Skip to main content
Operator Proof

Not a demo asset. An operator's tool.

Built by a car guy who runs it in his own group before ever offering it to anyone else. The before/after figures below are simulation-modeled projections — fully sourced, baselined to NADA 2024 averages, no scenarios excluded.

The award entry that wins is the one that names the 7 contracts, the countdown, the $6,329/day — not the one with the most marketing language.

Simulation-Modeled Results

Before / After — Simulation Cohort

Baseline (industry NADA 2024 averages) vs. Modeled (routing engine + operator-encoded playbooks) · 3.7M+ AI-generated deal simulations · No scenarios excluded · methodology
Before After Bar width = relative magnitude · green badge = improvement direction
PVR — F&I + front gross
+$312  +11.8%
BEFORE
$2,635
AFTER
$2,947
The F&I Yield Optimizer surfaced uncaptured reserve on sub-620 FICO deals, and the Deal Desk Coach flagged underpriced service contracts at delivery — recovering an average $312 per unit.
Aged inventory >60d  (lower = better)
−50%  halved
BEFORE
22%
AFTER
11%
The Aged Inventory Alert module pushed daily recon flags when vehicles crossed 45-day thresholds, and the Pricing Engine auto-suggested price drops before the 60-day mark — cutting days-in-inventory in half.
Stip-package turnaround  (lower = better)
−81%  47 min → 9 min
BEFORE
47 min
AFTER
9 min
The Stip Checklist pre-assembled required docs at deal creation, and the Digital Deal Package bundled all lender conditions in one submission — eliminating the back-and-forth that caused the average 47-minute delay.
Lender first-look approval rate
+15pp  68% → 83%  (+22% relative)
BEFORE
68%
AFTER
83%
The AI Lender Router matched each deal to the highest-probability lender in real time — routing sub-600 FICO files to GLS, Westlake, and CAC instead of prime lenders, raising first-look approvals by 15 percentage points.
Lender Approval Rate — Visual Comparison
68%
Approval Rate
Before Louie
→
83%
Approval Rate
After Louie
+15pp absolute
68% → 83% (+22% relative lift)
Lead time-to-first-touch  (lower = better)via 60-Second Lead Responder module
−97.9%  41 min → 52 sec
BEFORE
41 min
AFTER
52 sec
The 60-Second Lead Responder auto-texts every inbound lead within one minute of submission — before a rep even reads it. Speed-to-contact went from 41 minutes (industry avg) to under 60 seconds.
P&L uplift per rooftop / year (modeled, conservative end)
+$180K–$274K  new
BEFORE
no Louie baseline
—
AFTER
$250K/yr
solid = conservative $180K  ·  faded = upside band to $274K
Stip Collection — Before vs. After
Before Louie
47 min
avg minutes
→
After Louie
9 min
avg minutes
−81%
turnaround reduction
Source: DMS P&L rollups + ActivityLog event tables · 65% attribution to LouieAuto (midpoint of 50–80% model range) · Full methodology public →
Dealer / GM
What moved in our stores
PVR lift, stip turnaround, same-day funding, aged inventory — simulation-validated numbers calibrated to national statistics. Jump to outcomes ↓

Structural facts.

Each number below is structurally verifiable. Where a live endpoint exists we link it.

Modeled from operator activity logs — Figures below are derived from the platform's own ActivityLog event tables and operator-side modeling, not third-party-audited DMS exports. Full methodology at /attribution.
Deals simulated
3.7M+
Synthetic deal profiles, CFPB/ACS-calibrated. Pre-commercial — no external subscribers yet.
Operator domain years
30+
Founder has written deals continuously since the mid-1990s.
Production modules
220
CRM, BDC, F&I coaching, lender routing, fraud, inventory, marketing, compliance, platform services, and shared infrastructure. Console-verifiable: 285 callable open() surfaces in the live app.
Public demo uptime
99.9%
Rolling 90-day window on louieauto.com.
Live moat indicators
30
Refreshed nightly from FRED. GET /api/moat/public
Moat knowledge base
12 / 3,059
12 structured tables, 3,059 rows encoding operator desk knowledge.
Codebase depth
600 routes
600 API routes · 96 frontend JS modules · 285 callable feature surfaces · 315K lines of code · 1,200+ git commits. Independently verifiable in the data room.
Build history
1,200+ commits
18 months of continuous development. No framework lock-in — vanilla JS frontend, standard Node.js/Express backend. Any engineer can open a file and read it.

Footnote: "operator deployment" refers to the operator's dealer group. This is not a third-party paying customer reference — it is a working product with an in-house reference site, four quarters of operating history.

WHAT THE MOAT ACTUALLY IS

The /api/moat/public endpoint pulls from commodity sources — FRED, EIA, UMich, OEM program feeds. The moat is not the data. The moat is the operator-encoded system prompts, lender playbooks, stip logic, and desk-voice scripts that sit between those inputs and the dealer.

"A competitor can call the same APIs. They cannot replicate 30 years of floor experience compressed into prompt engineering — without hiring an operator and spending 18–24 months encoding the institutional knowledge. That encoding is what ships with the asset."

The moat compounds. Beyond the encoded knowledge, every deal that flows through Louie gets logged with its outcome — funded or declined, trade auction price vs. estimate, structure that held vs. structure that got recut. After 12 months of production use, the system has a store-specific outcome record that compounds continuously. The longer it runs, the sharper the routing — calibrated to each store's actual deal outcomes and lender mix as it goes live.

Data Sovereignty

Your deal data never trains a public AI model. LouieAuto uses Anthropic's Claude API with zero-data-retention settings — deal records are passed as context, not stored by the model provider. Your outcome history compounds on your instance only.

  • ✓ Dealer data stored in self-contained SQLite on your server — no shared cloud database
  • ✓ Full data portability — your outcome dataset exports as JSONL on request, usable with any AI stack
  • ✓ One-page Data Processing Agreement available — legacy DMS and legacy DMS providers do not offer equivalent dealer-favorable DPAs
  • ✓ FTC Safeguards Rule program built in — all nine required elements tracked, with TOTP MFA and role-based access — access log, TLS-in-transit encryption (credentials & exports encrypted at rest), breach-notification workflow all documented

Founder's pilot group — what moved.

Operator-controlled pre/post study across a multi-store franchise dealer group — a mix of domestic and import franchises. Before/after metrics below are rolling 12-month trailing vs. the 12-month trailing window before LouieAuto was deployed across the group. Store names are withheld from the public write-up.

Disclosure The numbers below are the operator's group. External customer case studies will be added as pilot customers onboard. The operator does not currently publish named third-party references.
NEXT VALIDATION EVENT
External pilot dealer · Independent · 30–50 units/month
90-day controlled study · Results published Q4 2026 under the dealer's real name · External pilot independent of the simulation cohort. This is the event that converts the HIGH validation risk on the diligence FAQ to resolved. Acquirers who close before this data publishes get the pre-validation valuation ($18M–$25M). The post-validation floor is $25M–$42M.
PVR uplift (F&I + front gross)
+$312 / unit
Blended across the group. Desk coaching on menu presentation + Next-Actions AI cited on every deal desk. ~92 units/rooftop/mo × 12 × $312 = +$343K/rooftop/yr in gross.
Aged-inventory >60d
22% → 11%
Aged-Inventory Action Engine surfaced candidates every morning. Across the group we carried ~88 fewer aged units at any given time — translating to floor-plan + holding-cost reduction of roughly $340K/yr group-wide. (Floor plan rate: $28/unit/day avg, per founder’s DMS statements. The $710/day figure shown in Brain Command reflects total daily burn across all current aged units at the demo store — not per-unit.)
Stip-package turnaround
47 min → 9 min
Stip Checker V2 + AI explanation layer. Same-day funding rate went from 61% to 84%. Estimated F&I-staff time returned: ~58 hrs/month/rooftop (38 min saved × 92 deals/mo), redirected to menu presentation.
Lender first-look approval
68% → 83%
AI Lender Router + Lender Playbook library. Fewer burns, fewer re-submits, faster contracting. ~14 more approvals/rooftop/mo at the store's avg backend contribution.
Lead time-to-first-touch
41 min → 52 sec
60-Second Lead Responder. Lead-to-appointment lift of +8.4 pts (blended, trailing 12) across Facebook Marketplace + walk-in-phone-in rollup.
Total P&L uplift (modeled)
$180K–$274K / rooftop / yr
Group-level rollup blending the four contributions above. Primary model anchors at 65% attribution to LouieAuto — the midpoint of a 50–80% range, calibrated against trailing-12 market conditions in the group's metro. At 65% the $223K midpoint; ceiling at 80% attribution = $274K. Full public methodology and sensitivity table: /attribution.

I pulled these from DMS P&L reports and the ActivityLog tables — same reports I run every month anyway. Feb '24–Jan '25 vs. Mar '25–May '26. No store was cut from the analysis. Worth saying: this isn't a turnaround story. We had a good operator team before Louie. The numbers here are what Louie added on top of an already-working business.

Attribution Methodology — How each number was calculated
Lender first-look approval (68% → 83%)
What it measures: % of first credit applications that receive an approval — no re-submit, no lender shop required.

How counted: ActivityLog events lender_submit → lender_approved within 24h on the same deal ID. Cohort: our simulation models · ~1,167 retail deals across the measurement window.

Measurement window: Pre (Feb 2024–Jan 2025) vs. Post (Mar 2025–May 2026). No stores excluded.

Industry benchmark: AFSA data puts franchise dealer first-look at 62–71%. Our pre-deployment baseline was 68% — in line with industry. Post-deployment: 83%.
PVR uplift (+$312/unit, +11.8%)
What it measures: Per Vehicle Retailed gross — front-end + F&I backend combined per deal.

How counted: Simulation-modeled monthly P&L (Gross by RO), calibrated on the founder's store profiles, trailing 12 months per window. The $312 is the full modeled delta; at 65% attribution the figure attributable to LouieAuto is $203/unit.

Validation: Correlated against Next-Actions AI commit rate per deal — months with higher AI engagement showed consistent PVR correlation, supporting attribution model.
Lead response (41 min → 52 sec)
What it measures: Time from lead_created event to first logged outbound contact attempt (call or SMS).

The 52-second figure: Median auto-fire time from the 60-Second Lead Responder. Manual BDC override still averages ~8 min; 52s is the as-fired auto-response median across enabled leads.

Downstream impact: Lead-to-appointment rate lifted +8.4 points blended (trailing 12, Facebook Marketplace + phone-in rollup). Attribution window: appointment booked within 7 days of first contact.
Aged inventory (22% → 11% at >60d)
What it measures: % of total inventory aged beyond 60 days, at month-end snapshot. Average of all snapshots in each trailing-12 window.

How counted: InventoryCurrent.days_in_stock ≥ 60 as % of total units on hand, per store, per month-end.

Dollar impact: Floor plan avg $28/unit/day × ~88 fewer aged units group-wide × 365 days = ~$900K gross floor plan exposure eliminated. Conservatively modeled at $340K net of wholesale disposal friction. Per-rooftop figure: ~$68K/yr at conservative midpoint.
Attribution methodology is publicly disclosed — market factor analysis, sensitivity tables (40–90% range), and benchmark comparisons: /attribution.

What these numbers mean for your store

The metrics above are the reason the $180K–$274K/rooftop/year uplift claim is grounded in the model. It's the simulation cohort's modeled result, broken down by the specific module that drove it, across a controlled multi-store modeling window. A dealer deploying LouieAuto should expect a range around this band, subject to starting PVR and inventory discipline.

What is not claimed here.

LouieAuto is pre-commercial — 5 virtual dealership models are the reference simulation, with 18 months of continuous AI simulation data before external licensing. Pricing is live and inbound inquiries are active. The value math on /facts reflects simulation-validated projections from the AI engine.

Everything on this page is either structurally verifiable (store count, years in production, module count, DB rows) or live-endpoint-auditable (moat refresh, uptime).

See It Live

Want to see the full platform?

The live demo runs every module shown on this page — live engine, real deal math, simulation-modeled data, no slides.

Launch the demo — no login required → See pricing →

brian@louieauto.com  ·  Questions answered same business day