One worker gave ChatGPT dozens of anonymised meeting notes and spreadsheet comments and asked for a project plan. It grouped the ideas, spotted missing information and asked better questions; they made the final decisions, but it saved hours of work.
Here are my notes from several meetings about [PROJECT], with names and confidential details removed: [PASTE NOTES]. Group the ideas into themes, list what's been decided and what hasn't, point out any missing information, and draft a project plan with tasks, rough deadlines and owners left as [OWNER]. Finish with the five questions I should ask the team next.
Replace the [PLACEHOLDERS] with your own details before you send it.
How to try it
- Remove names and anything confidential before you paste your notes in.
- Run the prompt and check every ‘decision’ against what was really agreed.
- Take its questions to your next meeting and fill in the owners together.
Why this works
Ideas from several meetings are easy to lose, and it’s hard to see what’s been agreed. Grouping notes into themes and separating decisions from open questions gives you a starting plan, and leaving owners blank keeps the real decisions with the team.
Tips for a better result
- Include dates on each set of notes so the AI can see how ideas changed.
- Ask it to quote the note behind each decision so you can check quickly.
- Ask for the plan in a format your team already uses, such as a table or checklist.
- For capturing meetings better in future, see our guide to meeting notes and summaries with AI.
Common mistakes to avoid
- Treating the AI’s ‘decisions’ as agreed without checking with the team.
- Pasting confidential details your company hasn’t cleared for AI tools.
- Setting deadlines without asking the people who’ll do the work.
Try a variation
Here’s a variation of our own (not from the original source) to try next:
Here are notes about [PROJECT], anonymised: [PASTE]. List the risks and assumptions hidden in them, rate each by how likely it is to cause trouble, and suggest one small step to reduce each of the top three. Keep owners as [OWNER].
For more ideas like this, have a look at our AI for work and business hub.
More work use cases
Write a tricky Excel formula with ChatGPT
One user had ChatGPT write an Excel formula for a billing file he fills in every month, saving him two to three hours of manual work each time.
In Excel, I have [DESCRIBE COLUMNS, E.G. COLUMN A: DATE, B: CLIENT, C: HOURS]. I want [WHAT YOU WANT, E.G. TOTAL HOURS PER CLIENT FOR LAST MONTH IN COLUMN F]. Write the formula, explain each part in plain English, and tell me how to copy it down safely.
Tailor your CV to a job advert with ChatGPT
A consultant's main CV advice to one jobseeker was to paste the job listing and CV into ChatGPT and ask it to reword the CV to get past automated filters. The jobseeker found that advice demoralising, so treat it as a tool, not a substitute for your own judgement.
Here's a job advert: [PASTE]. Here's my CV: [PASTE]. Show me which requirements in the advert my CV already covers, which are missing or buried, and suggest honest rewording of my existing experience to match the advert's language. Don't invent skills or experience I haven't listed.
Turn your job search into a tracker you can learn from with Claude
A recent master's graduate used Claude like a career coach: she pasted job adverts next to her CV to see which key words she was missing, built separate CVs for three kinds of role, and tracked 186 applications through to one offer. Seeing it as a funnel made rejections feel less personal and showed her what to improve.
Here's a job advert: [PASTE ADVERT]. And here's my CV: [PASTE CV]. List the main skills and key words the employer is looking for, which ones my CV already shows, and which ones are missing or not clear enough. Don't invent experience; suggest where my real experience could be described more clearly. Then add this application to my tracker table: company, role, date, stage, notes.
Reference: X @KAWLLF