System Overview — Vianney Godbout
Your Lead Machine,
End to End
This system is built in layers. Each phase unlocks the next — nothing is throwaway, everything compounds. Here's what's done, what's next, and what each piece is worth.
The Math Behind Everything Below
Every phase below is measured against that: does it make the next deal more likely, faster, or more automated?
Phase 1 — Foundation
CompleteWhat this is
The complete scoring and outreach engine. Takes raw MLS data, identifies who's most likely to need help, generates personalized outreach, and runs a follow-up sequence — all automatically. This is the machine everything else plugs into.
The 8-step pipeline
Lead Ingestion
LiveMatrix MLS expired/withdrawn listings parsed from Gmail alert emails. Property data, agent remarks, financial details, listing history — all extracted automatically.
Contact Enrichment
Needs ConfigWaterfall lookup: Evalweb (tax rolls) → Apollo (contact database) → Direct mail fallback. Code is built — needs API activation.
Scoring Engine (v3)
Live25 distress keywords from agent remarks + financial signals (price drops, tax liens, estate sales) + listing behavior (days on market, multiple expirations). 173 leads scored.
Tiered Routing
LiveMessage Generation
LivePersonalized Quebec French messages based on each lead's situation. Honest, observational, no fake urgency.
Outreach — SMS + Email + Mail
Needs ConfigSMS via Twilio (Day 1), Email via Resend (Day 3), Direct mail letters (always). Digital channels need account activation.
Follow-Up Sequence
LiveDay 1 SMS → Day 3 email → Day 7 follow-up → Day 14 last touch. Automated, spaced out, not spammy.
Response Detection
LiveSeller replies → sequence stops → you get an SMS alert. This is when a lead becomes “hot” — from their action, not a score.
To activate Phase 1 fully
These are config items on your side — the code is already built and waiting:
Everything included
Phase 2 — Contact Intelligence
NextWhy this matters
Right now you have 173 scored leads with addresses — but not names, phones, or emails. That means outreach is limited to direct mail. This phase solves the #1 gap in the system: turning an address into a person you can actually reach digitally.
What gets built
Evalweb scraping layer
Their API gives property data but not owner names. We build a scraping layer on their website to complete the address → owner name connection. This is the missing link that makes Apollo useful.
Apollo search integration
Once we have an owner name, Apollo's API returns phone + email. The enrichment waterfall runs automatically for every new lead. Old leads get retroactively enriched.
Quebec Land Registry fallback
$1/lookup backup when Evalweb can't resolve. Definitive ownership records. System auto-falls to this when primary sources miss.
How it builds on Phase 1
Phase 1's owner_lookup.py already has the enrichment interface built — it accepts an address and returns contact info. Right now it returns empty because the data sources aren't connected. This phase fills in the actual data layer underneath. Once connected, the existing outreach engine (SMS, email, follow-up sequences) activates automatically — no changes needed to the pipeline code.
What this unlocks
Your 173 leads go from mail-only to SMS + email + mail. The follow-up sequence that's already built starts running. Response detection activates. You start getting “hot lead” alerts on your phone. The entire bottom half of the Phase 1 pipeline — which is built but idle — comes alive.
At $15K average commission, this phase pays for itself if it contributes to closing even one of those 173 leads faster.
Phase 3 — FSBO / DuProprio
PlannedWhy this matters
Expired listings are people whose agent relationship already failed. FSBO sellers are people who never wanted an agent to begin with. Different psychology, different message — but often a stronger opportunity because they're actively selling right now and hitting real friction.
What gets built
DuProprio monitoring
Automated scraping of new DuProprio listings matching your criteria — Montreal area, $600K+, condos/duplex/triplex. Runs daily. Contact info (name, phone, email) is already on each listing — no enrichment needed.
FSBO-specific scoring adjustments
Different signals matter for FSBO — days on DuProprio, price changes, property complexity (multi-unit is harder to sell solo). Separate scoring weights, same engine.
FSBO message templates
Same principles (honest, observational, no pressure) but framed around the specific challenges of selling without an agent — pricing accuracy, legal complexity, negotiation, marketing reach.
How it builds on Phase 1 & 2
FSBO leads enter the same pipeline as expired listings — scored, routed, sequenced. The scoring engine gets a new source_type flag that switches scoring weights and message templates. The router, outreach engine, and response detection don't change at all. One new lead source, zero changes to the existing automation.
What this unlocks
An entirely new lead channel that comes with contact info already attached — no enrichment cost, no Apollo dependency. These sellers are actively trying to sell right now. And because the contact info is built into the source, this channel starts generating outreach immediately, even before Phase 2's enrichment is fully configured.
FSBO sellers who convert often become long-term clients. One $600K+ listing at 3-5% commission = $18-30K. The entire phase pays for itself on the first conversion.
Phase 4 — Scale
PlannedWhy this matters
Phases 1-3 are reactive — you're responding to listings that expire or go FSBO. Phase 4 flips it: you proactively build audiences of homeowners matching specific criteria and reach them before they even list. This is where the system stops being a tool and becomes a machine.
What gets built
Apollo bulk audience builder
Define filters (geography, property profile, owner demographics, ownership duration) → Apollo returns bulk contact lists. Thousands of potential leads per pull, pre-filtered to your exact ICP.
Geographic expansion
Beyond Montreal: Laval, Monteregie, Laurentians. Each new region gets its own filter profile. The scoring engine already handles any region — this adds the data sources and adjusts messaging for local context.
Cross-reference engine
Apollo leads get cross-referenced with Matrix listing data. Apollo tells us who — Matrix tells us what happened with their property. Leads that show up in both sources get scored higher.
How it builds on Phase 1, 2, & 3
Apollo leads enter the same pipeline — scored, routed, sequenced. But they come with contact info already attached (Apollo provides it), so they skip the enrichment step entirely. The scoring engine gets a new apollo_bulk source type with its own weights. The router, outreach, and response detection don't change. More leads in, same automation out.
What this unlocks
Volume. Phases 1-3 bring in leads as they appear (expired, FSBO). Phase 4 lets you pull thousands of targeted contacts on demand. The pipeline is already built to handle it — this just opens the top of the funnel wider. One filter change can add an entire new market.
If expanding to Laval + Monteregie adds even 2 deals per year at $15K average, that's $30K annual return on a $3K investment. And the data keeps flowing every month.
Ongoing — System Management
Why this matters
The system has moving parts — APIs change, data sources shift, new opportunities appear. The retainer means you never think about the infrastructure. Scoring gets refined based on which leads actually convert. New data sources get plugged in as they become available. The machine gets better over time instead of going stale.
What's included
Pipeline monitoring — uptime, error alerts, data quality checks
Scoring refinement — adjust weights based on actual response data
Message optimization — A/B testing, response rate tracking
Dashboard updates — new views, filters, export features as needed
API maintenance — when Evalweb, Apollo, or Twilio change, the system adapts
New data sources — plugged in as they become available in Quebec
$750/month is less than one hour of commission on a single deal. The retainer keeps the machine running and improving — so you focus on conversations, not infrastructure.
Full Roadmap
What a Typical Day Looks Like
Once everything is activated — you don't touch anything until someone responds.
Daemon wakes up. Checks Gmail for new Matrix alerts. Checks DuProprio for new FSBO listings. Parses everything.
Enrichment runs. New leads go through Evalweb → Apollo waterfall. FSBO leads already have contact info.
Scoring + routing. Each lead scored, routed to a tier. Messages generated. Dashboard updated.
Outreach fires. Contact First leads get their Day 1 SMS. Others get emails or letters queued.
Sequence runs. Day 3, 7, 14 follow-ups fire automatically. No intervention needed.
Seller responds. System catches it. Sequence stops. You get an SMS. You take it from here.