A1 — Outbound List Builder

Sprint 1 atom. Pipeline:

Apify (pipelinelabs lead-search) → process-classify (ICP segment rules) → store-postgres

Tool surface

• runA1(input, deps?) — imperative shell, returns Result<A1Output, A1Error>
• createA1McpServe(deps?) — MCP-tool wrapper for marketplace exposure

Cost moat

Replaces Apollo / ZoomInfo / Lusha subscriptions ($99-$15K/mo per seat) with $0.001 per lead PPE on Apify. A 10K-lead monthly outbound budget runs at $10/mo all-in (Apify) vs $99-$1500/mo (Apollo) — 90-99% cost cut.

Input

{
  filters: { /* pipelinelabs actor input — industry, role, location, size, etc */ },
  maxResults: 100,        // cap on leads per run
  persist: true,          // write to Postgres atom store; false = dry-run
}

Output

{
  prospects: [{ lead: LeadRow, segment: 'vp_decision_maker', matchedRule?: '...' }],
  summary: {
    fetched: 100,
    classified: 100,
    persisted: 100,
    skippedSchema: 0,
    bySegment: { vp_decision_maker: 42, founder_smb: 18, ... },
  },
}

ICP segment rules

Rule-based (not LLM) — see src/segments.ts. Categories:

| Category | Rule | |----------|------| | founder_smb | seniority=owner, OR company.employees < 50 | | c_suite_general | seniority=c_suite | | vp_decision_maker | seniority=vp | | mid_market_director | seniority=director | | manager_influencer | seniority=manager | | unsegmented | fallback |

Rules are exported (DEFAULT_SEGMENT_RULES) so consumers can extend. LLM-based segmentation is a v0.2 upgrade if rule precision plateaus — defer until rule-based hits a recall floor in production.

n8n flavor

n8n-flavors/a1.json — same 3-node flow (HTTP request to Apify → IF rules → Postgres node).

License

MIT.