Bi account brief

Account Briefing Generator

The most-used skill in the Book Intelligence pillar. Compiles a structured, comprehensive account snapshot from all available data sources. Designed to replace the 30-60 minutes a CSM spends manually assembling account context before a call, QBR, renewal conversation, or escalation.

This is a data assembly skill. It pulls, structures, and formats. No inference, no judgment, no action. The brief is an input to every human decision point across all pillars.

When to Run

• On-demand: CSM requests a brief before a call, meeting, QBR, renewal conversation, or escalation
• Auto-generated: Triggers ahead of scheduled meetings when linked to calendar integration (generates brief 30 minutes before meeting start)
• Called by other skills: Any skill that surfaces a human decision point can invoke this skill to provide context alongside the alert

Core Execution Logic

Step 1: Pull Latest Data from All Sources

Gather current data from every available source. If a source is unavailable, note the gap in the brief rather than failing entirely.

Source	Data Points
Health Score Engine	Composite score, component scores, trends, risk drivers
Usage Pattern Monitor	Key usage metrics, trend classifications, active pattern alerts
Risk Signal Detector	Active risk signals with evidence and severity
Expansion Signal Detector	Active expansion signals with evidence and timing
CRM - Contacts	Key contacts, roles, last engagement date, coverage gaps
CRM - Activities	Last 5 touchpoints across all channels (email, call, meeting)
CRM - Opportunities	Open opportunities, renewal dates, pipeline status
CRM - Notes	Last 3 CSM notes on the account
Support Platform	Open tickets (count, severity, age), recent resolved tickets
Product Analytics	Current key metrics snapshot (DAU/MAU, feature adoption, licence utilisation)
Contract/Billing	ARR, contract end date, payment status, pricing tier
Call Transcripts	Summary of last call (if transcript integration available)

Step 2: Assemble into Structured Brief

Organise data into a scannable format optimised for quick consumption:

Header Block (glanceable in 5 seconds):

• Account name
• ARR
• Health score + trend arrow
• Days to renewal
• Segment (tier / size / industry)
• CSM name

Health Summary (15 seconds):

• Composite score with component breakdown
• Trend direction (improving/stable/declining)
• Active risk drivers (if any)

Usage Summary (15 seconds):

• Key metrics vs. segment benchmark
• Overall usage trend
• Active pattern alerts (if any)

Active Signals (15 seconds):

• Risk signals with severity and evidence (from Risk Signal Detector)
• Expansion signals with timing window (from Expansion Signal Detector)
• Competitive intelligence flags (from Competitive Intel Monitor)

Stakeholder Map (15 seconds):

• Key contacts: name, role, last engagement date
• Coverage gaps: contacts not touched in 60+ days
• Stakeholder changes: recent departures or role changes

Recent Activity (15 seconds):

• Last 5 touchpoints across channels, most recent first
• Last call summary (if available)
• Last CSM notes (3 most recent)

Open Items (10 seconds):

• Open support tickets (count, top severity, oldest)
• Open opportunities or renewal status
• Pending action items from last interaction

Attention Flags (always at the bottom, highlighted):

• Anything requiring decision or action, sorted by urgency
• Data gaps or stale sources that affect brief reliability

Step 3: Format for Context

Adapt the brief based on the consumption context:

Context	Format Adaptation
Pre-call prep	Emphasise recent activity, open items, and stakeholder last-touch. CSM needs conversation context
QBR preparation	Emphasise usage metrics, health trends, and value delivered. CSM needs the story arc
Renewal preparation	Emphasise commercial data, health trajectory, risk signals, and stakeholder map. CSM needs negotiation context
Escalation	Emphasise support history, risk signals, and timeline of deterioration. CSM needs the evidence chain
Executive briefing	Condense to header + health summary + key risk/opportunity. Executive needs the headline, not the detail
Manager review	Full brief with all sections. Manager needs the complete picture

If context is not specified, default to the full format.

Output Format

Structured account brief:

{
  "account_id": "string",
  "generated_at": "2026-03-10T09:30:00Z",
  "context": "pre_call",
  "header": {
    "account_name": "Acme Corp",
    "arr": 85000,
    "health_score": 72,
    "health_trend": "declining",
    "days_to_renewal": 67,
    "segment": "Mid-Market / Medium / Established",
    "csm": "Jane Doe"
  },
  "health_summary": {
    "composite": 72,
    "components": {
      "usage": { "score": 81, "trend": "stable" },
      "engagement": { "score": 65, "trend": "declining" },
      "support": { "score": 58, "trend": "declining" },
      "sentiment": { "score": 80, "trend": "stable" },
      "commercial": { "score": 75, "trend": "improving" }
    },
    "risk_drivers": ["Support: P1 ticket open 12 days", "Engagement: no exec contact in 74 days"]
  },
  "usage_summary": {
    "dau_mau_ratio": { "value": 0.42, "vs_benchmark": "above_median", "trend": "stable" },
    "feature_adoption": { "value": 0.65, "vs_benchmark": "above_75th", "trend": "stable" },
    "licence_utilisation": { "value": 0.78, "vs_benchmark": "above_median", "trend": "growing" },
    "active_alerts": []
  },
  "active_signals": {
    "risk": [
      { "signal": "P1 ticket open", "severity": 4, "evidence": "Ticket #4521: API latency issue, open 12 days", "first_detected": "2026-02-26" }
    ],
    "expansion": [],
    "competitive": []
  },
  "stakeholder_map": {
    "key_contacts": [
      { "name": "Tom Chen", "role": "VP Engineering", "type": "champion", "last_touch": "2026-02-28", "days_since_touch": 10 },
      { "name": "Sarah Kim", "role": "CFO", "type": "economic_buyer", "last_touch": "2025-12-15", "days_since_touch": 85 }
    ],
    "coverage_gaps": ["Sarah Kim (CFO): 85 days since last touch"],
    "recent_changes": []
  },
  "recent_activity": [
    { "date": "2026-02-28", "type": "meeting", "summary": "Weekly sync with Tom Chen. Discussed API performance concerns" },
    { "date": "2026-02-25", "type": "email", "summary": "Follow-up on ticket #4521 with support team CC" },
    { "date": "2026-02-20", "type": "call", "summary": "QBR planning call with Tom Chen" }
  ],
  "open_items": {
    "support_tickets": { "count": 2, "highest_severity": "P1", "oldest_days": 12 },
    "renewal_status": "67 days out, no active renewal opportunity in CRM",
    "pending_actions": ["Follow up on API latency resolution", "Schedule QBR for March"]
  },
  "attention_flags": [
    { "flag": "P1 ticket open 12 days -- resolution needed before QBR", "urgency": "high" },
    { "flag": "CFO last touched 85 days ago -- coverage gap ahead of renewal", "urgency": "medium" }
  ],
  "data_gaps": []
}

Handoff to Human

This skill is a support function -- it generates context for human consumption. No decision is required from the agent. The brief feeds into every human decision point:

• Before calls: CSM reads the brief to prepare conversation points
• Risk triage: CSM uses the brief alongside risk alerts to understand the full picture
• Expansion review: CSM uses the brief to assess whether the account is ready for a commercial conversation
• Escalation: CSM uses the brief to build the case for internal escalation
• Manager 1:1s: Manager reviews the brief to understand a CSM's account portfolio

Confidence and Limitations

• High confidence for data assembly -- this is structured data retrieval and formatting, not inference
• Confidence degrades when source data is stale or incomplete. The brief explicitly flags gaps rather than presenting incomplete data as complete
• Cannot surface information that exists only in the CSM's head (relationship nuances, unlogged conversations, intuitions)
• The brief is a snapshot. It reflects the state at generation time. For fast-moving situations (active escalation, ongoing negotiation), regenerate before each interaction

Dependencies

Required:

• bi-health-score (health data)
• CRM API (contacts, activities, opportunities, notes)

Strongly recommended:

• bi-usage-monitor (usage profile)
• bi-risk-detector (active risk signals)
• bi-expansion-detector (active expansion signals)
• Support platform API (ticket data)
• Product analytics (usage snapshot)
• Calendar integration (for auto-generation ahead of meetings)

Optional:

• bi-competitive-intel (competitive flags)
• Call transcript integration (last call summary)
• Email integration (recent email threads)

References

• references/brief-templates.md -- Template variations by context (pre-call, QBR, renewal, escalation, executive)