How to Write Better AI Prompts (With Real Examples)
Most people get generic AI outputs because they write generic prompts. Here's how to fix that — with real before-and-after examples.
The quality of what you get from an AI tool is almost entirely determined by the quality of what you put in. This sounds obvious until you realise that most people write prompts the same way they write Google searches — short, vague, and keyword-heavy.
AI tools aren't search engines. They're more like a very capable colleague who needs context to do good work. The more clearly you brief them, the better the output.
Here are four principles — each with a real example — that will immediately improve your results.
1. Give it a role
Before you ask anything, tell the AI who to be. This isn't magic — it's context. AI tools are trained on vast amounts of human-generated content, and giving it a role helps it draw on the right patterns.
Try this
“You are an experienced HR professional helping a first-time manager navigate a difficult performance conversation. [Your question here]”
The quality difference is immediate. Try it on your next prompt and you'll see it right away.
2. Be specific about what you want
Vague prompts produce vague results. Every constraint you add — tone, length, audience, format — narrows the AI's output toward something actually useful.
Writing an email
❌ Weak
“Write an email about the project delay.”
✓ Strong
“Write a concise, professional email to my client explaining that the website project will be delayed by 2 weeks due to unexpected technical issues with the payment integration. Tone should be apologetic but confident. End with a revised timeline and next steps.”
Why it works: The bad prompt leaves everything open — who's it for, what's the context, what tone? The good prompt gives the AI a specific situation, audience, tone, and desired outcome. You'll get something usable on the first try.
Getting an explanation
❌ Weak
“Explain machine learning.”
✓ Strong
“Explain machine learning to me as if I'm a marketing manager with no technical background. Use a real-world analogy and keep it under 200 words.”
Why it works: "Explain machine learning" gets you a textbook answer. Specifying your background, asking for an analogy, and setting a length constraint gets you something you'll actually understand and remember.
3. Tell it what good looks like
If you have an example of what you're aiming for — a previous email you liked, a format that worked — paste it in. AI tools are excellent at matching patterns when you show them what pattern to match.
Editing your writing
❌ Weak
“Make this better: [paste text]”
✓ Strong
“Edit this paragraph for clarity and conciseness. Remove any jargon. Keep the core argument intact but tighten the language. The audience is non-technical business professionals: [paste text]”
Why it works: "Make this better" is too vague — better how? Clearer? More formal? Shorter? Defining what better means gives the AI a target to aim at.
Brainstorming ideas
❌ Weak
“Give me some ideas for a blog post.”
✓ Strong
“Give me 10 blog post ideas for a B2B SaaS company that sells project management software to construction firms. Focus on practical, how-to topics that address common pain points like scheduling delays and budget overruns. Avoid generic listicles.”
Why it works: Generic brainstorming request = generic ideas. Specific context = ideas you'd actually consider writing.
4. Iterate, don't start over
The first output is rarely the final output. The best use of AI is a conversation — you respond to what it gives you and refine from there.
Instead of starting a new prompt when the output isn't right, try:
"Make the third paragraph shorter and more direct."
"The tone is too formal — make it sound more conversational."
"Good, but the opening is weak. Give me 3 alternative opening sentences."
"This is close. Can you do the same thing but assume the reader has no technical background?"
Treating the conversation as iterative rather than transactional is the single biggest mindset shift that separates people who find AI genuinely useful from those who don't.
The four principles
Want to go deeper?
Kenivo has an entire module on this.
Interactive lessons. Real exercises. A built-in AI tutor to answer questions as you go. Start learning → and start for $9 USD/month — founding rate.
Start learning → →