Meeting Insights Analyzer

Transform meeting transcripts into concrete, evidence-backed feedback on communication patterns, leadership behaviors, and interpersonal dynamics.

Core Workflow

1. Ingest & Inventory

Scan the target directory for transcript files (.txt, .md, .vtt, .srt, .docx, .json).

For each file:

• Extract meeting date from filename or content (expect YYYY-MM-DD prefix or embedded timestamps)
• Identify speaker labels (patterns: Speaker 1:, [John]:, VTT/SRT cue formatting)
• Detect the user's identity: ask if ambiguous, otherwise infer from most frequent speaker
• Log: filename, date, duration, participant count, word count

Print a brief inventory table so the user confirms scope before heavy analysis begins.

2. Normalize Transcripts

Normalize into common internal structure: { speaker, timestamp_sec, text }[]

• VTT/SRT: Parse cue timestamps + text. Speaker labels may be inline or prefixed.
• Plain text: Look for Name: or [Name] prefixes. Warn if no speaker labels.
• Markdown: Strip formatting, treat as plain text.
• DOCX/JSON: Extract text content, treat accordingly.

If timestamps are missing, skip timing-dependent metrics but run text-based analysis.

3. Analyze

Run all applicable modules below. Each is independent -- skip any that don't apply.

See references/analysis-modules.md for detailed module specifications.

Module: Speaking Dynamics

Per-speaker: word count %, turn count, average turn length, longest monologue, interruption detection. Red flags: >60% speaking in 1:many meeting, <15% when facilitating, one participant never speaks, interruption ratio >2:1.

Module: Conflict & Directness

Scan for hedging language (qualifiers, permission-seeking, deflection, softeners) and conflict avoidance patterns (topic changes after tension, agreement-without-commitment). For each instance: full quote with context, severity tag (low/medium/high), rewrite suggestion.

Module: Filler Words & Verbal Habits

Count: "um", "uh", "like" (non-comparative), "you know", "actually", "basically", etc. Report rate per 100 words. Only flag if rate exceeds ~3 per 100 words.

Module: Question Quality & Listening

Classify questions: closed, leading, open genuine, clarifying, building. Good listening: clarifying/building questions, paraphrasing, referencing prior points, asking quieter participants.

Module: Facilitation & Decision-Making

Only when user is facilitator. Evaluate: agenda adherence, time management, inclusion, decision clarity, action item assignment, parking lot discipline.

Module: Sentiment & Energy

Track emotional arc: positive markers (enthusiasm, humor, praise), negative (frustration, sarcasm), neutral. Flag energy drops.

4. Output the Report

Structure as a cohesive report: Top 3 Findings, Detailed Analysis per module, Strengths, Growth Opportunities, Comparison to Previous Period (if prior analysis exists).

5. Follow-Up Options

After delivering: offer deep dive into specific meeting/pattern, 1-page communication cheat sheet, tracking setup for baseline, export as markdown or JSON.

Edge Cases

• No speaker labels: Warn upfront. Run text-level analysis only. Suggest re-export with speaker diarization.
• Very short meetings (<5 min or <500 words): Analyze but caveat that patterns may not be representative.
• Non-English: Filler word dictionaries are English-centric. Note limitation, focus on structural analysis.
• Single meeting: Skip trend/comparison language. Focus findings on that meeting alone.
• User not identified: Ask before proceeding. Don't guess.

Transcript Source Tips

• Zoom: Settings > Recording > enable "Audio transcript". Download .vtt.
• Google Meet: Auto-transcription saves to Google Docs in calendar event's Drive folder.
• Granola: Exports to markdown. Best speaker label quality.
• Otter.ai: Export as .txt or .json from web dashboard.
• Fireflies.ai: Export as .docx or .json.
• Microsoft Teams: Transcripts in meeting chat. Download as .vtt.

Recommend YYYY-MM-DD - Meeting Name.ext naming convention.

Anti-Patterns

Anti-Pattern	Better Approach
Analyzing without speaker labels	Ask user to re-export with speaker identification
Running all modules on a 5-min standup	Auto-detect meeting length and skip irrelevant modules
Presenting raw metrics without context	Compare to norms and show trajectory over time
Single meeting in isolation	Require 3+ meetings for trend-based coaching
Treating speaking time equality as the goal	Weight ratios by meeting type and role
Flagging every hedge word	Distinguish decision meetings from ideation