Tell us your idea

VeriTrust · an insights apps application for car dealerships

Identity verification software development

A dealership hands over a car against a license, a signature and a phone call. VeriTrust verifies who is on the other side across twelve layers and boils them down to a single number. insights apps built it as its own product.

OCR + AAMVAAnti-spoof livenessSIM-swapPlaidGeofenceI-94 / CBP

VeriTrust

12 LAYERS · 1 TRUST SCORE

Tap to open the app and walk through it layer by layer, just as the customer sees it at the counter.

Demo data · not a real person

0 / 11

What it solves

Identity fraud is discovered after the car has already left.

Verifying by hand means looking at a license, trusting a photo and moving on. Every check that's done separately today — the document, the phone, the home address, the account — lives in a different tab, and none of them talk to each other. VeriTrust runs them together, in the same session, and leaves a written trail.

01

The customer opens a link

There's nothing to install. The dealership sends a link and the verification runs on the customer's phone, at the counter or before they arrive.

02

Twelve layers, a single pass

Document, biometrics, phone, home address, bank, location and immigration status. Each layer returns a signal, not an opinion: match or no match, inside or outside the radius, valid or invalid.

03

A number and its reasoning

The engine synthesizes the 37 signals into a Trust Score from 0 to 100, with the why written alongside. On the dealership side, the full report is kept for compliance.

The twelve layers

Each layer returns a signal, not an opinion.

The customer walks through the first nine. The last three run on their own, without anyone touching anything.

  1. 00

    Consent

    Sensitive data · protected use

  2. 01

    License · front

    OCR with per-field confidence

  3. 02

    Back · AAMVA

    PDF417 decode + OCR comparison

  4. 03

    Phone · OTP

    6-digit OTP + SIM-swap check

  5. 04

    Selfie · liveness

    Anti-spoof + match vs. license

  6. 05

    Address

    USPS normalization + VoIP

  7. 06

    Bank reference

    Existence + account holder only

  8. 07

    On-site GPS

    Geofence vs. home address (Haversine)

  9. 08

    I-94 · CBP

    Verified immigration status

  10. 09

    Email forensics

    Domain age, provider reputation, and whether the mailbox was created just for this transaction. It's the only signal that still shows amber in the demo case.

  11. 10

    VeriTrust Network

    Cross-dealer matching: if the same document, phone or device has already shown up in another flagged transaction, the signal travels with it.

  12. 11

    AI Risk Engine

    Normalizes the 37 signals from the previous layers and synthesizes them into a single number, with the reasoning written alongside so a person can review it.

AI RISK ENGINE

37 signals go in. A number and its reasoning come out.

The score doesn't replace the person who decides: it saves them the work of gathering the evidence. That's why the report shows which layer contributed each signal and which one stayed amber, instead of just returning a traffic light.

12
layers in the report
37
normalized signals
0–100
Trust Score, with a configurable threshold

Where it stands

VeriTrust is in development. The twelve-layer flow is built end to end — it's the one you can try above — and each external integration is connected as legal review moves forward: biometrics, financial data and immigration lookups require explicit consent, end-to-end encryption and regulatory framing (BIPA, GLBA, CBP). That's why the app flags, inside each screen, which part is simulated. The data in the walkthrough is demo data.

See the rest of the insights apps applications›

Yours could be next

We build the system for you

If your business needs to know who it's dealing with, we'll build it. Tell us what you need and in 30 minutes we'll leave the call with the scope locked and a delivery date.