Posted on — Leave a comment

What We Chose Not to Build Into ID Verification

Most identity-verification platforms quietly collect more than they advertise: IP address, device fingerprint, browser fingerprint, even behavioral telemetry like how long you hesitated filling out a form. It’s standard practice, and there’s a real fraud-detection argument for it.

We looked hard at adding it to NEXIOS AML/KYC’s remote verification flow — the one your customer uses from their own phone, from wherever they are, to submit an ID photo and a selfie. We decided against it.

That flow exists specifically because the person being verified is remote and never physically present — the exact context with the most real legal exposure around biometric data and consent. Our protocol makes an explicit promise: no device identifiers, no IP, no geolocation, nothing beyond the document and the selfie someone actually agreed to submit. Adding invisible tracking on top, even well-intentioned, would mean quietly breaking that promise.

We kept what already does real fraud-detection work without that tradeoff: on-device liveness detection that catches a photo-of-a-photo or a virtual-camera replay, MRZ and barcode cross-checks against what the customer typed, a tamper-evident hash-chained audit log, and sanctions/PEP watchlist screening with a staleness policy that actually gates the verdict.

We also just shipped a real partner API — bearer-token auth scoped per firm, idempotent writes, the same audit trail, self-service token management, no engineering ticket required.

None of this is a pitch dressed up as a principle. It’s what we actually built, and the tradeoff we actually chose, when a faster path was sitting right there.

— Norman Fuchs, NEXIOS

Leave a Reply

Your email address will not be published. Required fields are marked *