Manual meeting notes are one of the most avoidable productivity taxes in modern work. AI tools can handle transcription, summarization, and action item extraction — leaving you to actually be present in meetings. Here’s how to set up an AI meeting notes system.

The Full AI Meeting Notes Stack

A complete system covers:

  1. Transcription — capturing what was said
  2. Summary — synthesizing the key points
  3. Action items — extracting tasks with owners and deadlines
  4. Distribution — sending to relevant parties
  5. Storage — making notes searchable later

Tool Options by Meeting Platform

Zoom

  • Zoom AI Companion (included in Zoom Pro) — automatic summaries, action items, and Q&A
  • Otter.ai — bot joins your Zoom and transcribes, better search and organization
  • Granola (Mac only) — combines your notes with auto-transcription for better quality

Google Meet

  • Otter.ai — works with Google Meet via calendar integration
  • Fireflies.ai — bot joins Meet, good CRM integration
  • Google Meet’s built-in transcription (Workspace plans) — basic, no AI summary

Microsoft Teams

  • Teams Copilot (M365 Copilot) — native transcription and AI summary
  • Otter.ai — also works with Teams

In-Person / No Video Platform

  • Otter.ai mobile app — record directly, auto-transcribes
  • Granola — Mac app that uses your mic, doesn’t require a video call
  • Claude / ChatGPT — bring rough notes, AI structures them

Setting Up Automatic Transcription

Otter.ai (cross-platform)

  1. Connect Otter to Google Calendar or Outlook
  2. OtterPilot will auto-join any calendar meeting with a video link
  3. After the meeting, you get:
    • Full transcript (searchable)
    • AI summary
    • Action items
    • Shared with participants if you want

Configuration tip: Set OtterPilot to “auto-join all meetings” to make this zero-effort. Review the “Meeting preferences” to exclude internal standup calls if transcript isn’t needed.

Granola (Mac, best quality)

Granola works differently — it records your meeting audio in the background and combines it with any notes you type during the meeting.

  1. Open Granola before a meeting
  2. Click “New Meeting”
  3. Take a few rough notes during the call (just keywords/highlights)
  4. After the meeting, Granola generates notes that combine your notes + transcription

The result is higher quality than pure-AI transcription because your notes guide the summary.

Processing Rough Notes with AI

If you don’t have an automatic transcription tool, process your rough notes with Claude:

Prompt:

I took rough notes during a meeting. Please structure them into proper meeting minutes.

Raw notes: [paste your rough notes]

Output format:
**Meeting:** [date, attendees if I noted them]
**Objective:** [what the meeting was trying to accomplish]

**Key Decisions:**
[bullet points]

**Action Items:**
- [ ] [Task] — Owner: [name] — Due: [date/timeframe]

**Discussion Summary:**
[2-3 paragraph summary of what was covered]

**Next Meeting / Follow-ups:**
[any next steps]

Action Item Extraction

For any meeting transcript, AI extracts action items reliably:

Prompt:

Read this meeting transcript and extract all action items.

Transcript: [paste transcript]

Format each action item as:
- [ ] [Specific task] — Owner: [person who agreed to do it] — Due: [deadline mentioned, or "TBD"]

Criteria for inclusion:
- Someone specifically agreed to do something
- A deadline was mentioned
- Include "check on" and "follow up" items

Note any items where ownership was unclear.

Automated Follow-Up Email

After a meeting, generate the follow-up email automatically:

Prompt:

Write a follow-up email for this meeting.

Meeting summary: [paste AI-generated summary]
Action items: [paste action items]
Recipients: [names and roles]
Tone: [professional/casual based on relationship]

Email should:
- Brief thank you (one sentence)
- Summarize the 2-3 key decisions/outcomes
- List action items clearly with owners
- Next meeting date if relevant
- Be under 200 words

Subject line options: give me 3 variants

Building a Searchable Meeting Database

Over time, your meeting notes become a knowledge base. Use Otter.ai’s search or build your own:

Notion-based system

  1. After each meeting, copy the AI summary into a Notion database
  2. Tag by: project, meeting type, attendees, date
  3. Use Notion AI to query: “What did we decide about the pricing strategy?”

Obsidian-based system

  1. Export meeting notes as Markdown
  2. Store in Obsidian vault
  3. Use the Daily Notes template to auto-create meeting note files
  4. Search across all notes with Obsidian’s built-in search

Meeting Notes Templates

Create standard templates for recurring meeting types:

Weekly Team Standup

**Team:** [name] **Date:** [date]

**Progress (since last week):**
[bullet points]

**Plans (next week):**
[bullet points]

**Blockers:**
[bullet points]

**Action Items:**
[auto-extracted]

Client Check-in

**Client:** [name] **Date:** [date] **Duration:** [X min]

**Account Status:**

**Issues Discussed:**

**Decisions Made:**

**Action Items:**
- [ ] [task] — Owner: [us/client] — Due: [date]

**Next Check-in:** [date]

ROI of AI Meeting Notes

For a knowledge worker in 10 meetings per week:

  • Manual notes: 15-20 min per meeting = 2.5-3 hrs/week
  • AI notes: 2-3 min review per meeting = 20-30 min/week
  • Saved: 2+ hours per week

At any professional hourly rate, this pays for the tool (Otter.ai Pro at $16/month) many times over.

The secondary benefit — being actually present in meetings instead of furiously taking notes — also improves meeting quality and relationships.