Prompting
A Practical Prompting Guide for Better AI Results
Better prompts are usually not about clever wording. They are about giving the AI the information it needs to do the job properly.
7 min read
In this guide
You do not need a collection of “magic prompts.”
A strong prompt usually answers a few simple questions:
Who should the AI act as? What exactly should it do? What context does it need? What should the result look like? What should it avoid?
1. Give it a useful role
A role helps set the perspective.
Weak:
“Write this.”
Better:
“You are a customer-care specialist for a professional training company.”
The role should help with the task.
Do not add unnecessary personas just to make the prompt longer.
2. Define the task clearly
Say what you actually want.
Instead of:
“Help me with this email.”
Try:
“Rewrite this email so it is shorter, professional and easy for the client to understand.”
Clear tasks reduce guessing.
3. Provide the context
AI cannot automatically know the information in your head.
Include the context that changes the answer.
For example:
- who the audience is
- what happened before
- what service or product is involved
- what information must be included
- what constraints exist
4. Describe the output
Tell the AI what a good answer should look like.
You can specify:
- length
- language
- tone
- format
- sections
- bullets
- table
- number of options
Example: “Give me three options. Each should be under 25 words and suitable for WhatsApp.”
5. Add boundaries
Tell the AI what it should not do.
Examples:
- do not invent statistics
- do not add information I did not provide
- do not use technical jargon
- do not change the approved policy wording
- ask me if critical information is missing
Constraints can be as important as the task itself.
6. Give examples when style matters
If you want a particular tone, one good example can be more useful than ten adjectives.
Instead of: “Make it premium, elegant, human, sophisticated and engaging.”
Show the AI one short example of writing you like.
Then explain what you like about it.
7. Separate information from instructions
When prompts become longer, structure them.
A simple format:
- Role
- Task
- Context
- Requirements
- Output format
- Constraints
This makes the prompt easier for both you and the AI to follow.
8. Improve the prompt through feedback
You do not need to get the final answer in one attempt.
Useful feedback sounds like:
- “Make it 30% shorter.”
- “The tone is too formal.”
- “Keep the first paragraph and rewrite the rest.”
- “Do not change the numbers.”
- “This sounds like advertising. Make it more natural.”
Specific feedback produces better revisions.
A reusable prompt structure
ROLE — Who should the AI act as?
TASK — What should it do?
CONTEXT — What does it need to know?
REQUIREMENTS — What must be included?
OUTPUT — What should the result look like?
CONSTRAINTS — What must it avoid?
Example
Instead of: “Write a caption about our AI course.”
Use:
ROLE: You are writing social media content for a professional AI training company in Abu Dhabi.
TASK: Write an Instagram caption promoting an AI agents diploma.
CONTEXT: The audience is professionals and business owners who already use AI tools but want to understand agents and practical workflows.
REQUIREMENTS: Focus on practical learning and building an AI agent.
OUTPUT: One Arabic caption, maximum 70 words, followed by a short CTA.
CONSTRAINTS: Use simple modern Arabic. Do not use exaggerated claims or generic phrases such as “the future starts now.”
Good prompting is not about finding the perfect sentence.
It is about making your expectation clear.
Ai MindUp
Want to go beyond prompting?
Our practical training covers prompting, AI workflows, agents and applied use cases.