meddic-call-prep-auto（医疗呼叫准备自动）

<quick_start> Before a call: "call prep Baylor University" → pulls all context → generates MEDDIC brief

Before a demo: "demo prep for University of Michigan, meeting at 2pm" → adds demo-specific flow

From calendar: "prep my next call" → reads next calendar event → auto-identifies company → runs full prep

Trigger phrases:

• "call prep [company]"
• "prep me for [company]"
• "demo prep [company]"
• "discovery prep [company]"
• "meddic brief [company]"
• "prep my next call"
• "get me ready for my [time] meeting" </quick_start>

<success_criteria>

• Complete MEDDIC brief generated in under 60 seconds
• All 6 MEDDIC dimensions populated (even if some are "UNKNOWN — ask this")
• Prior conversation context included from CRM activity history
• Competitive displacement angle identified if relevant
• Attendee roles mapped to MEDDIC roles (champion, economic buyer, etc.)
• 3 personalized discovery questions generated from enrichment data
• Brief fits on one screen (scannable in 30 seconds before call) </success_criteria>

Pipeline

TRIGGER              CONTEXT GATHER              SYNTHESIZE              OUTPUT
───────────────────────────────────────────────────────────────────────────────
"call prep X"  →  Calendar: meeting details  →  Map attendees to   →  MEDDIC Brief
               →  HubSpot: deal + contact    →  MEDDIC roles       →  Talking Points
               →  Apollo: enrichment         →  Identify gaps       →  Questions
               →  CRM: activity history    →  Competitive angle   →  Objection Prep
               →  Epiphan CRM: device check  →  Pain hypothesis     →  Next Steps

Stage 1: Context Gathering

Run these MCP calls in parallel:

1a. Calendar Context

Extract: attendee names + emails, meeting time, any agenda in description.

1b. HubSpot Context

Extract: deal stage, deal amount, lifecycle stage, prior notes, form submissions, page views.

1c. Apollo Enrichment

1d. Clari Conversation History

For each attendee email:

1. clari_search_calls(attendeeEmail=attendee_email, daysBack=90)
2. For up to 3 most recent calls: clari_get_call_summary(callId=call_id)
3. Extract:

- Prior conversation topics: What was discussed last time? - Unresolved action items: What did we promise to follow up on? - Competitor mentions: Were Extron, Blackmagic, Crestron, vMix, Teradek mentioned? - Objections raised: Price, features, integration concerns? - MEDDIC gaps from prior call: Which dimensions were NOT covered?

1e. Prior Conversation Context

Query ask_agent with: "Show recent activity, notes, and engagement history for [company name] in the last 90 days"

Extract: what was discussed, commitments made, objections raised, next steps promised.

1f. Competitive Intelligence Context

Check for stored competitive intel from the bdr-v3-competitive-intel daily task:

1. Search Gmail for recent competitive intel reports: gmail_search_messages(q="subject:competitive-intel from:me newer_than:7d")
2. If company or competitor was mentioned in recent Clari extractions, pull the relevant battlecard context
3. Integrate into MEDDIC brief:

- If competitor mentioned in Clari calls: Include displacement talking points - If no competitor mentioned: Include proactive competitive positioning for top 2 competitors in prospect's vertical - Specific objection rebuttals: Map Clari objections to prepared responses

Competitive Quick Reference (by vertical):

1g. Epiphan CRM Check

Golden Rules check: If ANY of these are true, STOP and skip this prep:

• lifecyclestage = 'customer' → Route to AE/account manager, not Tim's call
• device_count >= 1 → Existing customer, route to account manager
• is_channel = true → Channel partner, route to channel manager
• hubspot_owner_id IN ('82625923', '423155215', '190030668') → AE-owned deal (Lex Evans, Ron Epstein, Phillip Sandler), not Tim's to prep

Exception: If Tim explicitly says 'prep me for [company]' and the company is flagged, generate the brief but add a ⚠️ WARNING banner at the top: 'This is an EXISTING customer/channel partner/AE deal. Coordinate with the owner before reaching out.'

Stage 2: MEDDIC Synthesis

Map gathered context into MEDDIC framework:

M — Metrics

• Known: Any quantifiable goals mentioned in prior calls or deal notes
• Hypothesis: Based on vertical benchmarks (e.g., "Universities your size typically manage 50-200 rooms")
• Ask if unknown: "How are you measuring success for this project?"

E — Economic Buyer

• Known: Map attendees by seniority → VP/Director/Dean = likely EB
• ATL/BTL Validation (MANDATORY — see CLAUDE.md § ATL/BTL Classification v1.0):

- Confirmed EB (ATL tier): Title matches Chief (CIO, CTO, CFO, COO), VP (AVP, SVP, EVP), President, Provost, Vice Provost, Superintendent, Director (of IT, Technology, Facilities, Academic Technology, Procurement, Materials Mgmt, Medical Education, Court Administration), Dean, Court Administrator, Clerk of Court (Federal), City Manager, County Manager, Senior Pastor, Executive Pastor → Mark EB as ✅ IDENTIFIED - Gray Zone — confirm budget authority: Manager (AV/Facilities/IT) — only ATL if reports to Director+ AND has delegated budget >$25K. Ask: "Does [name] control the budget for this purchase, or do they need approval from someone above them?" - NOT an EB (BTL tier): Technician, Specialist, Coordinator, Support, Administrator (Systems/Network/Database), Engineer (AV/Network/Systems), Operator, Instructor/Professor/Faculty, Designer (Learning/Instructional/Graphic), Assistant, Clerk (non-Court Admin), Volunteer, Intern, Student, Resident, Help Desk → Mark EB as ⚠️ NOT IDENTIFIED — need to find the person who signs the PO - NEVER an EB: Warehouse Manager, Network Manager, Systems Administrator, AV Technician, Graphic Design Instructor, Program Administrator, Web Designer, Classroom Support, Lab Coordinator, Maintenance, Building Engineer, Multimedia Services Manager, Video Production Specialist, Streaming Crew → These roles CANNOT approve budget. Do not list them as potential EBs.

• Unknown attendees: Flag for discovery ("Who controls the budget for this?")
• Signal: If only BTL/technical staff on call, EB isn't engaged yet — flag as ⚠️ and add discovery question: "Who would sign off on this purchase?"

D — Decision Criteria

• Known: From prior call notes, form submissions, or deal properties
• Hypothesis: Based on vertical (Higher Ed cares about CMS integration + room scale; Courts care about compliance + reliability)
• Standard criteria: Price, features, CMS integration, support, ease of use

D — Decision Process

• Known: Timeline from deal properties, procurement notes
• Ask if unknown: "What's your evaluation timeline? Who else needs to weigh in?"
• Red flag: No clear process = deal likely stalls

I — Identify Pain

• Known: Pain expressed in prior calls, form submission context
• Hypothesis: Based on trigger signal (hiring = scaling pain; Extron exit = replacement pain)
• Ask if unknown: "What's broken about your current setup?"

C — Champion

• Known: Who initiated contact? Who's been most engaged?
• Test: "If this made sense, would you advocate for it internally?"
• Red flag: No champion identified after 2nd call

Stage 3: Output Format

╔══════════════════════════════════════════════════════════════╗
║  CALL PREP: [Company Name]                                   ║
║  [Date] [Time] | [Meeting Type: Discovery/Demo/Follow-up]    ║
╠══════════════════════════════════════════════════════════════╣

ATTENDEES & ROLES:
┌─────────────────────────────────────────────────────────────┐
│ [Name] — [Title] → MEDDIC Role: [EB/Champion/User/Coach]   │
│   LinkedIn: [url] | Email: [email]                          │
│   Context: [personalization note from enrichment]           │
├─────────────────────────────────────────────────────────────┤
│ [Name 2] — [Title] → MEDDIC Role: [...]                    │
└─────────────────────────────────────────────────────────────┘

DEAL STATUS:
- Stage: [stage] | Amount: [$XX,XXX] | Close: [date]
- Days in stage: [X] | Last activity: [date]

MEDDIC SCORECARD:
┌───────────────────────┬──────────┬──────────────────────────┐
│ Dimension             │ Status   │ Key Intel / Gap          │
├───────────────────────┼──────────┼──────────────────────────┤
│ M — Metrics           │ ✅/⚠️/❌ │ [what we know or need]  │
│ E — Economic Buyer    │ ✅/⚠️/❌ │ [identified or not]     │
│ D — Decision Criteria │ ✅/⚠️/❌ │ [known criteria]        │
│ D — Decision Process  │ ✅/⚠️/❌ │ [timeline/blockers]     │
│ I — Identify Pain     │ ✅/⚠️/❌ │ [stated pain or hypo]   │
│ C — Champion          │ ✅/⚠️/❌ │ [who + engagement]      │
└───────────────────────┴──────────┴──────────────────────────┘

PRIOR CONVERSATION CONTEXT:
[CRM activity summary or "No prior activity found"]
- Key commitments: [...]
- Objections raised: [...]
- Follow-ups promised: [...]

COMPETITIVE INTEL:
🆚 COMPETITOR STATUS
[Detection result from Clari calls or CRM]

If competitor detected:

Competitor detected: [name] (from Clari call on [date]) Displacement angle: [specific talking point] Objection prep: [specific rebuttal to raised objection] Key differentiator: [most relevant Pearl advantage]

If no competitor detected:

No competitor identified — lead with [vertical-specific value prop]

DISCOVERY QUESTIONS (ask these):
1. [Personalized question based on trigger/enrichment]
2. [MEDDIC gap-filling question]
3. [Competitive displacement question]

OBJECTION PREP:
| Likely Objection | Response |
|------------------|----------|
| "We're happy with current setup" | [response] |
| "Budget is tight" | [response] |
| "Need to evaluate others" | [response] |

CALL AGENDA (suggested):
1. [2 min] Rapport + confirm agenda
2. [5 min] Discovery: pain + metrics
3. [10 min] Demo / value prop aligned to pain
4. [3 min] Decision process + next steps

╚══════════════════════════════════════════════════════════════╝

<demo_prep_addon> When trigger is "demo prep" instead of "call prep", append:

Demo Flow (RECAP → AGENDA → SHOW VALUE → SUMMARIZE → NEXT STEPS)

1. RECAP (2 min): "Last time we discussed [prior activity summary]. You mentioned [pain]. Did I capture that right?"
2. AGENDA (1 min): "Today I'll show you [feature aligned to pain], then we'll discuss [decision criteria]."
3. SHOW VALUE (10 min): Demo path aligned to their vertical:

- Higher Ed: Multi-room capture → CMS auto-publish → room scheduling - Courts: Automated recording → compliance archival → failover - Corporate: Meeting room capture → NDI/SRT streaming → cloud management

1. SUMMARIZE (2 min): "You mentioned [metrics] — here's how Pearl delivers [quantified value]."
2. NEXT STEPS (2 min): "What would a pilot look like? Who else should see this?" </demo_prep_addon>

Sibling Skills Referenced

• sales-revenue-skill — MEDDIC framework, objection handling (LAER), demo flow
• prospect-research-to-cadence-skill — Shares enrichment logic, Golden Rules filter
• hubspot-revops-skill — HubSpot query patterns, deal stage definitions

Emit Outcome Sidecar

As the final step, write to ~/.claude/skill-analytics/last-outcome-meddic-call-prep-auto.json:

{"ts":"[UTC ISO8601]","skill":"meddic-call-prep-auto","version":"1.0.0","variant":"default",
 "status":"[success|partial|error]","runtime_ms":[estimated ms from start],
 "metrics":{"calls_prepped":[n],"meddic_fields_filled":[n],"contacts_enriched":[n],"economic_buyers_identified":[n]},
 "error":null,"session_id":"[YYYY-MM-DD]"}

Use status "partial" if some stages failed but results were produced. Use "error" only if no output was generated.