We find in-pain consumers the moment they need a service — and live-transfer them to the businesses that pay for them.
NYC pest control. One confirmed pest signal on a block — a 311 rat report, a DOH inspection fail, a DOB demolition filing — means the whole street is at risk. So we resolve every callable homeowner on it.
Brooke — a real Gemini-Live AI voice agent — holds a full homeowner conversation and warm-transfers interested buyers. Recorded live: the homeowner said "Definitely," the transfer fired to the buyer line.
CONFIDENTIAL — MAY 2026
The operator — and every local service business — pays a fortune to reach people who already said no to someone else. The lead market is built to recycle attention, not to find intent.
Angi, HomeAdvisor and Thumbtack sell the same form-fill to 3–5 competitors at once. Contact rates collapse to 15–30%, transfers to 10–25%. You're not buying a customer — you're buying a race.
A paid customer on Meta/TikTok runs $200–600 each. LSA "qualified" pest calls cost $20–80 apiece. The math only works if the homeowner overpays too.
Data-broker lists ($0.001–0.01) are demographic guesses — 5–10% contact, 1–3% transfer. They tell you who might own a home, never who has rats right now.
When 60–90% of a shared lead's value is gone before you dial, the operator's budget mostly funds the platform's resale margin — not their own pipeline.
Every incumbent buys attention after a consumer raises their hand on a form. HeatSeak reads the public record the day the pain appears, reconstructs the household behind it, and has Brooke on the phone — warm-transferring the interested ones — before any form is ever filled.
A single confirmed pest event condemns the block — so we resolve every callable homeowner on the street.
One signal condemns the block. A single confirmed pest event — a 311 rodent complaint, a DOHMH rat-inspection failure, a DOB demolition filing, a DOHMH restaurant citation — means the infestation isn't contained to one address.
So we resolve the whole street. From that one anchor we skip-trace every callable homeowner on the block. One public data point becomes dozens of in-pain, perfectly-timed leads no competitor knows exist.
Live now: 1,073 ranked infected NYC blocks · 219 callable homeowners (11 Bugatti-grade, 13 Premium) · $1,682 predicted pipeline — built entirely from public records + free people-search.
No single public source is a lead. We fuse partials — 311 + DOH + DOB + DOHMH records, matched on geo and time, then enriched with people-search — to construct a complete, callable, provenance-labeled lead no one source could give. Optional skip-trace adds verified contact at ~$0.07/home (85% verified). The lead is manufactured, not resold — which is why it's exclusive and near-zero cost.
A live Gemini-Live voice agent on dedicated SignalWire. She runs a full NEPQ conversation, AI-disclosed and honest, then warm-transfers interested homeowners straight to the buyer's line. Every call is recorded → analyzed voice-to-voice → distilled into learnings → injected into the next call's prompt. She compounds autonomously.
HeatSeak fuses scattered public records into callable homeowner leads no single source can produce — then Brooke works the entire infected street and warm-transfers the interested ones to the buyer.
A live lead token travels Fuse → Cluster → Resolve → Call → Qualify → Transfer → Learn, on a loop.
Pull NYC 311 rodent reports, DOHMH rat-inspection fails, DOB demolition filings, DOH restaurant citations. Fused on time + geo + address, they become ground truth: this block has a confirmed pest problem right now. Every field provenance-labeled.
One signal means the whole block is exposed — pests don't respect property lines. We rank 1,073 infected NYC blocks by signal density and severity, turning a point-incident into a street-wide opportunity.
For each block, fuse the address against free people-search → owner name + verified line. 219 callable homeowners LIVE in the CRM (11 Bugatti, 13 Premium), a $1,682 predicted pipeline. Near-zero lead cost.
Brooke dials on dedicated SignalWire, AI-disclosed and honest — she works only from public record. Proven live: she holds a full natural homeowner conversation, not a script-reader.
Brooke runs Jeremy Miner's NEPQ framework — questions that pull the homeowner to their own conclusion about the risk on their street. Confirms concern, urgency and interest.
Interested homeowner → Brooke transfers live. Already proven on a real recorded call: homeowner said "Definitely" → transfer fired. Pay-per-call, 50/50 with the buyer.
| Stage | Input | Output | Live number |
|---|---|---|---|
| Fuse | 311 / DOH / DOB / DOHMH records | Confirmed block signal | 4 public sources |
| Cluster | Confirmed signals | Ranked infected streets | 1,073 blocks |
| Resolve | Block address + people-search | Callable homeowner lead | 219 live in CRM |
| Call | Brooke + SignalWire | NEPQ conversation | ~$0.25/min all-in |
| Transfer | Interested homeowner | Buyer's phone | 50/50 pay-per-call |
Not a spreadsheet of leads — a live operating console where Luka and the buyer log in, see every lead, and launch a real AI call with one tap. Compliance, recordings and intelligence are built into the same screen.
Business leads and infected-street homeowner leads side by side — graded Bugatti / Premium / standard, with the fused provenance behind each.
No dialer, no setup. Tap a lead → Brooke calls live → watch the outcome land in real time.
Connected, qualified, transferred, callback, not-interested — updated per lead as calls happen.
Every opted-out number with transcript, recording and voice-intelligence. Opt-outs actually enforced — the number is suppressed and never called again. TCPA-clean.
Open any call: read the transcript, hear the recording, see per-call voice analysis (tonality, critical moments, what to improve).
A live record of what's shipping — the buyer watches the product compound in real time.
Why it matters to a buyer: JC doesn't get a CSV and a prayer. He gets a logged-in console where he can see the lead, hear the call that produced it, verify it was honest and compliant, and watch his pipeline fill — with one-click control over the AI doing the work. Access model: dedicated per client — isolated data, the buyer's own numbers. Shared brain, dedicated body.
Competitors ship a static script; every HeatSeak call makes the next one convert better. The data flywheel is real, not a slide.
The improvement loop runs on the calls we're already making — no labelers, no manual QA.
Re-aim the engine at a new vertical and the same self-improvement machinery sharpens Brooke there too. Plus a sparring arena: Brooke vs. AI prospect personas, judge-scored, between live calls.
For 30 years, qualified lead-gen meant a $20–80 shared lead or a $15–25/hr call center. Both are now obsolete. Three independent shifts hit maturity at the same moment.
Brooke is live Gemini-Live on dedicated SignalWire — proven holding a full NEPQ homeowner conversation and warm-transferring. Wasn't possible 18 months ago. A human-grade qualifying conversation now costs ~$0.25/min.
NYC 311, DOH, DOB, DOHMH records expose exactly who has a problem, for $0. Our fusion engine stitches partials + people-search into a complete callable lead no single source could give. Optional skip-trace ~$0.07/home (85% verified).
$5/lead in 2024 → ~$0.10 now. What used to be an enterprise budget line is now a rounding error.
Clippers deliver views at $0.10–1.00 / 1,000 — paid only for performance. No SDRs, no $200–600 Meta CAC. At $0.10/1K, every $1 of spend returns $3.40 of MRR inside 30 days.
Every Brooke call is recorded, analyzed voice-to-voice, and the learnings are injected into the next prompt. Brooke gets measurably better every single call, autonomously. The asset compounds while it works.
We don't sell "data." We sell qualified, ready-to-talk conversations — starting where intent is undeniable and competition is asleep: local home-services pest, then re-aim the same engine at every adjacent vertical.
Comps: Apollo $1.6B · Clay $1.3B · ZoomInfo $12B IPO. SOM at JC's capacity (50→200 transfers/day @ ~$35, 50/50). One buyer. Wedge already live.
$0.30/call → $17.50 per transfer = 96% margin. Zero human agents, 24/7 within TCPA.
Google LSA $20–80, Angi $15–30 (shared 3–5x), Thumbtack $8–25. We deliver hotter, exclusive, AI-qualified transfers at a fraction.
Premium pests carry premium economics: Termites $80–120, Bed Bugs $60–100 — year-round, recession-resistant.
| Adjacent vertical | Transfer value | Same infra? |
|---|---|---|
| Home insurance | $40–80 | ✓ |
| Auto insurance | $20–40 | ✓ |
| Health coverage | $30–100 | ✓ |
| Roofing / HVAC / plumbing | $40–120 | ✓ |
The lead-gen world splits into three camps, each with a structural flaw we don't share.
Angi, HomeAdvisor, Thumbtack sell the same form to 3–5 buyers. Cold by call-time, shared, no conversion loop. We deliver exclusive, AI-qualified, live-transferred conversations.
Can qualify — but at $15–25/hr, turnover, no nights/weekends, zero learning loop. Brooke is 24/7, $0.25/min, never has a bad day, improves after every call.
Apollo, Clay, ZoomInfo, Instantly give you data or a dialer — no intent, no qualification, no transfer, no loop. We close it: signal → call → qualify → warm-transfer → revenue → self-improve.
| HeatSeak | Brokers (Angi/HA) | Human call centers | AI dialers (Apollo/Clay) | |
|---|---|---|---|---|
| Cost / qualified contact | $0.30/call | $15–80/lead | $15–25/hr labor | $49–800/mo (no contact) |
| Intent source | Fused public-record (real-time) | Self-reported form | Whatever list you load | Stale DB, no intent |
| Exclusivity | Exclusive transfer | Shared 3–5 buyers | Depends | N/A |
| Qualification | AI NEPQ conversation | None | Human (variable) | None |
| Live warm-transfer | ✓ Proven | ✗ | ✓ (costly) | ✗ |
| Availability | 24/7, TCPA-bounded | Inbound only | Business hours | N/A |
| Margin | 96% pest / 78–85% SaaS | Thin / arbitrage | Negative on labor | Buyer bears cost |
| Gets smarter per use | ✓ Auto-train loop | ✗ | ✗ | ✗ |
| Compliance | DNC enforced, AI-disclosed | Varies | Varies | N/A |
The fusion layer — four disconnected public-record feeds into one callable, provenance-labeled lead.
The self-improving voice asset — every recorded call → voice-to-voice analysis → namespaced learnings → next-call injection, plus a judge-scored sparring arena. The product compounds while it runs.
The closed outcome loop — we own signal-to-revenue end to end, so we learn from conversion and they don't.
Honest flag: enrichment alone is replicable in weeks (our own advisory panel said it). That's exactly why we don't compete on enrichment — we compete on the closed, self-improving conversation loop, distributed by creators at 1/50th the CAC. That part nobody else has assembled — and it's already live.
We studied how each incumbent went to market then, how they still promote now, and where their model structurally fails. The pattern is identical across all four: shared/recycled leads, rising CAC, and zero real-time intent. That gap is the entire reason a fused-data, exclusive, AI-qualified transfer can take share.
Two opposite playbooks stapled together. Angie's List (1995) was a paywalled, door-to-door trust/subscription play. HomeAdvisor (ServiceMagic, 1998) was the brute-force pay-per-lead machine — national TV, PPC, and outbound sales armies cold-calling contractors — that monetized by selling the same homeowner lead to 3–5 competing pros at once. IAC merged them in 2017; that lead-resale mechanic became the reputational liability.
How they promote now: National TV/CTV + paid social + SEM; consumer marketing rose to ~$340M in FY2025 (+5%) even as revenue fell 13%. They launched "AI Helper" (Jun 2025) and put the marketplace inside ChatGPT (Mar 2026) — chasing an AI front door because the old one broke.
Where it's broken: the 2021 "Angie's List → Angi" rebrand torched 25 years of organic SEO — CEO Joey Levin: "the impact of the brand change was more severe than we expected." Forced migration to paid channels. The lead is shared, cold by call-time, and self-reported (no real-time intent). Contractors bake lead cost into bids → homeowners pay more.
Pure pay-per-lead marketplace — pro pays per contact, win or lose; dynamic weekly pricing. ~$400M revenue (FY2024), EBITDA-profitable since 2023, >90% gross margin on home, ~300K active pros. Last valued $3.2B (2021). Proves the model scales to nine figures — but leads are non-exclusive, sold to multiple pros.
Demand from the ad side. Yelp Services revenue $948M in 2025 (+8%), ~65% of its $1.46B total, with 15+ straight quarters of double-digit services growth while restaurant/retail shrinks. "Request a Quote" is structurally identical to Thumbtack's lead flow — and just as shared and cold.
The 800-lb gorilla — ranks above all ads + organic, charges pay-per-lead (~$53 avg, $20–180 by trade). Contractor adoption ~28%→~70% (2022→late-2025). But leads are non-exclusive and a reported ~45% are unbookable (spam, wrong number, out-of-area). Industry benchmark, not Google-reported.
| Lever | HeatSeak | Angi / HomeAdvisor | Thumbtack | Yelp Services | Google LSA |
|---|---|---|---|---|---|
| Pricing unit | $0.30/call → $17.50 transfer | $15–30/lead | $8–150 / contact | ad subscription | ~$53/lead avg |
| Intent | Real-time public-record | Self-reported form | Self-reported form | Self-reported | Search keyword |
| Exclusivity | ✓ Exclusive transfer | Shared 3–5 | Shared | Shared | Shared (~45% dead) |
| Qualification | ✓ AI NEPQ call | None | None | None | None |
| Self-improving | ✓ Auto-train loop | ✗ | ✗ | ✗ | ✗ |
| 2025 direction | Live & growing | Rev −13%, pros −23% | ~$400M, flat-private | Services +8% | Adoption rising |
Verification discipline: Angi $1.76B→$1.03B, ServiceTitan, IAB $294.6B, Yelp Services $948M and FTC $7.2M are primary-sourced. "~45% unbookable LSA," "70% adoption," and "CAC +40% from SEO loss" are industry/analyst benchmarks (BlueGrid, MatrixBCG), labeled as such — not audited filings.
The demand side isn't a hypothesis we have to create — it's a proven, recurring spend the incumbents already harvest. Our job is to deliver the same buyers a hotter lead at a fraction of the cost. Here's exactly who pays, and how much.
Home-service SMBs & contractors — plumbing, HVAC, roofing, electrical, cleaning. Fragmented, can't run sophisticated marketing, structurally addicted to a steady flow of new jobs. Most spend $500–$10,000/mo on digital marketing (Workiz/WebFX).
Pest-control operators — our wedge. ~34,000 US companies in a ~$29.7B industry (IBISWorld 2026), highly fragmented, recurring-revenue, needs constant new customers.
PI / legal firms — the richest lead vertical. Avg PI firm pays ~$284/lead, ~$468/signed case; exclusive leads $250–$600+ (First Page Sage 2026, $3.3M-spend study).
Agencies & lead resellers — buy in bulk, mark up, and route to their own SMB clients. The whole pay-per-call affiliate economy runs on this.
| Vertical | Lead / call payout | Buyer | Why they pay |
|---|---|---|---|
| Personal injury / legal | $250–$600+ (avg $284/lead) | PI firms | One signed case = thousands in fees |
| Water damage / remediation | $80–$180 | Restoration cos | High-ticket emergency jobs |
| Roofing | $60–$130 | Roofers | $10K+ average job |
| HVAC / plumbing | $35–$110 | Trades | Recurring + install revenue |
| Pest control (termites/bed bugs) | $60–$120 | Pest operators | Year-round, recession-resistant |
| Insurance (home/auto/health) | $20–$200+ | Carriers/agents | Lifetime policy value |
Sourcing: $657B is Angi's own self-published TAM estimate (label accordingly; independent US home-services estimates span ~$475–842B by methodology). Per-vertical payouts blend SearchLight/BlueGrid LSA benchmarks, Service Direct, and First Page Sage — directional industry figures, not audited.
The data sources are commodities. The moat is the four layers we stacked on top — each one compounds, none of them is on a price list, and the gap widens every day we operate.
Most tools resell a record they bought. We construct one — partial public signals fused with people-search into a complete, callable, provenance-labeled lead no provider has in inventory. Competitors who lose API access lose their product. We'd still have ours — the source is the public record itself.
One signal means the whole street is exposed — so we resolve every callable homeowner on it. Original IP, live: 1,073 ranked blocks · 219 callable homeowners · 24 Bugatti/Premium · $1,682 pipeline. Cost to build all 219: ~$0 on fusion, or ~$15 with optional skip-trace.
A real Gemini-Live agent, proven live warm-transferring ("Definitely" → transfer fired). Every call: record → voice-to-voice analysis → namespaced learnings → next-prompt injection, plus a judge-scored sparring arena. Call #10,000 is categorically better than call #1 — a new entrant starts at call #1.
Every buyer gets their own SignalWire, numbers, data, dashboard — isolated, TCPA-enforced, with full transcripts + recordings + voice-intelligence per lead. Intelligence flows across clients; identity and data never cross. Enterprise-grade isolation, hard to copy, harder to leave.
| Source | Cost / Lead | Notes |
|---|---|---|
| HeatSeak — open-data fusion + free people-search | ~$0.00 | the wedge |
| HeatSeak + optional skip-trace | $0.07 | 85% verified mobile |
| Brooke call, all-in (~90s) | ~$0.38 | $0.25/min |
| Angi / HomeAdvisor (shared) | $15–30 | 20–30% contact |
| Google LSA | $20–80 | per lead |
No sales team. Three creator-led channels, staggered, resting on one arbitrage: clipper distribution acquires a paying user for ~$26 while Meta charges $200–600. That's 6–23x cheaper — the entire reason this scales without burning capital.
Creators paid $0.10 / 1,000 views. Same funnel the deck is built on — 0.08% CTR → 8% trial → 6% trial-to-paid → $89 ARPU — and every budget produces a blended $26 CAC, 27x LTV pool.
| Budget | Views | Paid users | New MRR | CAC | LTV pool |
|---|---|---|---|---|---|
| $5K | 50M | 192 | $17,088 | $26 | $136,704 (27x) |
| $15K | 150M | 576 | $51,264 | $26 | $410,112 (27x) |
| $25K | 250M | 960 | $85,440 | $26 | $683,520 (27x) |
The one number that closes it: at $0.10/1K, $1 of spend returns $3.42 of MRR inside 30 days. Break-even = 56 paying users. Even at 10x hotter ($1.00/1K), the channel stays 2.7x LTV-positive.
The unfair advantage. 50+ MLM leaders across 50 countries, each 1,000–50,000 followers, activated via a 30% recurring affiliate — zero ad budget. 50 affiliates × 100–500 users = 5,000–25,000 users at $0 CAC. Directive: affiliate-first — prove 20 affiliates + a 50–100 user cohort before scaling paid.
Deliberately the backstop. Hook: "Find any business owner's phone number in 10 seconds." Target $5–12 CPL, 4x ROAS within 60 days. Turned on only once a channel clears Gate 2 (6%+ trial-to-paid, payback <90 days). Paid spend follows proof, never precedes it.
| Channel | CAC / paid user | vs. Meta | Budget posture |
|---|---|---|---|
| Klemen affiliate network | $0 | ∞ cheaper | activate first |
| Clippers @ $0.10/1K | $26 | 6–23x cheaper | scale on green-light |
| Clippers @ $1.00/1K (stress case) | $263 | parity | still 2.7x LTV-positive |
| Meta / TikTok ads | $200–600 | baseline | secondary throttle |
This isn't a better ad-buy or a thinner-margin reseller. Each layer below is a structural advantage that gets stronger the longer we run — and not one of them is on a price list.
Brokers resell a record they bought — so the same lead lands on 3–5 desks and goes cold. We construct the lead by fusing partial public signals no provider sells whole, then deliver it to exactly one buyer. Exclusivity isn't a policy we can be undercut on; it's a property of how the lead is made.
Angi/Thumbtack/LSA leads are self-reported and stale. Ours fire off live public signals — 311 reports, DOH/DOB filings — the moment a homeowner is in pain. Intent at the source means a call that lands, not a form filled out a week ago by someone now ignoring five callbacks.
Every call records → voice-to-voice analysis → namespaced learnings → next-call injection, plus a judge-scored sparring arena. Call #10,000 is categorically better than call #1. A new entrant starts at call #1. Incumbents reselling forms learn nothing from conversion — we own the outcome, so we own the learning.
Fused from free open-data + free people-search → ~$0 supply cost (~$0.07 with optional skip-trace). A Brooke call is ~$0.30–0.40 all-in. We're 50–250x cheaper at the top of the funnel than LSA/Angi/Thumbtack — and the lead is hotter. That's the asymmetry the whole model rests on.
One engine. Swap the keyword set and the same fusion + FIRE scoring + Brooke stack re-aims at roofing, HVAC, insurance, or legal — verticals where a transfer is worth $40–$600 instead of $35. Every dollar of R&D pays off across every revenue line, and each new vertical inherits the moat fully built. A reseller has to renegotiate supply for each vertical; we just change a config.
We could sell lead-gen as a service to one company at a time. It's easier to start and pays the bills. But it caps at a payroll. Building the platform is the harder thing — and that difficulty is the moat. It's the only path to a unicorn outcome, and here's the honest math on why.
| Dimension | Software Platform (what we're building) | Lead-gen Agency / Services shop |
|---|---|---|
| Priced on | Revenue multiple (SaaS median ~8.5x, 2026) | EBITDA multiple (3–6x; ~0.7–2.3x revenue) |
| Gross margin | 70–85% (SaaS); ~83% marketplace | 10–20% net; 50–60% blended delivery |
| Cost of next customer | ~$0 marginal | ~proportional new labor |
| Revenue type | Recurring (subscription/usage) | Non-recurring retainer, 30-day cancel |
| Scaling shape | Leveraged (Rule of 40 applies) | Linear (Rule of 40 doesn't apply) |
| Defensibility | Network effects / data moat / switching cost | Relationship-based, fragile, owner-dependent |
| Real comparable | ZoomInfo 1.8x rev / 84% GM (even depressed) | WPP 0.7x rev (largest agency on earth) |
Healthy SaaS trades at ~8.5x revenue (2026). The world's largest ad agency, WPP, trades at 0.7x. Even a depressed lead-gen platform — ZoomInfo at 1.8x revenue, 84% gross margin — is priced richer than the best-run agency holding company alive. The same revenue dollar is worth ~3x (vs depressed platform) to ~10x+ (vs healthy SaaS) more inside a platform.
You sell hours. Every new revenue dollar needs a new labor dollar. Margins cap at 10–20% net; growth is linear.
The asset walks out nightly. Clients cancel at 30 days; the moat is a personal relationship that exits with the account lead.
Concentration discount. Top client >20% of revenue = automatic valuation haircut. Exits at 3–6x EBITDA if you can sell at all.
Honest caveat: agencies aren't bad businesses and the platform build is genuinely harder — the cold-start and the R&D are real. The argument is structural and relative, not that services have no value. We use services revenue (pay-per-call, live today) to fund the platform build — best of both.
This is not a bet on a new market — it's a more capital-efficient, more defensible way to capture a market that has already produced four generations of billion-dollar outcomes. The pattern is clear, and we fit it.
| Generation | Company | Vertical | Value | What it owned |
|---|---|---|---|---|
| G1 · Listings/ads | Yelp | Local discovery | ~$1.46B rev (2025) | Discovery (thin moat) |
| G2 · Lead marketplace | Angi (HA+Angie's List) | Home services | ~$890M rev at 2017 merger | The lead (shared) |
| G2 · Lead marketplace | Thumbtack | Home services | $3.2B (2021) | The lead, >90% GM |
| G3 · Transaction OS | ServiceTitan | Trades/contractors | IPO ~$6.3B → ~$8.9B day-1 | The whole job, >110% NDR |
| G4 · AI-native | EvenUp | Personal-injury law | $2B+ (Oct 2025) | The workflow, AI |
| G4 · AI-native | Sierra / Decagon | AI agents | $15.8B / $4.5B | AI does the work |
EvenUp (AI for personal-injury — our exact adjacent vertical) went $1B+ → $2B+ in under a year (Bessemer, Oct 2025) on ~10,000 cases/week.
AI-voice re-rated fastest of all: Decagon $1.5B→$4.5B (2025→26); Sierra $4.5B→$10B→$15.8B; ElevenLabs $11B (Feb 2026). Capital is voting for AI that owns a vertical workflow end-to-end.
Bessemer + Euclid: AI-native vertical companies hit $100M ARR faster than any prior SaaS generation — "particularly in legal, healthcare, and housing." That's us.
A thin AI wrapper fetches ~8x ARR.
Genuine workflow automation + defensibility fetches up to 25x; >120% NDR + a proprietary data loop pushes to ~20x (Capitaly vertical-AI comp sheet, Feb 2026).
An AI caller that qualifies and converts the lead it generated, then feeds outcomes back to sharpen the next lead, is the textbook self-improving data loop — the engine that re-rates a lead-gen company into a vertical-AI company. We plug into PI's proven $284/lead, $468/case willingness-to-pay at a fraction of incumbent CAC.
Honest pre-empt: the first "AI SDR" generation (11x, Artisan) over-promised full human replacement and many customers reverted to hybrid by early 2026. Our framing is deliberately not "AI replaces the call center" — it's a self-improving AI caller that warms a fused-data lead with a human-grade validation gate before delivery. That captures the economics while sidestepping the failure mode that burned the pure-replacement plays. Private valuations (EvenUp, Sierra, Decagon, Thumbtack's $3.2B 2021 mark) are confirmed via funding announcements; circulating ARR figures are directional estimates, labeled as such.
Don't take our word for the economics. Drag the inputs and watch the math.
How paying clippers ~$0.10 per 1,000 views turns a small spend into millions of views, real leads, and transferred calls — at an effective CAC of pennies vs Meta's $200–600.
At an identical budget, how many customers Meta/Google buy at $200–600 CAC vs how many Clippers deliver at ~$0.10/1K views — and the multiple.
Brooke dials, connects, transfers — daily/monthly revenue, Brooke's all-in cost, gross margin, and our 50% share of the pay-per-call deal (net of Brooke's cost — the honest version).
The whole business hinges on one asymmetry: the lead costs almost nothing to build, and Brooke works it for pennies a minute. Everything else is upside.
Leads → connects → qualified → warm-transfers. Each stage labeled with its conversion rate (illustrative defaults — drag the calculator below to model your own).
Lead cost. Fused from free public open-data + free people-search. Optional skip-trace ~$0.07/home for 85% phone-verified.
Brooke ~$0.25/min, all-in. A qualifying call runs under 90s → ~$0.30–0.40 cost per call placed.
Pay-per-call, 50/50. Avg transfer ≈ $35 → $17.50 each side. Termites $80–120, bed bugs $60–100.
Gross margin on the call layer. No human agents — 100% AI, 24/7 within TCPA. And it compounds: cost stays flat, conversion climbs.
| Lead cost | ~$0 |
| Optional skip-trace | ~$0.07/home |
| Brooke call (≤90s) | ~$0.30–0.40 |
| Avg transfer revenue | $35 ($15–120) |
| Each side (50/50) | $17.50 |
| Gross margin (call layer) | ~96% |
vs the market: $0.30/call vs $20–80 LSA, $15–30 Angi, $8–25 Thumbtack. 50–250x cheaper at the top of the funnel — and the lead is hot, not cold.
One engine, three revenue models — so we capture both the operator who wants leads delivered and the self-serve user who wants the tool.
Live now, with JC. Brooke qualifies and warm-transfers. 50/50 split, ~$35/transfer. At 200 transfers/day = ~$105K/mo; conservative 50/day = ~$26K/mo.
Token-based enrichment — industry + location → enriched leads + FIRE scores in seconds. Starter $49 / Pro $149 / Scale $349, Enterprise $499 (M4–6). Blended ARPU ≈ $89/mo.
5–8 tokens per enriched lead vs 1–2 of API cost → 3–5x margin per transaction. Sold as packs or on top of subscriptions. (Market it as "leads/credits.")
| Revenue / enriched lead | $0.40–0.80 |
| API cost / lead (waterfall) | $0.08–0.15 |
| Gross margin / lead | 78–85% |
| Monthly usage / active user | 200–500 leads |
| Effective ARPU | $89–249/mo |
Pay-per-call gives predictable cashflow from operators today; SaaS gives a network-effect product that compounds. The same data fusion + FIRE scoring + Brooke stack powers all three — so a dollar of R&D pays off across every revenue line.
Discipline is the story: every dollar of capital is unlocked only after the prior milestone is proven. No blind scaling.
ARPU holds ~$89 across the curve. M12 cumulative revenue $4.1M · M24 cumulative $33.9M.
| Round | Amount | Timing | Gate to unlock next |
|---|---|---|---|
| R1 Seed (open now) | $200K | Mo 0–3 | 500 users · $45K MRR · positive unit economics |
| R2 Growth | $400K | Mo 4–8 | 3,000 users · $250K MRR · 2+ profitable channels |
| R3 Scale | $600K | Mo 9–14 | 10,000 users · $800K MRR · 3+ countries |
| R4 Domination | $500K | Mo 15–20 | 25,000 users · $2M MRR |
| R5 War Chest | $300K | Mo 21–24 | $3M+ MRR · Series A ready · $50–100M val |
→ $1–1.5M on the 10%
→ $5–10M on the 10%
stake → 8%
Full $2M across all rounds → blended 25–30% equity → at a $50–100M exit, a 25% stake = $12.5–25M (6.25–12.5x). Comparable exits: Apollo.io $1.6B (16x ARR), Clay $1.3B (50x ARR), ZoomInfo $12B IPO (40x ARR), Instantly ~$100M (8x ARR).
Not a pitch for what we'll build — a working product with a real call recorded and leads live in the CRM.
A Gemini-Live AI voice agent on dedicated SignalWire held a full NEPQ homeowner conversation, AI-disclosed and honest, and warm-transferred an interested homeowner. A real recorded call exists — the homeowner said "Definitely," and the transfer fired.
One confirmed pest signal (311 / DOH / DOB / DOHMH) → the whole street is at risk → we resolve every callable homeowner on it. Density is the moat.
It fuses partial public-record signals + people-search into a complete, provenance-labeled callable lead no single source could produce.
Every call → Gemini voice-to-voice analysis → namespaced learnings → injected into the next prompt. Plus a judge-scored sparring arena vs AI prospect personas.
Login for Luka + buyer: all leads, one-click "Call with Brooke," live outcomes, an Advancements/changelog tab, and a DNC view with full transcripts + recordings + voice-intelligence per lead. TCPA opt-outs enforced.
Opt-In/Signal Scanner · Brooke AI Caller · IVR + DTMF (Press 1/2/3, warm transfer, DNC) · Pipeline API · Call QA + auto-train loop · per-transfer 50/50 revenue tracking · TCPA engine · FIRE scoring.
The rare combination: someone who can build it, someone who can distribute it across 50 countries, and a paying operator already buying the output.
Built the entire platform — fusion engine, FIRE scorer, Brooke's voice agent, the self-improving loop, the members dashboard. Full-stack (Next.js, Supabase, VPS, AI automation). Track record: $8.5M revenue across 5 clients, 10x build speed. Full-time.
Capital partner + the distribution moat: 50+ MLM leaders across 50 countries, built over 15 years, each 1,000–50,000 followers → $0-CAC affiliate army. Slovenian Government COVID partnership (saved 50% taxpayer money, 100% success).
A $7M/yr operator already buying transfers on the 50/50 pay-per-call deal. Brings the dialer/routing platform, the buyer demand, and 200 transfers/day of stated capacity — revenue has a real destination on Day 1.
We're raising $2M across 5 milestone-gated rounds, opening with a $200K Seed at a $2M pre-money (10% equity). Capital is released only as each milestone is proven — the investor's downside is structurally capped.
| Product (auth, billing, Stripe) | $40K |
| Clippers (10–15 creators) | $35K |
| Meta + TikTok test | $30K |
| API credits | $25K |
| Luka stipend (3 mo) | $20K |
| Affiliate launch | $20K |
| Infra + DevOps | $15K |
| Legal + contingency | $15K |
| Total | $200K |
500 paying users, $45K MRR, positive unit economics — proven before another dollar is asked for.
Worst case: $75K max loss, kill-switch at Month 4 if MRR isn't tracking. The product already exists and the buyer is already live, so seed capital buys distribution, not invention.
The market: $127B Sales Intelligence TAM, growing 15% YoY. Buyers over-pay for cold, stale lists. We sell hot, fused, verified, called.
The wedge is pest, the engine is universal. Only the keyword set changes — home insurance ($40–80), auto ($20–40), health ($30–100), recruiting, agencies, coaches.
The moat compounds: data flywheel, infected-street density model, provenance-labeled fusion, Klemen's 50-country distribution, AI-first cost structure.
On Sales Intelligence comps of 4–16x ARR (Apollo, Clay, ZoomInfo).
A viable $2–5M/yr bootstrap business with a credible $15–25M acquisition exit in 3–4 years — and a real shot at far more if the network and the engine compound. We plan for conservative, and we're built to celebrate moderate. Exit paths: lifestyle ($15K–100K/mo each) · Series A at $35–80M (M12–18) · acquisition to Apollo/ZoomInfo at 4–8x ARR (M18–36) · IPO track at $100M+ ARR.
We're past the "will it work" question — Brooke has already called, qualified, and transferred a live homeowner. The next steps are about volume, not validation.
Live today: Data fusion engine · 219 callable leads in the CRM · Brooke proven on a recorded warm-transfer · self-improving train loop · members dashboard with TCPA enforcement · 50/50 revenue tracking.
Dialer platform — API/number for warm-transfer routing
Target geo markets — which cities first
Qualification criteria — what defines a "good" lead
Entity for 10DLC — EIN + company name for telephony compliance
Legal (LLC + operating agreement) · ship self-serve MVP ("already 70% built")
First 3 clipper contracts ($5K) · Klemen activates first 10 affiliates
Launch $5K ad test
First paying users → Gate 1 check (500 trials, 3+ clippers live)