AI 201 Apply: Everyday Tasks
Meeting Minutes with AI
17 June 2026
Learn how to turn messy meeting notes into clean, circulation-ready minutes in minutes.
Why this matters
Every meeting ends the same way: someone has to turn a page of raw shorthand notes into something comprehensive and fit for circulation. Formatting the headings. Untangling who-said-what. Separating the actual decisions from the discussion. Chasing down every action and its owner. It's necessary work, but it draws your best officers away from higher-value work. But what if you could just paste your raw notes and receive perfectly formatted minutes, tailored to your team's style, in a matter of minutes?
Your five-part recipe - C.R.A.F.T.
Use this structure to guide your thinking. Every effective AI task is built around five components — together, they spell CRAFT:
C.R.A.F.T.
C | 📂 Context | Your raw meeting notes, transcripts, or bullet points. |
R | 🎭Role (who the AI should act as) | A Meeting Minutes Writer for your team or committee. |
A | ⚡Action (what the AI needs to do) | Clean and concise minutes, using only what was said. |
F | 📋Format | Meeting details → discussion points → decisions made → Action table → Risks. |
T | 🔍Test | Ensure no decision or action was dropped, and that the minutes accurately reflect what was discussed. |
Here’s what you need for this exercise:
Tools and Input required
🛠️ Tool | ChatGPT Enterprise (WOG) |
🔗 Feature | File Upload |
📂 Your input | Raw notes, approved minutes |
Step-by-step guide
Step 1: Gather your notes. Collect the raw notes or transcript from your meeting. The clearer the notes (especially names written next to actions), the better the output. For practice, copy one of the sample notes blocks below.
Sample A: Tender evaluation discussion
Tender Eval mtg 14 jun, MR 4.2, ~2.30pm. present: Priya (chair, Procurement), Daniel (IT), Mei Ling (Finance), Hafiz (User Dept/Ops), Sarah (Legal) joined late. apologies: Wei (away)
re: ITQ for managed helpdesk svc, 3 yr. 4 bids in. shortlist down to 3 after compliance check - VendorA, VendorB, VendorC. VendorD knocked out (missed mandatory ISO cert, non-compliant).
pricing - VendorA cheapest ~$1.42m total, VendorB ~$1.58m, VendorC ~$1.69m. Mei Ling flagged VendorA pricing looks low on Yr1, possible lowballing, manpower rates below market? need clarification. action - Priya to send clarification q to VendorA.
technical - Daniel scored VendorB highest on technical (SLA 99.9%, local L2 team, strong transition plan). VendorA technical ok but transition plan thin, only 2 wks onboarding, Hafiz worried that's too short given our ticket volume. VendorC strong on tooling but proposed offshore L1 - data residency concern, Sarah to advise whether that breaches our data classification reqs.
discussion - Hafiz strongly prefers B on service quality, willing to pay premium. Mei Ling cautioned must justify not picking lowest price, need clear value-for-money writeup. agreed evaluation must follow PQ then price-quality 60/40 weighting as per tender doc.
decisions: VendorD confirmed non-compliant & excluded. proceed to clarification round with VendorA (pricing) and VendorC (data residency) before final scoring. final scoring meeting to be set next week.
actions - Priya: send clarifications to A and C by 17 jun. Sarah: advise on offshore L1 data residency by 18 jun. Daniel: finalise technical scores once clarifications in. Mei Ling: prep value-for-money template. Hafiz: confirm current monthly ticket volume figures for transition assessment.
risk - timeline tight, contract must be awarded before current vendor contract expires end Aug. if scoring slips we may need bridging arrangement. Priya to check bridging options w current vendor as contingency.
next mtg: 23 jun (tbc), final scoring + recommendation to approving authority.
Sample B: Project steering committee
PSC #6 - "OneService Portal Revamp" - 12 Jun, 10am, hybrid (MR 7 + Zoom). chair: Director Lim (Sponsor). present: Project Lead Aisha, Tech Lead Ben, Change/Comms Grace, Finance rep Kumar, Vendor PM Ravi (dialed in). note-taker: me.
status update from Aisha - project amber. UAT started but 3 wk behind original plan. 2 main reasons: (1) integration with legacy CRM more complex than scoped, (2) two key testers pulled onto another priority. overall delivery date may slip from Sep to mid-Oct unless we add resources.
Ben - integration issue: legacy CRM API undocumented, team reverse-engineering. needs 1 more backend dev for 6 wks. Ravi (vendor) can provide at additional cost ~$45k. Kumar said budget has contingency but needs Director Lim sign off as it's above his delegation.
decision - Director Lim approved engaging 1 additional vendor backend dev for 6 wks, funded from project contingency. Kumar to process variation order.
Grace - comms/change plan ready, but training rollout depends on go-live date. cannot finalise schedule until date confirmed. agreed to hold training comms until next PSC.
risk discussion - if Oct go-live, clashes with year-end peak service period, ops may resist. Director Lim wants a fallback: consider phased go-live (core features Sep, secondary Oct). Aisha to assess feasibility of phased approach with Ben.
Kumar raised budget - currently 68% spent, on track but the variation will push to ~78%. still within approved envelope. flagged we should not absorb further scope without a budget review.
decisions recap: (1) approve extra backend dev from contingency. (2) explore phased go-live. (3) hold training comms till date locked.
actions - Aisha: assess phased go-live feasibility, report by 19 Jun. Ben: onboard new dev, target start 16 Jun. Kumar: raise variation order this week. Grace: draft two comms timelines (Sep full vs phased) for next PSC. Ravi: confirm dev availability by 13 Jun.
next PSC: 26 Jun 10am. Aisha to circulate phased go-live assessment beforehand.
Step 2: Open ChatGPT Enterprise.
Start a new chat in your WOG workspace.
Step 3: Enter the prompt below then paste your notes underneath it.
Use this sample prompt
You are a Meeting Minutes Writer for [your team / committee e.g. "the Tender Evaluation Committee"]. Your role is to transform raw meeting notes into professional meeting minutes.
Follow these rules:
Use only information provided in the notes. Do not invent decisions, actions, or attendees.
Consolidate duplicate points and remove conversational repetition.
If information is missing, indicate "[Not stated]".
Format every meeting using the structure below:
Meeting Details: Date, Title, Chairperson, Attendees
Key Discussion Points: (grouped by topic)
Decisions Made: (bullet points)
Action Items: (table with Action, Owner, Due Date)
Risks or Issues Raised: (bullet points)
Here are the raw notes: [paste your notes here]
Enter the prompt into a new chat
where to paste your notes
For readability, you can paste the sample notes on a new line
Send your prompt and watch ChatGPT work
Step 4: Review the output. Check that no decisions or context were missed, and that every action has the right owner and date. Anything the notes didn't state should appear as "[Not stated]" rather than AI’s guess.
![Due dates were not mentioned in the original messy minutes. As such, AI states it as “[Not stated]](https://isomer-user-content.by.gov.sg/543/6298662e-0087-48ff-ad7d-d9d93aceaa17/5_RefineAIOutput.png)
Due dates were not mentioned in the original messy minutes, as such, AI states it as “[Not stated]
Step 5: Refine and fill the gaps. The AI flags anything missing from your notes as "[Not stated]" rather than guessing. Complete the record with a follow-up prompt: supply the real details and ask the AI to slot them in. Enter this follow-up prompt:
Use this sample prompt
For every field you marked "[Not stated]", fill in these details:
1. The meeting was held on 14 June 2026
2. Due date for all deliverables in 2026
3. Daniel and Mei Ling - both 24 June
4. Hafiz - 18 June
5. Priya - 20 June
Update the minutes accordingly and keep everything else unchanged.
Iterate with AI with a follow-up prompt
You can also tidy the Action Items table using this next follow-up prompt:
Use this sample prompt
Sort the Action Items table by Due Date (earliest first). Keep everything else unchanged.
Outputs not ordered
Iterate again
Check final output
Step 6: Save as a Skill. Package the instructions so you can reuse the exact format next time:
Use this sample prompt
Can you package this set of instructions as a reusable Skill called 'Meeting Minutes Writer'? Next time, I will just paste raw notes and you will apply this format.
Enter sample prompt
ChatGPT building the Skill
Click install
Step 7: Install the Skill.
After installing, you can try it in a new chat.
Now it's your turn to try!
Take 10 minutes to generate a clean and concise set of minutes, ready to be circulated.
📂 Prepare your raw notes and templates. If possible, prepare a set of meeting notes that were already approved to help the AI align the tone and exact format that is expected.
⚡ Review Critically. Never let AI invent points. Ensure that it did not miss anything as well.
🔍 Test and Share. If this worked well for you, consider saving the workflow as a Skill and sharing it with others.
You've just cut hours of work producing a clean and concise set of meeting minutes, ready to be circulated.
📅Following the 12-Week Learning Plan?
Congratulations on reaching Week 6! This guide is part of a 12-week learning plan designed to help you build practical AI skills. You're well on your way to becoming AI fluent.
Preview the 12-Week Learning Plan
12 Week Learning Plan
Week 1 | |
Week 2 | Webinar: Introduction to AI-First |
Week 3 | Webinar: Open AI ChatGPT Foundations |
Week 4 | Webinar: Introduction to AI-First |
Week 5 | Using AI responsibly |
Week 6📍 | Meeting Minutes with AI 📍 |
Week 7 | Generate a Daily Morning Brief |
Week 8 | AI Apply 201: Domain Specific Tasks |
Week 9 | Workflow Mapping & Peer Sharing |
Week 10 | |
Week 11 | Your AI Playbook |
Week 12 | |
Week 13 | CAT Brownbag at Lorong |
⏭️Next Guide: Building PowerPoint Decks with AI
↩️Return to: AI 201 Apply Learning Pathway