Customer Meeting Intelligence

Customer Meeting Intelligence System

A comprehensive meeting intelligence platform that transforms reactive meeting notes into proactive customer engagement through automated briefings, context gathering, and intelligent follow-up workflows.

Core Features

🌅 Pre-Call Intelligence (8:30 AM daily)

• Calendar scanning identifies customer meetings for the day
• Account context gathering from billing, support, and communication history
• Customer intelligence includes spend trends, recent escalations, previous meeting notes
• Proactive briefings posted to Slack DM before each customer call
• Thread continuity links pre-call brief with post-meeting follow-up

📝 Post-Meeting Processing (3:33 PM daily)

• Action item extraction from Fellow.ai meetings and Gmail Gemini notes
• Draft approval workflow prevents accidental posting with review process
• Enhanced context enriches meeting notes with customer account health
• Smart filtering only processes meetings where you were a participant
• Thread linking connects back to morning pre-call brief

🔄 Draft Review & Approval

• Slack draft generation with meeting summary and action items
• Approval reactions (✅ approve, ✏️ edit, ❌ skip) for controlled posting
• Context preservation maintains account intelligence across meeting lifecycle
• Team coordination splits action items by assignee for clear ownership

Quick Start

Morning Pre-Call Briefings

# Generate today's pre-call briefings
bash scripts/pre-call-brief.sh

# Generate briefings for specific date  
bash scripts/pre-call-brief.sh --date 2026-02-21

# Dry run — see what would be generated
bash scripts/pre-call-brief.sh --dry-run

Post-Meeting Action Items

# Process last 24h of meetings with enhanced context
bash scripts/meeting-next-steps.sh --slack

# Check specific time period
bash scripts/meeting-next-steps.sh --days 3

# Generate draft for approval workflow
bash scripts/meeting-next-steps.sh --output meeting-data.json
bash scripts/draft-approval.sh meeting-data.json

Draft Approval Workflow

# Generate and post draft for review
bash scripts/draft-approval.sh meeting-data.json

# Post to specific customer channel after approval
bash scripts/draft-approval.sh meeting-data.json "#customer-channel"

# Test draft format without posting
DRY_RUN=true bash scripts/draft-approval.sh meeting-data.json

Configuration

Copy assets/config-template.json to your working directory as config/config.json and customize:

{
  "fellow": {
    "api_key": "your-fellow-api-key",
    "base_url": "https://api.fellow.app"
  },
  "filters": {
    "attendee_email": "your-email@company.com",
    "skip_keywords": ["1:1", "internal", "standup"]
  },
  "slack": {
    "dm_channel": "your-dm-channel-id",
    "bot_token_env": "SLACK_BOT_TOKEN"
  },
  "integrations": {
    "billing_agent_url": "http://your-billing-agent:8000/a2a/billing-account/rpc",
    "escalation_tracker": "~/memory/escalation-router-tracker.json",
    "salesforce_enabled": true
  }
}

Environment Variables

Customer Intelligence Sources

Account Context

• Billing data via A2A billing agent (spend trends, credit status, payment history)
• Support tickets from escalation tracking and ENGDESK systems
• Usage patterns from customer spend monitoring dashboards
• Relationship status from previous meeting notes and communication history

Communication History

• Gmail integration via gog CLI for recent email threads
• Fellow.ai meetings for previous call context and outcomes
• Slack conversations for ongoing support discussions
• Calendar analysis for meeting frequency and engagement patterns

Opportunity Intelligence

• Salesforce integration for open opportunities and pipeline status
• Renewal tracking for upcoming contract discussions
• Expansion signals from usage growth and feature requests
• Competitive intelligence from support tickets and meeting notes

Advanced Features

Smart Customer Detection

Automatically identifies customer meetings by:

• External attendees not from your company domain
• Customer keywords in meeting titles (configurable)
• Domain mapping for known customer domains
• VIP flagging for strategic accounts requiring extra context

Context Enrichment

Enhances meeting data with:

• Recent escalations from support ticket tracking (last 7 days)
• Billing alerts including credit risk and payment issues
• Usage anomalies from spend monitoring systems
• Previous commitments from historical meeting notes
• Account health scores from multiple data sources

Thread Continuity

Maintains conversation flow by:

• Morning brief threads that connect to post-meeting notes
• Customer history linking all meetings for the same account
• Action item tracking across multiple meetings and time periods
• Commitment monitoring ensuring follow-through on promises made

Integration Points

Required Tools

• gog CLI for Gmail/Calendar access (/opt/homebrew/bin/gog)
• Fellow.ai API for meeting transcripts and action items
• Slack API for posting briefs and collecting approvals
• curl/jq for API interactions and JSON processing

Optional Integrations

• Billing A2A agent for customer account financial health
• Escalation tracker for recent support ticket activity
• Salesforce API for opportunity and renewal context
• Memory files for historical meeting notes and customer insights

Best Practices

Customer Names & Privacy

• Internal communications (Slack) use actual customer names
• External communications use "partner carrier", "losing carrier", etc.
• Thread subjects include customer name for easy identification
• Context sharing respects customer confidentiality boundaries

Meeting Efficiency

• Pre-call preparation reduces time spent gathering context during calls
• Action item clarity ensures clear ownership and follow-through
• Draft review prevents miscommunication and incomplete notes
• Historical context builds on previous meeting outcomes

Account Management

• Relationship tracking maintains continuity across team members
• Escalation awareness surfaces recent issues before they become problems
• Opportunity recognition identifies upsell and expansion possibilities
• Risk mitigation flags accounts with declining engagement or usage

Troubleshooting

Common Issues

• No meetings found: Check calendar permissions and date filters
• Missing context: Verify API keys and integration endpoints
• Draft not posted: Confirm Slack bot permissions and channel access
• Customer not detected: Review domain mapping and detection rules

Error Recovery

• Partial context: System continues with available data sources
• API failures: Implements retry logic with exponential backoff
• Missing participants: Filters ensure only relevant meetings are processed
• Channel errors: Graceful fallback to DM delivery

File Structure

scripts/
├── pre-call-brief.sh       # Morning customer meeting preparation
├── meeting-next-steps.sh   # Post-meeting action item extraction  
├── draft-approval.sh       # Review workflow for meeting notes
└── format-actions.sh       # Slack formatting utilities

assets/
└── config-template.json    # Configuration template

references/
└── (dynamically populated with customer-specific context)

Automation Schedule

Set up these cron jobs for full automation:

Morning Pre-Call Briefings:

30 8 * * 1-5  # 8:30 AM CST, Monday-Friday

Post-Meeting Action Items:

33 15 * * 1-5  # 3:33 PM CST, Monday-Friday

Both can run as isolated sub-agent sessions with automatic delivery to your Slack DM for review and action.