Lead Engine Agent

Zip AI Free Blueprint →
MODULE_02 · LEAD_ENGINE

Stop chasing leads. Build a fleet.

The 5-agent Lead Engine that replaces an SDR team: how it’s wired, the funnel math behind 12 booked meetings a week, and the fleet config you can fork for your own business.

~24 min · 5 lessonsFREE · NO SIGN-UP

// LESSON 01Why lead gen is the perfect first fleet

Lead generation is high-frequency, low blast radius, and ruthlessly measurable — the three properties of a perfect first deployment (see Module 01). Every business needs it, almost every business does it inconsistently, and the cost of a miss is invisible: the lead just goes quiet and buys from whoever answered first.

78% of buyers go with the company that responds first. Your agents respond in 90 seconds, at 2am, on Christmas.

A human SDR costs $4,500–$7,500/month loaded, works 40 hours a week, and gets demoralized by rejection. A lead fleet costs a few hundred dollars a month in inference and tooling, works 168 hours a week, and never stops following up on day 7.

Key Takeaways
  • Lead gen = high frequency + low blast radius + measurable. The perfect first fleet.
  • Speed-to-lead is the whole game — and agents win it by default.
  • The fleet works 168 hrs/week for a fraction of one SDR’s cost.

// LESSON 02The 5-agent architecture

The Lead Engine isn’t one agent — it’s five specialists in a relay. Each does one job well, hands off cleanly, and can be debugged in isolation. Monolithic “do everything” agents fail; fleets compound.

01

HUNTER

Finds prospects matching your ICP — scrapes directories, monitors intent signals, watches job boards and permits. Output: raw lead rows.

↓
02

SCORER

Grades every lead against your ICP rubric (A/B/C/D). A-leads route instantly; D-leads never waste a human minute.

↓
03

ENRICH

Fills the gaps — company size, stack, recent news, the hook for personalization. Turns a name into a dossier.

↓
04

OUTREACH

Writes and sends the sequence — personalized first touch, 3–7 day cadence, books straight onto the calendar. Human approves drafts until accuracy is earned.

↓
05

TRACKER

Owns the funnel numbers — reply rate, meeting rate, cost per meeting — and reports weekly so you can re-tune the rubric.

fleet_config.yaml
fleet: lead_engine_v1
agents: [hunter, scorer, enrich, outreach, tracker]
icp: “service businesses · 5–100 staff · US”
cadence: [day0, day2, day4, day7] # stop on reply
guardrails: human_approves_first_touch: true
handoff: calendar_link + crm_row + slack_ping # to a human closer
Key Takeaways
  • Five specialists beat one generalist. Always.
  • Each agent has one job, one input, one output — debuggable in isolation.
  • Humans stay in the loop on first touches until the fleet earns autonomy.

// LESSON 03The funnel math

Here’s the real arithmetic from a production fleet. Run your own numbers through the same shape — the ratios matter more than the absolutes.

2,400
leads sourced /wk
320
A/B-graded /wk
38
replies /wk
12
meetings booked /wk
Cost per booked meeting: ~$47 with the fleet vs. ~$145 with a human SDR team. Same calendar, one-third the cost, zero turnover.
Tune These Three Ratios First
  • Grade rate (sourced → A/B): below 10%? Your HUNTER targeting is too loose.
  • Reply rate (contacted → reply): below 8%? Your OUTREACH personalization is generic.
  • Meeting rate (reply → booked): below 25%? Your booking flow has friction — fix the calendar, not the copy.
Key Takeaways
  • Watch ratios, not totals — the ratios tell you which agent to tune.
  • ~$47 per booked meeting is achievable at small-team scale.
  • The fleet’s edge compounds: it never skips day-7 follow-up.

// LESSON 04Where humans stay in the loop

The Lead Engine doesn’t remove humans — it removes busywork. Two places where humans are non-negotiable:

The first-touch gate

Until your OUTREACH agent sustains a 70%+ draft-acceptance rate for two weeks, a human approves every first touch. This is the Assist stage from Module 01 — don’t skip it.

The meeting itself

Agents book the meeting; humans run it. The fleet’s job is to put qualified conversations on a closer’s calendar — not to close.

The weekly rubric review

15 minutes with TRACKER’s report: re-grade 10 random leads by hand and compare to SCORER. Drift means re-tuning.

The voice check

Read 5 random outbound emails weekly. The moment they smell like AI, tighten the style guide and examples.

Key Takeaways
  • Agents book, humans close.
  • Weekly 15-minute reviews keep the fleet honest.
  • Autonomy is earned with accuracy data, never assumed.

// LESSON 05Ship your own in 21 days

You don’t deploy all five agents at once. The sequence: week 1 — SCORER on your existing inbound (instant value, zero risk). Week 2 — ENRICH + OUTREACH drafts with human approval. Week 3 — HUNTER + TRACKER, and the fleet is live end-to-end.

The 21-Day Checklist
  • Day 1–3: Write your ICP rubric (one page, A/B/C/D criteria).
  • Day 4–7: SCORER live on inbound. Compare to your gut on 50 leads.
  • Day 8–12: ENRICH wired to your CRM. Dossiers on every A-lead.
  • Day 13–17: OUTREACH drafting, human approving. Track acceptance rate.
  • Day 18–21: HUNTER sourcing + TRACKER reporting. Full relay running.

Want the rubric template, the fleet YAML, and the cadence copy as a starting kit — mapped to your business? That’s literally what the free Blueprint is.

Scroll to Top