The official debrief is for the team. The subtext is for you—honest feedback on what you could do better, patterns you’re repeating, and political dynamics you should be aware of. The Meeting Debrief is for the team—shareable, DSGVO-compliant, focused on decisions and actions. The Personal Subtext is for you—brutally honest feedback that helps you grow.

Insight
It’s okay if AI destroys you in the subtext. That’s the point. You want the unfiltered feedback that helps you grow—just not in the official record.

Why a Separate Layer?

Official Meeting Debrief:

  • Shared with team
  • DSGVO-compliant
  • Focus on decisions + actions
  • Neutral, blameless

Personal Subtext:

  • Only for you
  • Brutally honest
  • Focus on your improvement
  • Can include uncomfortable truths

Note
Coming Soon: Personal Context for AI
I’ll write about how to give AI personal context so it knows you better—your patterns, your blind spots, your goals. The better AI knows you, the more personalized (and useful) the feedback becomes.

What Goes in the Subtext

1. Your Own Manöverkritik

Questions AI answers about YOU:

  • What could you have done better?
  • Did you dominate the discussion?
  • Did you listen enough?
  • Were you prepared?
  • Did you enable or block decisions?

Example output:

  • Du hast 3x unterbrochen bevor andere ausreden konnten
  • Topic X hast du zu schnell abgehandelt — Team hatte noch Fragen
  • Gute Moderation bei der Entscheidung zu Y
  • Vorbereitung auf Z war dünn — nächstes Mal Zahlen parat haben

2. Pattern Recognition

Your recurring behaviors:

🔥 Firefighter Mode detected:

  • Du hast das Deployment-Problem selbst gelöst
  • Wer hätte es lernen können? → DevOps-Team

📥 Desk-Vortex:

  • Task X ist auf deinem Schreibtisch gelandet
  • Eigentlicher Owner: [Team/Person]

⚡ Speed Mismatch:

  • Du warst 3 Schritte voraus
  • Team brauchte mehr Context

3. Political Awareness

Things you should know but shouldn’t document officially:

  • [Person A] war auffällig still bei Topic X — möglicherweise Widerstand?
  • Dynamik zwischen [Team 1] und [Team 2] angespannt
  • Entscheidung Y wurde “durchgewunken” — prüfen ob wirklich Konsens

Important: This is YOUR awareness layer. Not for sharing, not for HR, not for official systems.

4. Self-Improvement Actions

  • Nächstes Mal: Mehr Raum für Fragen lassen
  • Topic X mit [Person] separat besprechen
  • Vorbereitung: Zahlen zu Y recherchieren
  • Pattern beobachten: Unterbreche ich zu oft?

Implementation

Two-Output Strategy

Phase 3a: Official Debrief
├── Entscheidungen
├── Action Items
├── Insights (anonymisiert)
└── Meeting Effectiveness Review

Phase 3b: Personal Subtext (GETRENNT)
├── Manöverkritik (dein Verhalten)
├── Pattern Alerts (Firefighter, Desk-Vortex)
├── Political Awareness (privat)
└── Self-Improvement Actions

Prompt for Subtext Generation

Note
This is a high-level example. Adapt it to your context, your patterns, and what actually helps you. You can add personal context (your known blind spots, your goals, patterns you want to watch) to make it more specific.

## Personal Subtext (NUR FÜR MICH)

Analysiere MEIN Verhalten in diesem Meeting:

1. **Manöverkritik**
   - Was hätte ich besser machen können?
   - Wo habe ich zu viel/zu wenig geredet?
   - War ich vorbereitet?

2. **Pattern Check**
   - Firefighter Mode? (Habe ich selbst gelöst statt delegiert?)
   - Desk-Vortex? (Habe ich Tasks übernommen die anderen gehören?)
   - Speed Mismatch? (War ich zu schnell für das Team?)

3. **Political Awareness**
   - Welche Dynamiken sollte ich beachten?
   - Wo könnte Widerstand entstehen?
   - Was wurde NICHT gesagt?

4. **Actions für mich**
   - Konkrete Verbesserungen
   - Follow-ups die ich machen sollte
   - Patterns die ich beobachten will

Sei direkt. Das ist nur für mich.

The Value of Honest Feedback

Why AI for Self-Critique?

Problem with self-reflection:

  • We have blind spots
  • We rationalize our behavior
  • We forget specifics quickly

Problem with asking others:

  • Political consequences
  • People soften feedback
  • Asymmetric power dynamics

AI advantage:

  • No consequences for honesty
  • Consistent criteria
  • Immediate after meeting
  • Based on actual transcript, not memory

Example: Uncomfortable but Valuable

Insight
When AI Feedback Hurts (In a Good Way)

Du hast das Meeting dominiert:
• 62% Sprechanteil bei 5 Teilnehmern
• 4x Ideen anderer “verbessert” bevor sie ausreden konnten
• Entscheidung zu Y praktisch alleine getroffen

Das mag effizient sein, aber: Team ownership sinkt, du wirst zum Bottleneck, andere entwickeln sich nicht.

Suggestion: Nächstes Meeting bewusst 30% max Sprechanteil anpeilen.

This kind of feedback is gold. But you’d never put it in official notes.

Privacy & Storage

Rules for Subtext

  1. Separate files — Never in the same document as official debrief
  2. Local only — Don’t sync to shared systems
  3. Regular cleanup — Delete old subtext after learnings internalized
  4. Never reference — Don’t cite subtext in official communications

Suggested Naming

work/meetings/
├── 25•330-daily-standup.md          # Official (shareable)
└── 25•330-daily-standup-subtext.md  # Private (never share)

Sources

  • Military debrief tradition: After-action review methodology
  • Pattern framework: Firefighter Mode, Desk-Vortex, Speed Mismatch
  • Self-improvement research: Immediate feedback effectiveness

Deep Dives

01

Example: Meeting Debrief + Subtext

Full example of a Meeting Debrief (shareable) and Personal Subtext (private) from a real onboarding session.

02

Meeting Effectiveness Review: Neutral Feedback for Teams

When a neutral tool says 'this wasn't a Daily'—that lands differently than when a person says it. AI becomes the honest mirror your team needs.

03 Here

Personal Subtext: Private Manöverkritik

The official debrief is for the team. The subtext is for you—honest feedback on what you could do better, patterns you're repeating, and things you need to know.

04

DSGVO-Compliant Meeting Analysis

You can analyze meetings with AI—but not people. Focus on structure, decisions, and format. Never on individuals.