It’s been years since I opened a new bank account. The last time was in London. I was required to make two trips to the local bank branch, carrying a stack of documents. When I recently decided to switch banks, I was excited to try a 100% digital onboarding process for the first time. In counties such as Cornwall (where I live) you’d be lucky to find a brick and mortar bank branch within 20 miles. Using the bank’s new customer portal, my smartphone camera, and my UK driving license, I was approved in under 20 minutes. This despite several minutes of high anxiety waiting for the “approval pending” screen to change and confirm my wobbly selfie matched my driving license photo. It seemed like magic.
No doubt you too have experienced that moment of “will this actually work or will I be rejected again?” Fiddling with the overhead light glare and shadows to get that one good shot of your passport while trying to keep it flat with books that don’t ruin the picture. According to a study commissioned by ABBYY, some 40% of users have abandoned a digital onboarding process after an ID verification fail. If that figure is even half-right, a huge amount of potential sales are missing from any business with dodgy digital onboarding.
That’s the top-line value proposition for the growing number of proof of identity (POI) vendors who promise to eliminate those lost sales with their AI-driven solutions to read and validate IDs and other supporting documents. POI-focused startups such as OCR Labs have attracted sizable investments. Meanwhile, long-time intelligent document processing (IDP) vendors like ABBYY have also joined the fray.
Lost sales are not the only motivation driving POI purchases. Fraud detection and prevention also require a strong defense to catch the foxes before they enter the henhouse. This inevitably leads to the classic sales vs. IT tension of “make it strong enough to catch the bad guys but friendly enough to retain the real customers.” POI implementation thus needs to be both user-friendly and painless, or people will continue the abandonment and go to your competitors.
Today we caught up with ABBYY, whose first POI solution has just gone GA. In a sea of existing POI solutions that all claim to read any passport or driving license, what makes ABBYY a unique fish? Their undeniable expertise at reading more than just the ID. ABBYY software is well-qualified for validating those extra supporting documents often required for upload in onboarding processes; documents such as utility bills, bank statements, income verification letters, etc.
ABBYY’s POI solution is available as a cloud service and when it comes to ID validation, there’s no additional training of the AI models needed because the ABBYY POI algorithm is pre-trained on every conceivable ID format. For use cases where supporting documents must be analyzed and validated, customers will install ABBYY Vantage and continue to train the AI models (or “Skills”, as ABBYY calls them) on domain-specific document types. However the customer does not have to train from a zero basis, because the Skills are pre-trained for common documents such as tax forms, invoices, bank statements, etc.
Who is the buyer for this? Surprisingly (at least to us), it’s not the fraud prevention team. ABBYY is finding more traction with the customer experience (CX) team lead. CX definitely has a dog in this hunt: without a great POI piece, there will be no “C”. In ABBYY’s case, we think this may also be because the fraud team are already customers of competitors who offer much more than just the POI piece, selling a complete fraud prevention/KYC platform.
If you’re evaluating proof of identity for your business process, and the use case includes IDs and other document types, it’s worth a look at the ABBYY solution.