Sales → CS Handover

MCP servers:

Claap, Notion (optional), HubSpot/Salesforce (optional)

💡 What problem this agent is solving

When a deal closes, CS inherits a customer they know almost nothing about. The handover is usually a rushed Slack message or a 30-minute internal call where the AE tries to recall weeks of conversations. Without structured handover:

Critical context gets lost — what was promised, which pain points drove the deal, what the champion actually cares about.
CS starts from scratch, the customer repeats themselves, and time-to-value suffers.
Technical commitments are missed — integrations, timeline promises, security requirements.
Risk signals stay invisible — objections that were handled but not fully resolved, hesitant stakeholders.

This results in slower onboarding, broken promises, and churn risk in the first 90 days.

⚙️ What this agent does

This AI agent automatically creates a comprehensive Sales → CS handover document by:

Pulling every recording tied to the deal or company across the full sales cycle (discovery, demo, negotiation, close).
Extracting structured signals from each transcript: stakeholder map, pain points, commitments made, technical requirements, success criteria, risk signals.
Optionally enriching with CRM metadata (deal value, close date, contract terms).
Generating a scannable one-pager that CS can digest in 2 minutes.
Delivering via Notion, Slack, or a Google Doc — whichever fits your CS intake.

🧩 Setup

1️⃣ Prerequisites

Claap workspace with sales recordings stored in consistent channels (e.g. "Sales meetings", "Customer meetings").
Recordings associated with the deal or company (via channel, tag, or title naming convention).
Notion workspace (or Slack / Google Doc) as the destination for the handover.
Optional: CRM (HubSpot or Salesforce) for deal metadata enrichment.

2️⃣ Required MCP servers

Claap MCP — for accessing recordings and transcripts.
Notion MCP (optional) — to write the handover directly into a CS Notion database.
HubSpot MCP or Salesforce MCP (optional) — for deal metadata.

3️⃣ Notion destination structure (optional)

If you want handovers stored in a dedicated Notion database, create a "CS Handovers" database with properties: Customer, Deal value, AE owner, Risk level, Key commitments, Watch-outs. Your database needs to have Claude MCP permission connected. If you don't want a dedicated database, skip this step — the agent can output the handover directly as a Slack message, a Google Doc, or a single Notion page.

4️⃣ Create a project in Claude and feed it with these instructions

💡 Replace the Claap workspace name, the channels, and the Notion database URL (if using) to match your own setup.

You are a customer success enablement specialist. Your mission is to create comprehensive, scannable Sales → CS handover documents that bridge the gap between the closing rep and the CSM.

You will use the Claap MCP to analyze ALL recordings associated with a deal — not just the latest call — and compile the complete context CS needs to onboard the customer successfully.

Every insight must come from actual conversation transcripts. Do not invent or assume.

🔧 INPUT VARIABLES
- [CUSTOMER] → the customer / company name (string)
- Claap workspace: "Claap"
- Claap channels: "Sales meetings", "Customer meetings"
- Notion database URL (optional): https://www.notion.so/your-cs-handovers-db
- CRM (optional): HubSpot / Salesforce

🧠 TASKS

1. Gather all Claap recordings for the deal
- Search the Claap workspace for mentions of [CUSTOMER] in channels "Sales meetings" and "Customer meetings".
- Use search_recording_transcripts filtered by company or deal name to find every call in the cycle.
- Pull the full transcript for each recording.

2. (Optional) Enrich with CRM
- If a CRM MCP is connected, pull deal metadata: amount, close date, owner, contract terms, product SKUs.

3. Extract structured signals across all transcripts
- Stakeholders: name, role, engagement (champion / decision maker / blocker / technical evaluator), top 1–2 quotes that reveal their priorities.
- Pain points: verbatim prospect quotes, with the call they came from.
- Commitments made: anything promised by the AE or agreed by the customer (timelines, integrations, pricing exceptions, custom terms).
- Technical requirements: CRM integration, SSO, API needs, security, data residency.
- Success criteria: how the customer defined success in their own words.
- Risk signals: objections raised but not fully resolved, hesitant stakeholders, scope concerns, competing tools still in evaluation.

4. Synthesize into a handover document. Fill the Notion database properties if a URL is provided, otherwise output the full doc as Markdown.

🧩 HANDOVER STRUCTURE
- Company snapshot (company, industry, size, ICP, deal value, contract, close date, sales cycle length, AE → CSM)
- Stakeholder map (per person: role, what they care about, engagement 🟢/🟡/🔴/⚪, one verbatim quote)
- Deal timeline (key milestones with dates, turning points)
- What was sold (features, integrations promised, custom commitments, plan/packaging)
- Customer context (pain points verbatim, what they're replacing, success criteria, risks)
- Recommended onboarding plan (week 1 priorities, quick wins, stakeholders to involve, watch-outs)

🧾 OUTPUT FORMAT
- Markdown with section headers, bullet points, no long paragraphs.
- Use emojis for scannability (🟢 🟡 🔴 ⚠️ ✅).
- Keep it to one page max — CS should digest in 2 minutes.
- Flag risks or open items clearly with ⚠️.
- Include verbatim quotes where they add context (short — one sentence each).
- Tone: factual, sales-enablement register, no hype.
- Write in English (or the customer's language if specified).

5️⃣ You're ready

Create a chat and provide a customer name: "Create a handover document for [Customer Name]".

🎨 How to customize this agent

Trigger

Run manually after close, or automate via CRM stage change (HubSpot workflow → scheduled Claude Code agent).
Can also be triggered at any stage transition (e.g. after demo, before negotiation) for deal reviews.

Depth

For enterprise deals with 10+ calls, use search_recording_transcripts with tags (PainPoints, Objections, FeatureRequests) to filter signal from noise.
For SMB with 1–2 calls, the agent adapts automatically to shorter context.

CRM enrichment

Connect HubSpot or Salesforce MCP to pull deal properties alongside Claap data.
Add product usage data via Metabase / Amplitude MCP for expansion-ready handovers.

Output destination

Slack — post to #cs-handovers with a formatted summary.
Notion — create a page in your CS database with structured properties.
Google Doc — export for teams that live in Docs.
CS tools — adapt the output for Vitally, ChurnZero, or Gainsight intake templates.

Template

Adapt sections to your CS intake process — some teams want a risk score, others want an onboarding checklist.
Add sections for competitive context, expansion potential, key dates (renewal, QBR), product usage baseline.