Spinach AI and Zoom transcript ingestion turned into source-backed summaries, action items, blockers, and a persistent daily/weekly operations dashboard with human review before any downstream action.
Privacy-safe system map. Private records, identifiers, and workspace details are excluded.
System Summary
Built a workflow that turns Spinach AI and Zoom meeting records into source-backed summaries, decisions, scheduling commitments, follow-ups, blockers, and action items, then surfaces them in a persistent daily and weekly operations dashboard.
Business Problem
Meeting decisions and commitments were buried in long transcripts, so follow-ups were missed and blockers were not visible by the next day.
What Czar Studios Built
Transcript ingestion, long-transcript chunking strategy, extraction logic, normalization into a structured dashboard, and human-review gates before Slack or Hermes actions.
Architecture and Data Flow
Complete transcripts are analyzed with overlapping map-reduce for long sources so commitments mentioned in the middle or end are not missed. Outputs are normalized into a React/API/SQLite operations dashboard with owner, status, due date, source, blocker, next action, and proof link.