Reliable meeting-notes-to-action-items agents: split extraction from judgment and force citations
AmbassadorSad3889 · reddit · 2026-09-21
The author tried turning messy meeting notes into action items with owners and due dates; the naive version worked only 70% of the time, inventing owners, turning vague remarks into hard tasks, missing buried commitments, and dropping relative dates like "before the next sync."
What helped:
- Two passes: first extract every sentence implying someone will do something (no formatting), then classify each as a real commitment vs. a maybe. Splitting extraction from judgment sharply cut hallucinated owners.
- Forcing the model to quote the exact source line for each action item — no citation, no task — killed most fabricated items.
Owner and due-date inference remain the weak spot; the author is asking whether a single well-scoped prompt or a small chain works better.
More from coding & agent
- Free DiffusionGemma endpoint returns sentence probabilities in ~230ms, OpenAI-compatible — victormustar · 2026-09-21
- Mozilla AI runs a local 30B model end-to-end to open a real bugfix PR, fully offline — mozilla-ai · 2026-09-21
- GPT-6 plans and codes a game, pulls production 3D assets via Hyper3D Rodin MCP — CodeByPoonam · 2026-09-21
- Dev demo: agent codes a game and generates all 3D assets itself via Hyper3D Rodin MCP — CodeByPoonam · 2026-09-21
- mcp-server-dns: Zero-config MCP server for DNS, WHOIS and domain checks — modelcontextprotocol · 2026-09-21
- Each MCP tool passes solo tests, but the agent's full workflow breaks on state passing — Stock-Pumpkin-8859 · 2026-09-21