AI Meeting Follow-Up Brief

Problem

Meeting follow-ups often come from scattered bullets, partial transcripts, or chat context. The hard part is separating real commitments from discussion: who owns what, by when, what was explicitly decided, and what still needs confirmation. Relying on memory can turn ambiguous statements into decisions or miss owners and deadlines.

How it works

This skill is a prompt-only documentation workflow. It does not join meetings, access calendars, send messages, or read private transcripts. Given meeting notes, transcript excerpts, or chat logs, it processes the text in a structured way:
- Assess source quality: note whether the input is complete, partial, or fragmented.
- Extract confirmed decisions: keep only decisions explicitly supported by the notes, with context where useful.
- Build action items: capture owner, deadline, dependency, and status, marking ambiguous items as needs confirmation.
- Separate open questions and risks: list unresolved topics, blockers, conflicting statements, and assumptions separately.
- Produce a brief and recap draft: include meeting snapshot, decisions, action items, open questions, risks, items needing confirmation, and a verification checklist.

Boundaries

It does not send recaps, contact participants, or verify facts outside the provided text. If the input contains confidential, regulated, or third-party sensitive material, confirm processing permission first. For legal, medical, HR, financial, or safety-critical meetings, recommend human review before acting.