Meeting Transcription
Author/Source: Chris Blattman, claudeblattman.com
Key Ideas¶
- Meeting transcription turns ephemeral decisions, action items, and context into a searchable, actionable record -- solving the problem of institutional memory loss.
- Granola is the author's recommended tool (~$10/month): it transcribes Zoom calls, phone calls, and in-person meetings (phone in pocket), provides AI summaries, and exports raw transcripts.
- The real value comes from feeding transcripts to Claude Code for targeted extraction ("What decisions were made about the survey instrument?") and appending results to project meeting logs.
- Over time, project meeting logs become searchable institutional memory, especially valuable for distributed teams across time zones.
- Granola's main limitation is unreliable speaker identification, which matters for larger meetings.
- Alternatives include Otter.ai (strong speaker ID), tl;dv (good free tier), Zoom built-in transcription (basic but free), Notion AI, and Rev (human transcription).
Summary¶
This article makes the case that meeting transcription is a foundational tool for any professional workflow, independent of whether you use Claude Code or any other AI system. The author recommends Granola for its ability to transcribe Zoom meetings, phone calls, and in-person meetings without requiring a bot to join the call. The tool costs about $10/month and works on Mac, Windows, and iOS.
The article distinguishes between standalone use (transcripts and AI summaries after every meeting) and integrated use with Claude Code (feeding raw transcripts for specific information extraction and project meeting log maintenance). The integrated approach transforms meetings from ephemeral events into a searchable project record. The setup is straightforward (5 minutes), with the real investment being the habit of reviewing and processing transcripts after meetings.
Relevance to Economics Research¶
Meeting transcription is directly applicable to research team management, especially for projects involving multiple collaborators, field teams, or advisory boards. The ability to search across months of meeting transcripts ("When did we decide to drop the third treatment arm?") addresses a chronic problem in multi-year research projects. For economists running field experiments or collaborating across institutions and time zones, having a searchable record of decisions and their rationale prevents relitigating settled questions and helps onboard new team members.