Foundations of Prompt Engineering
Common Prompt Mistakes That Cost Time
Most people waste hours going back and forth with AI because of avoidable mistakes. Learn to spot these 5 errors and save yourself significant time.
Mistake #1: Being Too Vague
The Problem:
"Help me with my presentation"
AI doesn't know: For whom? About what? What format? What help do you need?
The Fix:
"Create an outline for a 15-minute sales presentation to C-level executives about our new CRM integration. Include 5 key slides with talking points."
Mistake #2: Overloading with Instructions
The Problem:
"Write a blog post about AI, make it engaging, SEO-optimized, include statistics, add a call-to-action, mention our competitors, keep it professional but casual, around 1500 words, include headers, make sure it's mobile-friendly..."
Too many competing demands confuse the AI.
The Fix: Break it into steps:
- First: Create an outline
- Then: Write section by section
- Finally: Add SEO elements and polish
Mistake #3: Missing Critical Context
The Problem:
"Write a response to this customer complaint"
AI is guessing your company, your policies, and your voice.
The Fix:
"I'm a customer service rep at [Company], a B2B software provider. Our policy allows full refunds within 30 days. Write a response to this customer complaint that acknowledges their frustration, explains what happened, and offers a solution. Maintain a professional but empathetic tone."
Mistake #4: Ignoring Output Format
The Problem: You ask for a comparison and get paragraphs of text when you needed a table.
The Fix: Always specify format:
| What You Need | How to Ask |
|---|---|
| Easy scanning | "Present as bullet points" |
| Comparison | "Create a table with columns for..." |
| Step-by-step | "Provide numbered instructions" |
| Quick summary | "Summarize in 3 sentences" |
Mistake #5: Not Iterating
The Problem: Accepting the first output even when it's not quite right.
The Fix: AI conversations are iterative. Use follow-ups:
- "Make it more concise"
- "Add more detail to the second point"
- "Rewrite this for a technical audience"
- "Give me 3 alternative versions"
Pro Tip: When you get a great output, save the prompt. Build your own library of what works.
Quick Diagnostic
When AI output isn't right, ask yourself:
| Symptom | Likely Cause | Fix |
|---|---|---|
| Too generic | Missing context | Add who, what, why |
| Wrong format | No format specified | State exactly how you want it |
| Missed the point | Vague task | Be more specific about what you need |
| Too long/short | No length constraint | Specify word count or sections |
Key Takeaway
Most prompt problems come from one of these 5 mistakes. Before hitting enter, quickly check: Is it specific? Is there context? Is the format clear? Did I avoid overloading? And remember — you can always iterate.
Congratulations! You've completed Module 1. Take the quiz to test your understanding of prompt engineering foundations. :::