The Problem with Meetings
The average business meeting lasts 45-60 minutes. Of that time, between 30 and 40% is repetition, tangents, or clarification of points already discussed. And at the end, half the attendees leave with a different understanding of what was decided and who needs to do what.
AI does not fix poorly planned meetings. But it can eliminate the manual work of transcribing, summarizing, and distributing what was discussed.
This tutorial covers the complete workflow: record → transcribe → summarize → distribute action items.
The Workflow
Meeting on Zoom/Teams/Meet
↓
Automatic recording + transcription (Otter / Fireflies / tl;dv)
↓
Summary with AI (ChatGPT / Claude + prompt)
↓
Action items with owners and dates
↓
Distribution (email, Slack, Notion)
Step 1: Record and Transcribe Automatically
Free option: Otter.ai
- Create account at otter.ai (free, 300 min/month)
- Connect your Google or Outlook calendar
- Enable "Auto-join meetings": Otter will automatically join all your Zoom/Meet calls
- After the meeting, you will have the transcript ready in 5-10 minutes
- Download as TXT or copy the full text
Free alternative: if you use Zoom Business or Teams with Office 365 Plan, both platforms have built-in automatic transcription you can enable in settings.
Alternative option: Fireflies.ai
- Create account at fireflies.ai
- Add the Fireflies bot to your calendar
- The bot automatically joins meetings on Zoom, Teams, Meet, and Webex
- After the meeting: transcript + automatic summary in your dashboard
The advantage of Fireflies over Otter is that it generates a basic automatic summary. The disadvantage: transcription is not real-time.
Step 2: Summarize with ChatGPT or Claude
Once you have the transcript, the next step is converting it into an actionable summary. AI models like ChatGPT or Claude are excellent for this.
Base prompt for summarizing meetings
You are an expert meeting assistant. Analyze the following transcript and generate:
1. EXECUTIVE SUMMARY (3-5 sentences)
2. MAIN TOPICS DISCUSSED (bulleted list)
3. DECISIONS MADE (bulleted list with context on why)
4. ACTION ITEMS (table: Owner | Task | Deadline)
5. UNRESOLVED QUESTIONS (list)
6. NEXT MEETING: date/time mentioned if any
Transcript:
[PASTE TRANSCRIPT HERE]
Advanced prompt for sales teams
Analyze this sales call transcript and extract:
1. CUSTOMER INFORMATION:
- Main problem they mentioned
- Current solution they are using
- Objections raised
- Buying signals detected
2. NEXT STEPS AGREED
3. CRM INFORMATION:
- Deal status: [prospect/interested/proposal/closing]
- Next contact date agreed
- Budget mentioned (if applicable)
Transcript:
[PASTE TRANSCRIPT HERE]
Prompt for project/team meetings
Summarize this team meeting in a format that can be sent via email.
Structure:
- One context sentence (what the meeting was about)
- What was decided (bullets, maximum 5)
- Who does what (table with name and task)
- When we meet next
Tone: direct and professional. No internal jargon.
Transcript:
[PASTE TRANSCRIPT HERE]
Step 3: Tools That Automate the Full Flow
If you prefer not to do step 2 manually, these tools integrate transcription + summary in a single flow:
Fireflies.ai (Pro, $18/month)
Automatically generates a structured summary at the end of each meeting. Not as detailed as a well-crafted prompt with Claude, but fully automatic and requires no action on your part.
tl;dv (Free / Pro $29/month)
In addition to the summary, tl;dv allows creating clips of the most important moments. Ideal for sharing with people who did not attend and need to catch up without watching the full video.
Notion AI (add-on $10/month)
If you use Notion for meeting notes, Notion AI can summarize the notes directly within your workspace.
Step 4: Distribute the Action Items
The summary has no value if it stays in the organizer's email. The final step is distributing it so responsible parties receive what they need to act on.
Distribution Options
Direct email: Copy the AI-generated summary, review it (30 seconds to verify no errors), and send it to the meeting's channel.
Slack/Teams: Many teams prefer to receive the summary in the project channel. You can automate this with Zapier or Make:
- Fireflies finishes the meeting → generates summary → Zapier posts it to the project's Slack channel
Notion/Confluence: If you have a meeting notes page per project, the summary goes directly there. With Notion AI you can even have the summary generated automatically.
Task management: Action items can automatically become tasks in Asana, Linear, or Jira through Fireflies integrations or via Zapier.
Comparison: Free vs Paid Options
| Solution | Cost | Automation | Summary quality |
|---|---|---|---|
| Otter Free + ChatGPT Free | $0 | Manual | High (with good prompt) |
| Fireflies Pro | $18/month | High | Good (automatic) |
| tl;dv Pro | $29/month | High | Very good + clips |
| Otter Pro + Claude | $17 + $20 = $37 | Semi-manual | Very high |
The Otter Free + ChatGPT Free option with the prompts in this guide produces very high-quality results at no cost. It requires 5-10 minutes of manual work post-meeting, but for most teams that is acceptable.
Summary Email Template
Once you have the generated summary, this email format works best:
Subject: Meeting Summary [Project/Topic] — [Date]
Hi team,
Summary from today's meeting:
DECIDED:
• [Decision 1]
• [Decision 2]
TASKS:
• [Name]: [task] — by [date]
• [Name]: [task] — by [date]
NEXT MEETING: [date and time, if applicable]
Full transcript: [link to Otter/Fireflies if you want to add it]
Conclusion
The complete AI meeting management workflow can be set up in under an hour: configure Otter or Fireflies, test the prompts with your next meeting, and establish the distribution channel. From that point on, every meeting automatically generates its own documentation.
The time you save on note-taking, transcribing, and distributing goes toward real work. That is what makes AI genuinely useful.
Ir a la herramienta Ir a la herramienta