Common Prompt Mistakes & How to Fix Them
I've reviewed hundreds of prompts from frustrated users. And honestly? The same mistakes show up again and again.
The good news: once you know what to look for, these are easy fixes. Let's go through the 7 most common prompt mistakes and exactly how to solve them.
Mistake #1: The "Mind Reader" Prompt
What it looks like:
"Write something good about my product"
"Make this better"
"Create a nice email"
Why it fails:
You're expecting the AI to know what "good," "better," or "nice" means to you. It doesn't. So it guesses. And its guess is usually the most generic, middle-of-the-road option.
YOUR BRAIN AI'S BRAIN ┌──────────────┐ ┌──────────────┐ │ "Good" means │ │ "Good" means │ │ • Concise │ ≠ │ • ??? Maybe │ │ • Funny │ │ anything? │ │ • Edgy │ │ • Generic │ └──────────────┘ └──────────────┘
The fix:
Replace vague adjectives with specific criteria.
❌ "Write a good product description"
✅ "Write a product description that is under 100 words, highlights the top 3 benefits, uses punchy short sentences, and ends with urgency (limited time offer)."
Mistake #2: The "Kitchen Sink" Prompt
What it looks like:
"Write a blog post about productivity, also include tips for time management, and maybe add something about work-life balance, oh and can you make it SEO-friendly and also include a call to action for my newsletter?"
Why it fails:
Too many competing objectives confuse the AI. It tries to do everything and ends up doing nothing well.
TOO MANY TASKS
│
┌─────────────────┼─────────────────┐
│ │ │
▼ ▼ ▼
┌─────────┐ ┌─────────┐ ┌─────────┐
│ Task A │ │ Task B │ │ Task C │
│ 50% │ │ 30% │ │ 20% │
│ effort │ │ effort │ │ effort │
└─────────┘ └─────────┘ └─────────┘
│ │ │
└─────────────────┼─────────────────┘
▼
┌──────────────┐
│ MEDIOCRE │
│ OUTPUT │
└──────────────┘The fix:
One prompt, one primary goal. Chain multiple prompts if needed.
❌ "Write and optimize and format and add CTAs"
✅ Prompt 1: "Write a 500-word blog post about the Pomodoro technique for remote workers."
✅ Prompt 2: "Now optimize this for SEO. Target keyword: 'productivity tips for remote workers'"
✅ Prompt 3: "Add a compelling CTA for newsletter signup at the end."
Mistake #3: No Constraints = No Focus
What it looks like:
"Explain quantum computing"
Why it fails:
Without constraints, the AI doesn't know if you want a children's book explanation or a PhD thesis.
The fix:
Add boundaries: length, audience, complexity level, what to avoid.
❌ "Explain quantum computing"
✅ "Explain quantum computing to a smart 12-year-old. Use exactly one analogy. Keep it under 150 words. Avoid any math."
Mistake #4: Ignoring the Power of Examples
What it looks like:
"Write tweets in my brand voice"
Why it fails:
The AI has no idea what your "brand voice" is. It's not psychic.
WITHOUT EXAMPLE WITH EXAMPLE
┌─────────────┐ ┌─────────────┐
│ "Brand │ │ "Brand │
│ voice" │ │ voice" │
│ │ │ │ │ │
│ ▼ │ │ ▼ │
│ ??? │ │ "Here's a │
│ │ │ sample: │
│ (guesses) │ │ [tweet]" │
└─────────────┘ └──────┬──────┘
│
▼
(Matches style)The fix:
Provide 1-3 examples of what you want. This is called "few-shot prompting" and it's incredibly powerful.
✅ "Write 5 tweets about our new product launch. Match this style:
Example 1: 'We didn't just update the app. We rebuilt it from scratch. Here's why that matters →'
Example 2: 'Hot take: Most productivity apps make you less productive. We fixed that.'"
Mistake #5: Forgetting Negative Constraints
What it looks like:
You ask for marketing copy and get something filled with "revolutionary," "game-changing," and "synergy."
Why it fails:
You told the AI what to do, but not what to avoid. Without anti-patterns, it defaults to clichés.
The fix:
Explicitly state what you DON'T want.
✅ "Write a product description. Do NOT use these words: revolutionary, game-changing, best-in-class, synergy, leverage, or any superlatives like 'best' or 'greatest.' Keep it factual and specific."
Mistake #6: Wrong Level of Specificity
What it looks like:
Too vague: "Help me with my resume"
Too specific: "Write exactly 47 words about my 3.5 years of experience at Company X where I managed 4.2 people on average"
Why it fails:
Too vague = generic output. Too specific = constrained, awkward output.
SPECIFICITY SCALE
Too Vague ◄──────────────────────► Too Specific
│ │
▼ ▼
Generic, Awkward,
Useless Forced
┌───────┐
│ SWEET │
│ SPOT │
└───────┘
│
Clear goals + Room to executeThe fix:
Be specific about WHAT you want, flexible about HOW the AI achieves it.
✅ "Rewrite my resume summary to highlight my transition from marketing to product management. Emphasize transferable skills. Keep it to 3-4 sentences. Make it compelling for a hiring manager at a tech startup."
Mistake #7: Not Iterating
What it looks like:
You get a mediocre response and give up (or start from scratch).
Why it fails:
You're leaving value on the table. The AI already understands your context—use it.
The fix:
Treat your first prompt as a draft. Refine from there.
Iteration prompts that work:
Quick Diagnostic Checklist
Before you hit enter, ask yourself:
☐ Have I defined what 'good' looks like?
☐ Am I asking for one thing at a time?
☐ Have I set constraints (length, audience, style)?
☐ Did I provide examples if needed?
☐ Did I say what to avoid?
☐ Is my specificity level in the sweet spot?
☐ Am I prepared to iterate?
Fix these 7 mistakes, and you'll instantly get better outputs from any AI model.