Plan K: AI Meeting Notes Workflow

Record & transcribe → AI summary → extract action items → archive & distribute → follow up & review

Time
15-30 min / session
Cost
Low
Difficulty
Beginner

Best for

Best for regular meetings, client calls, brainstorming and interviews — turning 'meetings you had' into executable knowledge assets.

Steps

Record and transcribe

Automatic

Otter / Fireflies / Tongyi Tingwu

Auto-record and transcribe the meeting. For Chinese meetings, domestic tools like Tongyi Tingwu transcribe more accurately. Confirm recording consent — that's the compliance baseline.

💡Skim the transcript first and fix mis-transcribed names/jargon so they don't pollute the summary.

AI summary

5 min

ChatGPT / Claude

Feed the transcript to an LLM for a structured summary: meeting goal, key discussions, decisions. Give it an output template (goal / discussion / decisions / todos) for consistent format.

💡Provide a 'notes template' as a few-shot example so output archives consistently.

Extract action items

5 min

ChatGPT / Claude

Have the model pull every action item from the transcript: what was decided, who owns it, deadline. Action items are the most valuable part of meeting notes.

💡Ask for 'owner + task + due' triples so you can import them straight into a task tool.

Archive and distribute

5 min

Notion / Feishu Docs

Archive the summary plus the raw transcript in the knowledge base with date, attendees and related docs; auto-share to attendees and sync tasks to calendar/board.

💡Keep a link to the raw transcript so people can check details without replaying the whole call.

Follow up and review

Weekly

Task tools + calendar

Check last round's action items at the start of the next meeting and do a short review; escalate anything that keeps slipping.

💡Make action items a fixed opener in your meeting template — clear old debts first, then new business.

Tools involved

Otter / Fireflies / Tongyi TingwuChatGPT / ClaudeNotion / Feishu DocsCalendar / task tools

Pitfalls to avoid

Compliance and privacy

Recording and transcription need attendee consent; be cautious with cloud transcription for confidential meetings.

Speaker misattribution

Multi-speaker meetings often mislabel who said what; manually confirm 'who said what' for key decisions.

Missed action items

AI can miss implicit commitments; walk through an 'owner + task + due' list line by line.

Over-compression loses context

Too-tight summaries drop discussion background; keep the raw transcript link for reference.

Related Workflows

Plan I: AI Presentations (Slides)

Topic & audience → outline & storyline → per-slide generation → visual consistency → speaker notes → rehearsal

Time：
30-60 min / deck
Cost：
Low
Difficulty：
Beginner

Plan J: AI Data Analysis / Research Report

Define the question → clean data → exploratory analysis → deep modeling → visualize & narrate → conclusions

Time：
2-6 hours / report
Cost：
Low-Medium
Difficulty：
Advanced

Plan M: AI Learning / Paper Reading Workflow

Define the goal → gather materials → AI deep reading → active quizzing → knowledge base → spaced review

Time：
2-3 hours / topic
Cost：
Low
Difficulty：
Beginner