentation), a specific format (three points with examples), and a defined audience (classmates).
Examples: Purpose-built Prompts
Task | Prompt |
|---|---|
Summarize an article, document, or PDF. | “Summarize the main points of the [attached document], focus on key changes and implications …” “... for our marketing strategy” or |
Revise text for clarity and enhanced readability. | “Revise the provided technical document for clarity and enhanced readability. Focus on simplifying complex technical language without compromising the accuracy of the details.” |
Create a presentation outline. | “Create a presentation outline for an upcoming [our department] strategy workshop. The outline should include sections for an executive summary, market analysis, competitive landscape, strategic objectives, implementation plan, and expected outcomes. Ensure the structure flows logically and is designed to engage senior management.” |
Create a meeting recap (minutes).
| “Create a recap of our project team meeting on [Date]. The recap should cover updates on project progress, key discussion topics, challenges encountered, and decisions made. Include a bullet-point list of action items with the names of the team members responsible and their deadlines. Aim for a clear, structured format that helps keep the team aligned.” |