Best use case
Use this pattern when the goal is clear but the details are incomplete, scattered, or easier to answer through a guided interview.
This lecture page shows how to build a prompt in sections, then demonstrates how ChatGPT can ask one question at a time to gather the information needed to write a professional email requesting time off over the next holiday.
Instead of the user guessing every detail needed for a complete prompt, ChatGPT leads the interaction by asking targeted questions.
Use this pattern when the goal is clear but the details are incomplete, scattered, or easier to answer through a guided interview.
Tell ChatGPT to ask questions one at a time, continue until enough information is gathered, then produce the final output.
The user wants a professional email asking for time off over the next holiday, but does not yet know what information the email needs.
The prompt below is assembled from the six checklist items shown here.
Each section in the simulated prompt maps directly to one part of the good-prompt checklist.
Give ChatGPT the expertise needed to guide the user and write workplace communication.
Define the final deliverable: a professional email requesting time off.
Name the situation: requesting time off over the next holiday.
Use the flipped interaction: ask one question at a time until enough information is gathered.
Tell ChatGPT how to structure the final email.
Set standards for tone, clarity, and completeness.
Act as a professional workplace writing coach.
Help me create a professional email asking for time off.
The email will request time off over the next holiday.
Ask me questions one at a time to gather the information you need. Do not write the final email until you have enough information.
When ready, write the final answer as an email with a subject line, greeting, concise body, and professional closing.
Make the email polite, specific, concise, and ready to send. If any detail is uncertain, use wording that avoids overcommitting.
The learner sees that ChatGPT controls the interview, asks only one question at a time, and uses each answer to decide what to ask next.
The final email uses the information gathered during the one-question-at-a-time conversation.
Hi Priya,
I hope you are doing well. I would like to request time off from Wednesday, November 27 through Friday, November 29 for the Thanksgiving holiday.
I will make sure my urgent tasks are completed before I leave, and Jordan can help handle anything time-sensitive while I am out.
Please let me know if this time off is approved or if you would like me to adjust anything before then.
Thank you,
[Your Name]
The final email is not generic. It includes recipient, holiday, exact dates, coverage plan, tone, and the request for approval because ChatGPT gathered those details before drafting.
Use these expandable notes to guide class discussion after the demonstration.
The normal interaction starts with the user asking questions and ChatGPT answering. In this pattern, ChatGPT asks the questions needed to complete the user’s goal.
One question at a time lowers cognitive load, makes the interaction feel conversational, and helps the model adapt each next question based on the previous answer.
The model might still produce an email, but the result may be too wordy, too casual, vague about approval, or missing careful wording around uncertain details.
The same pattern can guide performance-review writing, project updates, troubleshooting summaries, meeting agendas, customer responses, and any task where the desired output depends on details the user may not provide upfront.