decision-maker-finder

Decision Maker Finder

Finds 5 decision makers at a target company using Apollo.io UI via Chrome DevTools MCP. Search only, no enrichment, no credits consumed.

Prerequisites

• agency.config.json at repo root with icp.segments[].titles, icp.segments[].seniority, and tools.lead_enrichment
• Chrome DevTools MCP available (mcp__chrome-devtools__*)
• Apollo.io logged in on Chrome Beta
• Target company name and/or website

Phase 0: Intake

1. Read agency.config.json from the project root.
2. Extract:
   • icp.segments[] -- each segment's titles and seniority arrays
   • tools.lead_enrichment.tool -- must be "apollo"
   • tools.lead_enrichment.access -- must be "chrome_devtools"
   • tools.lead_enrichment.contacts_per_company -- number of contacts to find (default: 5)
   • tools.browser_automation -- Chrome DevTools MCP config
3. Accept parameters:
   • company_name -- (required) the target company
   • website -- (optional) company website domain
   • company_size -- (optional) if known, skip the size detection step
   • segment -- (optional) which ICP segment to use for title targeting
   • max_contacts -- (optional) how many contacts to return (default: 5)

Phase 1: Company Size Detection

If company_size is not provided, determine it:

WebSearch:

• Search: "{{company_name}}" "employees" OR "team size" OR "people" site:linkedin.com/company
• Search: "{{company_name}}" employees site:crunchbase.com
• Search: "{{company_name}}" "about us" "team"

Classify into size tiers:

• Early-stage: < 50 employees
• Mid-market: 50 - 500 employees
• Large: 500+ employees

Phase 2: Role Targeting

Based on company size, determine which titles to target:

Early-stage (< 50 employees)

Primary targets:

1. Founder / Co-Founder -- they make all decisions
2. CEO -- same person as founder in most cases
3. Head of Marketing / Growth -- if they exist, they own the ecommerce stack

Rationale: Small companies have flat hierarchies. The founder decides on agency partnerships.

Mid-market (50 - 500 employees)

Primary targets:

1. Head of Ecommerce / Director of Ecommerce -- owns the Shopify store
2. VP Marketing / Head of Marketing -- owns the growth budget
3. Head of Digital / Director of Digital -- owns the online experience
4. CMO -- if the company has one, they approve agency spend

Rationale: Mid-market companies have functional leads. Target the person who owns the ecommerce P&L or the marketing budget.

Large (500+ employees)

Primary targets:

1. VP of Ecommerce / VP of Digital -- senior enough to approve budgets
2. Director of Ecommerce -- operational decision maker
3. Head of D2C / Head of Online -- if they have a dedicated D2C function
4. SVP Marketing -- if no dedicated ecommerce leadership

Rationale: Large companies have multiple layers. Target VP+ who can approve agency contracts without escalating.

Cross-reference these defaults with the icp.segments[].titles and icp.segments[].seniority arrays from config to ensure alignment with the agency's actual target persona.

Phase 3: Contact Search via Apollo.io (Chrome DevTools MCP)

Use Chrome DevTools MCP to search Apollo.io UI. This is the ONLY path. No API, no WebSearch fallback.

Chrome DevTools MCP Workflow:

1. mcp__chrome-devtools__navigate_page to https://app.apollo.io/
2. Verify logged-in state
3. Navigate to People Search
4. mcp__chrome-devtools__fill company filter with {{company_name}}
5. mcp__chrome-devtools__fill title filter with target titles from Phase 2
6. Apply seniority filters matching ICP config
7. mcp__chrome-devtools__click search/apply filters
8. mcp__chrome-devtools__wait_for results to load
9. mcp__chrome-devtools__take_snapshot to read results

Extract from each result:

• Name
• Title
• Company
• LinkedIn profile URL (Apollo shows this)

IMPORTANT:

• Pick 5 decision makers per company
• DO NOT click "Enrich" or "Access Email" on Apollo (no credits used)
• DO NOT use Apollo enrichment features
• We only use Apollo for SEARCH, not enrichment
• Email discovery happens via Snov.io (separate skill)

If fewer than 5 results: accept what Apollo returns. Note the shortfall.

Phase 4: Contact Data Assembly

For each found contact, record:

• Name: Full name from Apollo search results
• Title: Current title at the target company
• Company: Company name
• LinkedIn URL: Direct profile link from Apollo
• Email: NOT from Apollo. Will be found via email-validator (Snov.io) skill.
• Phone: NOT from Apollo. Will be found via phone-finder (Lemlist) skill.

DO NOT attempt to enrich or find emails at this stage. That's handled by downstream skills.

Phase 5: Prioritize Contacts

Rank the found contacts:

1. Title match score: How closely does their title match the target titles? Exact match = 10, partial match = 5, related = 3.
2. Seniority alignment: Does their seniority match the company size rules? Aligned = 10, adjacent = 5.
3. Recency: Recently joined (< 6 months) = bonus, as they may be actively building their vendor stack.

Sort by total priority score, descending.

Phase 6: Output

Return structured JSON:

{
  "company": "BrandX",
  "website": "https://brandx.com",
  "company_size": "Mid-market (120 employees)",
  "size_tier": "mid-market",
  "targeting_rationale": "Mid-market company, targeting Head/Director of Ecommerce or VP Marketing.",
  "contacts": [
    {
      "name": "Jane Doe",
      "title": "Head of Ecommerce",
      "linkedin_url": "https://linkedin.com/in/janedoe",
      "confidence": "HIGH",
      "method_used": "apollo_chrome_devtools",
      "priority_score": 27,
      "notes": "Joined 4 months ago. Likely building her vendor stack."
    }
  ],
  "recommended_contact": "Jane Doe -- Head of Ecommerce, highest priority score",
  "searched_at": "2026-03-16T09:30:00Z"
}

Phase 7: Log to CRM

Use crm-writer to log the decision maker data to the Pipeline tab. Write: company, contact name, title, LinkedIn URL, confidence, found_at timestamp.

No approval gate. Contacts are logged directly.

Example Usage

Trigger phrases:

• "Find the decision maker at [company]"
• "Who should I reach out to at [company]?"
• "Find the right contact at [company]"
• "Get me the founder of [company]"
• "Find the Head of Ecommerce at [company]"

User: Find the decision maker at BrandX
Assistant: [reads config, determines company size = mid-market, targets Head of Ecommerce / VP Marketing, searches via Apollo Chrome DevTools MCP, returns 5 prioritized contacts]

User: Who should I email at this 10-person startup?
Assistant: [size = early-stage, targets Founder/CEO, searches via Apollo Chrome DevTools MCP, returns contacts]