All use cases
Meeting Note Summarization
Auto-extract summaries, decisions, and action items from every meeting. Zero meetings without notes.
The problem
Product team running 12+ meetings per week. Notes were inconsistent — some people took them, most didn't. Action items got lost between Slack threads and half-written docs. Decisions made in meetings had no paper trail.
The result
Zero meetings without documented notes. Action item completion rate jumped from 40% to 78%. New team members can now read meeting history to get up to speed.
The workflow
How the system actually runs.
- 01 Meeting recorded and transcribed via Otter.ai
- 02 Transcript sent to Claude via Make automation
- 03 Claude extracts: 3-line summary, key decisions, action items with owners, open questions
- 04 Structured output pushed to Notion meeting log as a new entry
- 05 Action items auto-created as tasks in Notion with assignees
Tools used
The stack behind it.
Otter.aiClaude APINotion APIMake
What worked
Why this setup held up.
- Structured JSON output format made Notion integration clean and reliable
- Forcing a "decisions made" section in every summary changed how people ran meetings
- Weekly digest of unresolved action items created accountability without nagging
What did not
The friction to watch.
- Otter transcription errors caused bad summaries for highly technical discussions
- Had to switch to manual transcript upload for critical architecture and planning meetings
- Some meetings had no clear decisions — AI would infer decisions that weren't actually made
Verdict
The short version.
Changed how our team operates. The real win isn't the AI summary — it's the forced structure that makes meetings actually productive.