<< All versions

Skill v1.0.0

currentAutomated scan100/100
kami-community/kami/icp-segmentation
──Details
PublishedNovember 4, 2026 at 06:36 AM
Content Hashsha256:f54952509fa1261b...
Git SHA5a02f476bb73
──Files
Files (1 file, 1.6 KB)
SKILL.md1.6 KBactive
SKILL.md · 35 lines · 1.6 KB

name: icp_segmentation description: Derive distinct outbound ICP segments (B2B sales-assisted vs PLG self-serve) from the seller's validated domain positioning. Use before prospecting or when confirming who Kami will sell to. metadata: version: "1.0.0"


ICP Segmentation (outbound)

Produce 3–5 distinct segments for this seller's actual product (canonical domain + validated positioning). Do not inject a hardcoded vertical (healthcare, scheduling, etc.).

Method

  1. Lock identity to the submitted canonical domain. Ignore same-name companies in other industries.
  2. Read positioning + all ICP buckets + competitors (not just bucket[0]).
  3. Split by motion:
  • b2b_sales_assisted — teams; email a champion + economic buyer.
  • plg_self_serve — individuals; personalize from public signals; never invent emails.
  1. Technographic + firmographic filters must match this product category from evidence.
  2. B2B segments: real candidate_companies with domains (not news/listicles, not the seller domain).
  3. PLG segments: example_user_personas with hooks — not fake addresses.
  4. Set target_count budget per segment.

Hard rules

  • Never invent contact emails.
  • Never treat URL shorteners, job boards, or "Top N startups" pages as companies.
  • Never invent biotech/healthcare/scheduling/Calendly narratives unless first-party evidence says so.
  • Prioritize by Fit × Intent later (separate axes; decay intent after ~2 weeks).
  • Wait for founder confirmation before discovery.

Output

Fenced JSON matching Kami's SalesSegment shape.

All versions