Aulafy02 / Everyday practice / Lesson 3 of 8
Write a clear request
Write instructions that specify the task, context, limits and format.
01 / Understand
A prompt is simply your instruction
You do not need special vocabulary or a secret phrase. Say what you want done. Include the information needed to do it, who the result is for and how you want it presented. 'Write an email' leaves many decisions open. 'Draft a polite request to reschedule a meeting, under 80 words, without choosing a new time' is easier to evaluate.
Give facts and mark the boundaries
Supply facts you want the draft to preserve. State what the assistant must not invent, such as names, prices or dates. When a fact is missing, ask for a question or a visible placeholder. Instructions reduce ambiguity, but do not guarantee accuracy. You will still compare the response against your requirements.
Improve one thing at a time
If the first draft is too long, ask for a shorter version while preserving the facts. If the tone is wrong, name the desired tone and audience. Changing one requirement at a time helps you see what improved and what was lost. Recheck the facts after each revision.
02 / Worked example
Look closely at the result
A written teaching example, not a live AI response.
Draft a polite email to my project team asking to reschedule a meeting. Under 80 words. The only confirmed fact is that I cannot attend the planned time. Do not invent a reason or propose a time. Ask for availability.
Hi team, I cannot attend our meeting at the planned time. Could we find an alternative? Please share your availability so we can agree on a new time. Thank you.
Compare with the tutor’s review
The draft preserves the one supplied fact, asks rather than decides, and adds no reason or date. You would still check the recipients and original meeting details before sending.
03 / Your turn
Put it into practice
Write your own request for a short draft using the four fields below. Use fictional information, then compare your request with the worked example.
Use fictional information. Nothing you write here is sent to a server or an AI. Download your work before leaving; it is not saved automatically.
04 / Check your understanding
One decision before you go
Take this with you
Clear task + useful context + explicit limits + output format.
Complete the practice, review each criterion and check your answer to finish your self-review. Every lesson remains open.
Next: Check before you trustSources & course notes
Reviewed . Examples and exercises are original Aulafy teaching material under MIT. These references inform the course’s approach to reliability and learning; they are not endorsements.
- NIST: Generative AI risk profileBackground on unreliable outputs and evaluating risks.
- UNESCO: Generative AI in educationBackground on human agency, privacy and learning.