AI converts your raw meeting notes into a clean debrief summary with every action item assigned to a named owner and a specific deadline, so nothing discussed in a meeting goes undone.
A complete operational system. Built to run immediately.
Convert raw meeting notes into a structured debrief with action items, owners, and deadlines in under 5 minutes.
Why commitments do not get done after meetings and how AI extracts ownership gaps from raw notes automatically.
Roles and four prerequisites before Phase 1: action item tracker, AI access, note format guide, and team contact list.
Three phases: set up note format and context inputs, build the extraction and debrief workflow, review and calibrate.
Three targets: action item completion rate, debrief turnaround time, and extraction accuracy rate.
Prompt library for every debrief stage, a debrief quality checklist, and an extraction accuracy tracker.
How to fix missed implicit commitments, ambiguous ownership, and debrief summaries nobody reads.
Version history and an extraction improvement log so every calibration is traceable.
Reads raw meeting notes in any format and extracts every explicit and implicit action item with a named owner and deadline.
Classifies every substantive topic from raw notes as Decision, Action Item, Discussion, or Parking Lot with a one-line summary.
Takes the extracted action list and decision log and produces a clean debrief summary formatted for one-click distribution.
Generates one personalized follow-up message per action item owner containing only their assigned items and due dates.
Reviews monthly action item completion data and surfaces the top two extraction or ownership gaps reducing completion rate.
Generates a retrospective note with a completion accuracy score and one specific extraction improvement for next cycle.
A clean debrief summary with every action item in a table: owner, description, deadline, and status at a glance.
One follow-up message per action item owner listing only their assigned items, deadlines, and relevant context.
A persistent record linking this meeting's commitments to the tracker, with completion status updated each cycle.
Six copy-paste-ready prompts covering extraction, classification, debrief summary, owner follow-ups, and calibration.
A configured workflow that sends personalized owner follow-ups automatically within 30 minutes of meeting end.
A persistent record linking every meeting's commitments to the tracker with completion rates by owner and meeting type.
Every meeting produces commitments. Most of them disappear within 24 hours because the recap is vague, the owners are not named, and the deadlines are aspirational. The problem is not that people are unreliable. It is that the debrief system does not give them what they need to follow through. A structured debrief with AI-extracted action items, a clean decision log, and personalized owner messages changes that before the next meeting even starts.
An action item without a named owner is a suggestion. Without ownership, the same item appears in three consecutive meetings.
A recap blending decisions and discussions makes every reader sort the relevant parts themselves. Structure does that automatically.
A four-page debrief sent to six people gets skimmed by none of them. One-page owner-specific follow-ups get read and acted on.
Commitments with no tracking home disappear after the meeting. A structured tracker closes the loop on every item.
Expected Outcomes
These targets come directly from the SOP. Realistic benchmarks, not aspirational claims.
80%+
The primary signal that the debrief system is driving follow-through, not just producing better-formatted records. A rate below 60% for two consecutive months usually points to deadlines that are too aggressive, ownership that is still ambiguous after review, or the action item tracker not being updated within 30 minutes of the meeting ending.
<30 min
Measures whether the debrief is reaching participants while the meeting is still fresh. A summary sent the next day has already lost half its value as commitments and context fade. Track per meeting. Consistent misses over 30 minutes mean the AI prompts need to run before the meeting officially ends, not after.
95%+
Measures whether the action item extractor is catching real commitments rather than only the obvious ones. A rate below 95% for explicit items means commitment signals are not being used consistently in notes. Below 80% for implicit items means the note-taking format needs more specific attribution guidance for statements that create expectations without using the word 'will.'
System live. First 3 meetings debriefed through the full extraction-to-owner-follow-up workflow with completion tracking active for each session.
First optimization version shipped. One note format or prompt improvement based on extraction accuracy review and owner feedback from the first three debriefs.
Stable debrief habit with action item completion rate tracked. Monthly calibration review producing one consistent improvement to note format or AI prompts each cycle.
Common Questions
QuickStart gets your first debrief summary and owner follow-ups generated in about 30 minutes. That includes formatting one set of meeting notes, running AI ASSIST 1, and reviewing before sending.
The action item extractor catches implicit commitments by scanning for agreement language, not just task phrases. If a specific item gets missed, flag it in the calibration log and add the pattern to the note format guide.
Use AI ASSIST 2 to generate a neutral resolution process note. It classifies the item as a Parking Lot topic and produces a two-line owner assignment for the next session without taking a side.
They are paired systems. MTY-0403 produces the action item table and owner follow-ups. MTY-0402 uses the open action item tracker as an input to the next agenda. Together they close the loop between every meeting.
The 6 prompts cover action item extraction, topic classification, debrief summary, owner follow-ups, monthly calibration, and retrospective. Each prompt is copy-paste ready with labeled input placeholders.
