Effective AI Prompting Basics
The quality of AI output depends heavily on how you ask. Better prompts = better results. This guide covers fundamental prompting techniques.
Why Prompting Matters
AI responds to what you ask. Vague questions get generic answers. Specific, well-structured prompts get useful responses.
Weak prompt: “Write about marketing” Strong prompt: “Write a 500-word blog post explaining email marketing basics for small business owners who have never used email marketing. Focus on getting started with limited budget.”
Same AI, vastly different results.
The Five Elements of Good Prompts
1. Task
Clearly state what you want:
- “Write a…”
- “Explain…”
- “Summarize…”
- “Create a list of…”
Be specific: Not “help me with email” but “write a follow-up email to a client who hasn’t responded in two weeks.”
2. Context
Provide relevant background:
- Who is the audience?
- What’s the purpose?
- What’s the situation?
Example: “I’m a freelance designer reaching out to a client who missed our last two meetings.”
3. Format
Specify how you want the output:
- Length (500 words, 3 paragraphs)
- Structure (list, essay, steps)
- Style (formal, casual, technical)
Example: “Format as a numbered list with brief explanations for each point.”
4. Constraints
Set boundaries:
- What to avoid
- What to include
- Limitations
Example: “Don’t use jargon. Include at least one example per point. Keep each section under 100 words.”
5. Examples
Show what you want:
- Sample outputs
- Style examples
- Reference material
Example: “Here’s an example of the tone I want: [paste example]“
Prompting Patterns
The Role Pattern
Ask AI to adopt a perspective:
“You are an experienced product manager reviewing a feature specification. Provide feedback on this spec, focusing on potential edge cases and user experience issues.”
The Step-by-Step Pattern
Break complex tasks into steps:
“Help me plan a product launch:
- First, identify key audience segments
- Then, suggest messaging for each segment
- Finally, recommend marketing channels”
The Refinement Pattern
Iterate on responses:
- First prompt: Get initial output
- “Make it more concise”
- “Add specific examples”
- “Adjust tone to be more casual”
The Template Pattern
Create reusable structures:
“Write a product description using this format:
- Headline (benefit-focused, under 10 words)
- Problem statement (1-2 sentences)
- Solution (how product solves it)
- Three key benefits (bullet points)
- Call to action”
Common Prompting Mistakes
Too vague: “Help me with my business” → AI doesn’t know what you need
Too broad: “Explain everything about marketing” → Output is shallow overview
No context: “Is this good?” → AI doesn’t know what “this” is
No format: “Write something about X” → Output may be wrong length/style
Asking once: Accepting first response → Iterating usually improves output
Examples: Before and After
Example 1: Email Writing
Before: “Write an email to my boss”
After: “Write a brief email (under 150 words) to my manager requesting a meeting to discuss my career development. Tone should be professional but not formal. I want to discuss taking on more responsibilities in the upcoming project. Include a specific ask for a 30-minute meeting this week.”
Example 2: Code Help
Before: “Fix my code”
After: “This Python function is supposed to sort a list of dictionaries by the ‘date’ key, but it’s returning an error. Here’s the code: [paste code]. The error message is: [paste error]. What’s wrong and how do I fix it?”
Example 3: Research
Before: “Tell me about market research”
After: “Summarize 5 methods for conducting market research as a startup with limited budget (under $1000). For each method, include: what it is, approximate cost, time required, and what insights it provides. Format as a comparison table.”
Tips for Different AI Tools
ChatGPT
- Use plugins for web access and specific tasks
- Create custom GPTs for repeated workflows
- System prompts in Plus allow persistent instructions
Claude
- Excels with longer context — include more background
- Good at nuanced tasks — be specific about subtleties
- Can handle documents — paste relevant materials
Perplexity
- Phrase as research questions
- Specify what kind of sources you want
- Follow up for deeper exploration
Frequently Asked Questions
How long should prompts be?
Long enough to be specific, short enough to be clear. A good prompt might be 2-5 sentences. Complex tasks need more context.
Should I be polite to AI?
It doesn’t matter for results, but “please” and “thank you” don’t hurt. Whatever feels natural to you.
What if the first response is wrong?
Iterate. Tell AI what’s wrong, what you want different. “This is too formal” or “I need more specific examples.” Refinement is normal.
Can I save good prompts?
Yes. Keep a document of prompts that work well for you. Reuse and refine them.