Live Event
June 10, 2026
APAC Disaster Management AI Skills Jam: Learn Sessions Companion
Practice prompts and short exercises for Deep Research, data analysis, and spreadsheet workflows.
How to use this page
Use this page during the Learn sessions, or come back later to practice. You do not need to use every prompt. Start with the section for the session you are in.
- In ChatGPT 201 + Deep Research: start with the Deep Research practice flow.
- In Data Analysis + Excel: start with the dataset practice flow.
- After the workshop: pick one practice flow and try it with public, non-sensitive material from your work.
Quick links
Before you start
- Use public, fictional, minimized, or workshop-approved information only.
- Check sources, dates, numbers, names, and assumptions before sharing or acting.
- Keep human review visible, especially for public-facing or operational work.
Learn 1: ChatGPT 201 + Deep Research
This session helps you move from one-off prompting to more reusable ways of working: Deep Research, better prompts, Projects, and practical outputs.
Learning goals
- Use Deep Research to explore a public disaster-management question.
- Turn research into an output someone can review, share, or adapt.
- Use simple verification habits before sharing results.
Core practice flow
Step 1: Start with a Deep Research question
Use Deep Research when you need a sourced research brief across public information. The goal is not to accept the result automatically. The goal is to get a structured draft that you can inspect.
Use Deep Research to compare how cities around the world are adapting to flooding. Focus on what has been tried, what seems to be working, what has not worked well, and what lessons may be useful for disaster-management teams in Asia.
Try your own version:
Use Deep Research to help me understand [topic or decision] for [country or region]. Focus on practical lessons for disaster-management teams. Include credible sources, key risks, and what a human reviewer should verify.
What good looks like: the result names sources, explains what is known and uncertain, and gives practical lessons without pretending every city or country is the same.
Step 2: Improve the prompt before you run it
Meta-prompting means asking ChatGPT to help you write a better prompt. This is useful when you know the goal but are not sure how to ask.
Help me write a better image prompt for a public flood preparedness flyer for residents in Bangkok. The image should be calm, practical, locally appropriate, and avoid panic or inaccurate safety guidance.
You can use the same pattern for Deep Research: “Help me improve this Deep Research prompt so it is clearer, more specific, and easier to verify: [paste prompt].”
Step 3: Turn research into a usable output
Research is more useful when it becomes something a person can review or use.
Turn this research into a one-page briefing for disaster-management leaders.
Other useful outputs: a public FAQ, a checklist for local teams, a 5-slide outline, or a list of questions to ask local experts.
Step 4: Organize related work in a Project
Use Projects when you have related chats, files, or repeated outputs. For example, create a Project called Flood adaptation research and move related research, briefing, and image-prompt chats into it.
Project instruction: “Keep outputs practical for disaster-management teams. Use plain language, cite sources where relevant, and clearly separate evidence from assumptions.”
Optional activity: image challenge
If there is extra time, choose one public-facing image idea and use meta-prompting before creating it. Keep the image calm, locally appropriate, and easy to understand.
- Flood preparedness flyer for residents.
- Heat safety announcement for older adults and families.
- Evacuation center sign.
- Cyclone preparedness poster for coastal communities.
- Community volunteer recruitment image.
After-session practice
- Pick one public question from your work.
- Run a Deep Research prompt.
- Turn the result into one practical output.
- Ask ChatGPT what a human should verify before using it.
Learn 2: Data Analysis + Excel
This session helps you use ChatGPT to understand an existing dataset and create a useful spreadsheet from scratch. You can use ChatGPT for Excel, Google Sheets, or ChatGPT Data Analysis.
Learning goals
- Ask simple first questions before drawing conclusions from data.
- Turn a dataset into a short briefing or chart recommendation.
- Create a first-draft spreadsheet or tracker that a human can review.
Choose your starting point
- If you have Excel or Google Sheets: use the spreadsheet you opened in the session.
- If you do not have Excel: upload a CSV, screenshot, pasted table, or use ChatGPT Data Analysis.
- If you need a sample file: use the natural-disasters dataset linked above.
Core practice flow
Step 1: Understand the dataset
Start simple. Ask what the file is before asking for findings or recommendations.
Can you tell me what this dataset is?
Useful follow-ups:
- “What might be notable here for disaster-management professionals in Thailand?”
- “What might be notable here for disaster-management professionals in Asia?”
- “What should a disaster-management team pay attention to in this dataset?”
What good looks like: the answer explains the dataset plainly, points out limits, and does not jump to a recommendation before checking what the data can and cannot show.
Step 2: Create a short briefing
Prepare a 2-minute briefing for a disaster-management team from [country]. Find one useful insight, one surprising pattern, and one caveat we should verify before sharing.
Then ask: “What chart would help explain the most important pattern?” If time allows, ask ChatGPT to create the chart or explain how to create it in the spreadsheet.
Step 3: Build a spreadsheet from scratch
ChatGPT can also help create structure: trackers, budgets, reporting templates, and simple summary views.
Create a simple spreadsheet for tracking flood response coordination across districts.
Useful follow-up: “Add sample rows, status options, formulas, and a simple summary view.”
Step 4: Make the spreadsheet more useful
Improve this spreadsheet by adding one useful feature.
Useful features might include a summary tab, priority score, validation checklist, drop-down status options, or a simple chart.
What good looks like: the spreadsheet has clear columns, a clear purpose, simple status options, and a short checklist for what a human should verify before using it.
Practice options
Choose one. The goal is a useful first draft, not a finished operational tool.
- Shelter status tracker.
- Public information message tracker.
- Resource request tracker.
- First-draft budget for a disaster response activity.
After-session practice
- Open a public dataset or an approved spreadsheet.
- Ask ChatGPT what the file is.
- Ask for one insight, one caveat, and one chart recommendation.
- Ask ChatGPT to create or improve a simple spreadsheet based on the same use case.
Ready to go further? Try Codex
Codex is an AI teammate for building, editing, and improving digital work products. During the Builder Lab, builders will use Codex to help create prototypes, websites, dashboards, and other reusable tools.You do not need Codex for today’s Learn sessions. But if you are curious, check out Codex on OpenAI Academy to learn more abotu what it can do and how non-technical users can work with builders more effectively.
Which tool should I use?
Use this | When you need |
|---|---|
ChatGPT | A quick draft, summary, checklist, or first-pass analysis. |
Deep Research | Public sources, citations, comparisons, and caveats. |
Projects | A place to keep related chats, files, and repeated outputs together. |
ChatGPT for Excel or Google Sheets | Help inside a workbook. |
ChatGPT Data Analysis | Analysis of a CSV, spreadsheet, screenshot, or pasted table. |
Codex | A working prototype, website, dashboard, script, or lightweight tool you can inspect, test, and refine |
Simple rule of thumb
Start with the simplest tool that gets you to a useful first draft. Then verify before sharing, publishing, or acting.
Table Of Contents