Skip to main content
Operator Proof

Not a demo asset. An operator's tool.

Built by a car guy from his own floor experience. There are no measured before/after results on this page: the figures it used to show were simulation projections and the original pilot store's data, and they're withdrawn.

Dealer / GM
Outcome numbers
None published yet — why, and what will produce real ones. 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.
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 — most legacy DMS vendors don't offer an equivalent dealer-favorable DPA
  • ✓ 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

Measured results — not yet.

This section used to show a founder's pilot group before/after (approval rate, PVR, stip time, lead response, aged inventory, P&L uplift). Those figures mixed simulation output with the original pilot store's history and can't be reproduced for a dealer today, so they're withdrawn.

NEXT VALIDATION EVENT
External pilot dealer · controlled 90-day study
Baseline captured before anything changes; results published only under the dealer's real name and with their written approval.

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