Safe, an established provider of digital identity and fraud verification solutions, introduced the âfirstâ predictive document and identity verification solution with native signals and intelligence on fraud risks.
Predictive verification of Socure documents intended to go beyond simply authenticating a government issued ID and matching a selfie with the photo on the ID. It aims to predict “whether the identity itself is secure or not by using the real-time data link for better accuracy and real-time identity trust decisions”.
With Socure’s predictive document verification service, hundreds of validation checks can be performed on every document and selfie “resulting in unmatched automatic decision rates and risk insight”. The ad also mentioned that its multidimensional predictive signals “inform ML-driven decisions to identify more good customers and eliminate scammers in real time.”
Alex Faivusovich, fraud prevention manager at Lili, said:
âSocure’s predictive document verification helps us onboard new customers quickly and scale faster than with other tools. It is the only solution that offers a single, comprehensive view of digital identity, not only document authenticity and match, but also actual risk and fraud decisions. No other solution compares in terms of accuracy or information. With DocV, we gain confidence knowing that we are dealing with consumers who are who they say they are and will not put our business at risk.
Socure Predictive Document Verification’s advanced image capture “delivers up to 98% automatic decision-making rate in seconds,” the announcement revealed.
Meanwhile, less sophisticated image capture tools “only result in 65-70% automatic decision-making,” the update notes.
Socure’s Predictive Document Verification was also designed to thwart ‘impersonation’ attacks with 99.5% accuracy using NIST PAD Level 2 vividness detection and ‘enhanced biometrics to match selfie to the photo on the ID â.
The liveliness detection checks “work passively in the background, so scammers can’t detect what’s going on,” the announcement explained. No user blink or head rotation is required, dramatically reducing consumer friction. Unlike other tools, a single selfie image “fulfills both facial recognition and alertness detection,” the statement noted.
David Mattei, senior analyst at Aite-Novarica Group, said:
âThe use of document identification and verification solutions to verify a user’s identity is growing due to the increase in digital interactions online. Socure’s predictive document verification product goes beyond traditional document validation and photo-selfie correspondence. It examines other risk attributes such as device fingerprints, phone number, and other digital identity signals to provide a more comprehensive user risk profile.
Using proprietary computer vision technology, Socure’s predictive document verification “extracts the PII document and verifies it by linking it to authoritative data sources, such as credit header, utility, telecommunications records, etc.
From there, device, phone, and address data is also “collected and evaluated during the onboarding process, and then Socure derives predictive signals like phone-to-name correlation, address correlation. -name, device intelligence and IP distance from the physical address, all of which are almost impossible to fabricate, âthe statement said.
The merging of these predictive signals “informs a final decision (pass, reject, or resubmit), blocking risky attempts to enter a customer’s ecosystem while helping them onboard more legitimate customers faster, or identify good existing users with less friction. “
The announcement further noted that hundreds of predictive signals “incorporated into the document verification and biometric matching process constitute the industry’s only digital identity continuum measuring fraud risk.”
Socure’s predictive document verification was developed to use machine learning classification models “trained with 530 million known good and bad identities and over 400 offline / online data sources to assess risk.”
As very large amounts of data are critical to improving the overall accuracy of Socure’s predictive document verification, Socure has created a continuous improvement process “using feedback data from a wide range of customers representing multiple industries” .
Integrating hundreds of unique highly predictive features with identity and device components that can indicate fraud intent to generate more predictive, data-driven results is also essential for Socure’s fraud detection engine.
Johnny ayers, founder and CEO of Socure, said:
âOlder vendors who specialize only in verifying identity documents are unable to build in transparent risk controls because they don’t have historical data, the machine learning engine, or thousands of signals. predictive to provide a nearly perfect fraud classification. This means their customers either miss critical information or end up with disparate products that are costly to deploy and maintain because they are loaded with false positives.
âWorse yet, the traditional approach still leaves gaps in determining whether it is safe to do business with a consumer and predicting the likelihood that that person will commit fraud in the future while frequently requiring multiple document capture attempts. With Socure’s predictive document verification, we offer a unique solution that takes the guesswork out of it, with the highest degree of accuracy and customer experience.