Verify new customers in one sitting

Data and DVS checks return in seconds. A full ID Pass flow (photograph the document, take a selfie, run the checks) typically takes the customer under two minutes end to end.

Speed is the whole argument for verifying at onboarding instead of afterwards. Applicants asked to come back later frequently do not come back at all, and a verification step that finishes inside the session is the difference between a completed sign-up and an abandoned one.

Four steps, run in order of increasing friction, with each one only reached if the previous left doubt:

• Pre-screen — name, date of birth and address checked against Australian data sources, with no document required
• Document check — the identity document verified against the issuing authority through the DVS
• Face match and liveness — the document photo compared to a selfie, confirming a live person is present
• PEP and sanctions screening — the onboarding watchlist check

Most customers complete the sequence without noticing the first step happened. Consent is recorded at each stage, not collected once at the start.

Start with a document-free data check and escalate only on doubt.

A pre-screen against name, date of birth and address resolves the large majority of applicants invisibly: no document photograph, no selfie, no waiting. Those customers never see a second step, because there was never a reason to ask them for one.

Only mismatches escalate to document and selfie checks. The effect is that friction is concentrated on the applications that warrant it, instead of spread evenly across everyone as a precaution.

Spreading it evenly is the more common design and the more expensive one: it costs you completed sign-ups from exactly the customers you were trying to attract.

You decide, and because failures carry per-field results, you can decide intelligently.

The common designs are to retry with a different document type, escalate to a biometric step, or route the application to manual review. Which one applies is a policy question, and it usually depends on which field failed.

That detail matters because most first-time failures are not fraud. A superseded address, a name recorded differently on the document than in your form, or a mistyped licence number all produce a non-match against a completely genuine identity. Per-field results let you tell those apart from an application where nothing resolves at all.

Yes. ID Pass runs entirely on the customer's own device from a link you send, and there is no app to install.

The customer photographs their identity document and takes a selfie in their mobile browser. OCR extracts the document fields, the DVS check runs against the issuing authority, face matching confirms the selfie belongs to the document, and the result returns to your workflow.

For customers who are physically with your staff, IDFEX ID Check covers the assisted case: your team runs the same document and data checks from a browser while the customer is present.

The checks verify identity against the DVS and independent Australian data sources, which is the sourcing standard AUSTRAC expects: reliable and independent, not self-reported.

Your AML/CTF program defines which combination applies to each customer type, because a risk-based program is required to make exactly that distinction. There is no single fixed check set that satisfies the obligation for every customer.

One point of sequencing is worth being precise about: reporting entities must verify identity before providing a designated service. Onboarding is not merely a convenient place to run the check; for many services it is the only compliant place to run it.

Consent is confirmed before each check, not collected once at the start of the application and assumed to cover everything after it. DVS checks require it as a matter of law.

What the system stores is your confirmation that consent was held, attached to that specific check and its result. Obtaining and retaining the evidence of consent is your responsibility, and you may be audited and asked to produce it for any DVS check run.

The implication for an onboarding flow is that the consent wording, timestamp and customer record need to live in your own system. Doing it per step also handles the case where a customer consents, completes some checks and abandons at another: the trail then reflects what happened, not what the flow intended.

They still get verified. The flow falls back to document and data verification, which establishes the same identity without a selfie, and assisted checks are available through IDFEX where someone needs help.

This is worth designing deliberately instead of leaving to chance. A customer may have no working camera, poor connectivity, or a disability that makes the capture difficult, and none of those are reasons to decline an application.

Treating a failed biometric capture as a routing decision, not a rejection, keeps legitimate customers from being turned away by a technical limitation.