How to Build Your Claude Sales Team (for B2B GTM)

export LEADOCEAN_API_KEY="paste-your-key-here"
mkdir sales-team && cd sales-team && mkdir -p .claude/agents

claude mcp add --transport http leadocean https://api.leadocean.io/mcp
claude mcp list

Sales team for Acme (project management software)

## ICP
- Buyer: VP of Sales, US, software companies with 51 to 200 staff.

## Team rules
- Every fact about a person or company comes from the leadocean MCP server.
- Missing value: write "unknown". Never guess.
- Never send anything. Drafts go to drafts/ as files.
- Records cap per run: 13 (10 from one search, 3 full lookups).

---
name: sizer
description: Counts how many people match the ICP and reports the number. Use before any search.
tools: mcp__leadocean__leadocean_count_leads, mcp__leadocean__leadocean_list_enum_values, Read
model: haiku
---

Count the ICP with leadocean_count_leads. If a filter value is rejected or the total is 0,
resolve the exact string with leadocean_list_enum_values and retry once.
Report: total, distinctCompanies, and which email statuses were counted.
Do not search. Do not look anyone up.

curl -s "https://api.leadocean.io/v1/people/search?count=true" \
  -H "x-api-key: $LEADOCEAN_API_KEY" -H "content-type: application/json" \
  -d '{"jobLevel":["VP"],"jobFunction":["Sales & Business Development"],
       "country":["US"],"industry":["Software Development"],
       "employeeRange":["51-200"],"limit":1}' | jq '.meta'

---
name: researcher
description: Finds ICP buyers and pulls full records for the top three. Use after the sizer returns a non-zero total.
tools: mcp__leadocean__leadocean_search_leads, mcp__leadocean__leadocean_get_lead, Write
model: sonnet
---

1. Call leadocean_search_leads with the ICP filters and limit 10. One call only.
2. Pick the 3 people with email_status verified whose title best matches "VP Sales".
3. Call leadocean_get_lead once per pick. Three calls at most.
4. Write the three records to research/YYYY-MM-DD.json. Keep only fields that came back.

---
name: drafter
description: Writes one first email per researched person. Use after the researcher has written its file.
tools: Read, Write
model: sonnet
---

Read the latest file in research/. For each person write a first email of 60 words or fewer.
Use one fact from the record: current role, company, or a past role from resume_data.
Quote only fields present in the file. If a field is missing, say so in a note, not in the email.
Save to drafts/YYYY-MM-DD.md.

claude
> Run the sizer. If the total is above 0, run the researcher, then the drafter.

claude -p "Run the sizer, then the researcher, then the drafter." \
  --allowedTools "Agent,Read,Write,mcp__leadocean__leadocean_count_leads,mcp__leadocean__leadocean_list_enum_values,mcp__leadocean__leadocean_search_leads,mcp__leadocean__leadocean_get_lead"

curl -s https://api.leadocean.io/v1/account -H "x-api-key: $LEADOCEAN_API_KEY" | jq

{
  "total": 1649,
  "totalIsExact": true,
  "totalAll": 4141,
  "distinctCompanies": 1125,
  "credits": 0,
  "defaults": "Counted only addresses worth sending to (emailStatus: verified or catch_all_valid or catch_all)",
  "facets": {
    "country": [{ "value": "US", "count": 1649 }],
    "industry": [{ "value": "Software Development", "count": 1649 }],
    "seniority": [{ "value": "vp", "count": 1649 }],
    "company_size": [{ "value": "51-200", "count": 1649 }]
  }
}

{
  "data": [
    {
      "profile_data": { "profile_full_name": "Jane Doe", "profile_headline": "VP Sales at Acme" },
      "contact_data": {
        "has_email": true, "email_status": "verified", "email_type": "work",
        "contact_current_experiences": [{ "job_title": "VP Sales", "company_name": "Acme", "company_domain": "acme.com" }]
      },
      "meta": { "source_ids": { "person_id": "<person_id>" } }
    }
  ],
  "meta": { "count": 10, "limit": 10, "nextCursor": "<opaque cursor>" }
}