Use when extracting product signals from meeting transcripts - systematically identifies asks, problems, quotes, and evidence across time windows with customer attribution
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Um comando direto ignora o prompt de revisão. Verifique a origem antes de executá-lo.
Instruções da origem · Visualização somente leitura
name
meeting-synthesis
description
Use when extracting product signals from meeting transcripts - systematically identifies asks, problems, quotes, and evidence across time windows with customer attribution
allowed-tools
Read, Grep, Glob
Meeting Synthesis
Purpose
Extract actionable product signals from meeting transcripts:
Feature requests and capability asks
Pain points and friction areas
Customer quotes and verbatim feedback
Evidence for epic creation and prioritization
Pattern detection across multiple customers
When to Use This Skill
Activate automatically when:
product-planning workflow gathers signals for epics
cs-prep workflow compiles customer context for QBRs
strategy-session workflow needs customer insights
User explicitly requests meeting synthesis
Any workflow requires customer evidence extraction
Signal Types
1. Feature Requests ("Asks")
Explicit requests for new capabilities or enhancements.
Indicators:
"We need..."
"Can you build..."
"It would be great if..."
"Our team wants..."
Extract:
What capability is requested?
Why is it needed? (business impact)
Who requested it? (customer, role)
When? (meeting date)
2. Pain Points ("Problems")
Friction, blockers, or inefficiencies users experience.
Indicators:
"It's frustrating that..."
"We struggle with..."
"The current process is..."
"Our team spends too much time..."
Extract:
What is the pain point?
How does it impact workflow/outcomes?
Frequency/severity?
Who experiences it?
3. Onboarding Friction
Specific difficulties during setup or first value realization.
Indicators:
"Setup was confusing..."
"Took us X weeks to..."
"Couldn't figure out how to..."
"Documentation unclear on..."
Extract:
What step caused friction?
How long did it take?
Was it eventually resolved?
Suggestions for improvement?
4. Integration Gaps
Missing connectors or data sync capabilities.
Indicators:
"We use [Tool X] but can't connect..."
"Data sync with [Platform Y]..."
"Export to [System Z]..."
Extract:
What integration is needed?
What data needs to flow?
Current workaround (if any)?
Business impact of gap?
5. Performance Needs
Speed, scale, or reliability requirements.
Indicators:
"Too slow when..."
"Times out on..."
"Can't handle X records..."
"Need real-time..."
Extract:
What operation is slow?
Current performance vs. needed?
Scale requirements?
Impact on usage?
Synthesis Process
1. Determine Time Window
Inputs:
days: Explicit lookback window (e.g., 7, 14, 30)
last_run: Timestamp from state file (e.g., datasets/product/.meetings-to-backlog-state.json)
Default: 3 days if no inputs provided
Calculate cutoff date:
If days provided:
cutoff_date = current_date - days
Else if last_run exists:
cutoff_date = last_run timestamp
Else:
cutoff_date = current_date - 3 days
2. Collect Meeting Files
Scan paths:
datasets/meetings/Customers/**/YYYY/*.md (where date >= cutoff_date)
datasets/meetings/Internal/Product/YYYY/*.md (where date >= cutoff_date)
{
"cluster_label": "Action-style theme name",
"mention_count": N, # Total signal occurrences
"unique_accounts": N, # Distinct customers
"unique_functions": N, # Customer vs. internal sources
"signals": [array of signal objects],
"recentness_weight": calculated_score # Newer = higher
}
5. Compute Signal Metrics
For each cluster:
A. Mention count:
Total occurrences across all meetings.
B. Unique accounts:
Distinct customer names in signal sources.
C. Unique functions:
Count of different source types:
External customers
Internal product team
Partner/agency feedback
D. Recentness weight:
For each signal in cluster:
days_ago = current_date - signal.date
recentness_score = max(0, 30 - days_ago) / 30 # Linear decay over 30 days
Average recentness across all signals in cluster
E. Source diversity:
Mix of meeting types:
Sales calls
Product feedback sessions
Customer success check-ins
Onboarding calls
Higher diversity = stronger signal.
6. Output Synthesized Data
Return structured object:
{
"time_window": {
"start_date": "YYYY-MM-DD",
"end_date": "YYYY-MM-DD",
"days_scanned": N,
"meetings_processed": N
},
"clusters": [
{
"label": "Cluster name",
"mention_count": N,
"unique_accounts": N,
"unique_functions": N,
"recentness_weight": 0.0-1.0,
"source_diversity": 0.0-1.0,
"signals": [
{
"text": "Quote or paraphrase",
"customer": "Company",
"date": "YYYY-MM-DD",
"meeting_type": "...",
"source_file": "..."
}
]
}
],
"raw_signals": [all unclustered signals for reference]
}
Filtering and Scoping
Customer Filter
Input:include_customers (comma-separated list)
Behavior:
If provided: Only process meetings where customer field matches list
If not provided: Process all customers
Example:
include_customers: "PrettyBoy,CompoundStudio,Joy Organics"
→ Only synthesize signals from these three customers
