Skip to main content

How It Works

IDV Platform is a managed identity verification service. Integrators drive it through a REST API — there are no biometric models to deploy, no capture infrastructure to run, and no model updates to manage. The platform handles document capture, OCR, face matching, liveness, anti-fraud, and decisioning as one orchestrated pipeline, and runs either as an Innovatrics-managed SaaS or self-hosted on your own infrastructure.

This page explains the whole flow and points to the detail. For the vocabulary it uses — Company, User, Customer, Digital Identity, Trust Factor, Workflow — see Key Concepts.

A verification, end to end​

Every verification runs as a workflow: an ordered sequence of steps the platform executes for one session.

  1. Capture — the user photographs their identity document and takes a selfie with a liveness check, on a phone via the Mobile SDK or in a browser via Web Components; chip documents can add NFC reading. Real-time guidance keeps image quality high. Capture happens on the device — all analysis runs server-side.
  2. Checks — the platform reads and validates the document (OCR, authenticity), compares the selfie to the document portrait and confirms a live person (face & liveness), guards against manipulated or injected media (security), and — where enabled — searches for existing or blocked identities (duplicity check).
  3. Decision — the check results are combined by configurable Trust Factors into one explainable Accept / Review / Reject, and any duplicate hits resolve the identity as unique, merged, or blocked.
  4. Result — the outcome returns to your backend (synchronously or by webhook) and, on the stored tier, is recorded as a Digital Identity on the person's Customer record.

These controls are deliberately layered and independent, so fraud that slips past one is caught by the next — see Multi-layered defence.

What runs, and when​

The step sequence isn't fixed. The platform ships ready-made workflows for the common needs — onboarding, authentication, identification, personal-data update, age verification, document enrolment — and Administrators can adapt them or compose their own from vetted building blocks. Each maps to a business scenario under Use Cases.

Records and their lifecycle​

A person has one Customer record per tenant; each verification attempt adds a Digital Identity, one of which is the Customer's primary identity. How outcomes create, update, merge, or block records — and the states a Digital Identity moves through — is covered in Digital Identity Lifecycle, with the full two-axis state model in Digital Identity States. Results that need a human decision wait in Manual Review.

Integrating, operating, configuring​

Start here​

New to the platform? How to start walks through provisioning, configuration, and a first verification. Accuracy benchmarks and compliance are listed in Standards & Certifications.