closed-won-replication-play

Template placeholders

Replace every {{...}} before enabling. See the setup checklist reference for the full setup list.

• {{CRM}} — Your CRM (e.g. HubSpot) — source of the closed-won event and target of writes
• {{LOOKALIKE_COUNT}} — Lookalikes per win (default: 5)
• {{REVIEW_SURFACE}} — Where the batch review task is created (e.g. your desk/task queue)

Purpose

Every closed-won deal is a targeting signal. The moment a deal closes, use the winning company's profile to find {{LOOKALIKE_COUNT}} lookalike companies and draft outreach while the story is fresh. The rep reviews and approves before anything sends.

---

Input

You receive a {{CRM}} deal that just moved to Closed Won. The payload includes the deal record and its associated company/contact.

---

Step 1 — Extract the Win Profile

From the {{CRM}} deal and its associated company, extract:

• Industry / vertical
• Company size (headcount range)
• Geography (country/region)
• Tech stack (if available — look for CRM, marketing tools, data tools)
• Go-to-market model (B2B SaaS, services, marketplace, etc.)
• Use case / pain point that drove the purchase (check deal notes and contact activity in {{CRM}})
• Deal owner (the {{CRM}} user who owns the deal — you'll need their name and email for later)

If a field is missing or unclear, make a reasonable inference based on what's available. Don't block on incomplete data.

---

Step 2 — Build the Lookalike Search Profile

Construct a specific search profile. Good example: "B2B SaaS, 50–200 employees, US-based, uses Salesforce, scaling a GTM team." Generic is weak — be specific.

Use the extracted attributes to run a company search. Look for companies that:

• Match the industry and company size
• Are in the same geography or equivalent market
• Share a similar GTM model or tech stack when inferable
• Are NOT already in {{CRM}} (not a customer, not active pipeline, not a contact record's company)

Find exactly {{LOOKALIKE_COUNT}} qualifying lookalike companies. If you're having trouble hitting the count, broaden one criterion at a time (geography first, then size range). If you still can't after broadening, proceed with however many you found and flag the count in the review task.