Before smartphones, high-speed internet, and a pandemic shifted the world to digital-first onboarding, banks, financial institutions, and other businesses could confirm identities by comparing a face in person to a physical photo ID. Of course, the verification could not be taken fully at face value—no pun intended. The identity still needed to be validated by matching unique identifiers, such as name, date of birth, and address, against independent credit header and utility data using a back-end system.
So how does document verification work in a digitally transformed world?
Document Verification Process: Verifying Physical Documents and Attributes Digitally
Document verification is the process of verifying the authenticity of a document. Officially issued documents, such as a driver’s license, bank statement, or other state/federal documents, are usually the most accepted, verifiable documents. The process authenticates document features, such as stamps, watermarks, fonts and carrier materials. Further, PII data from the front of the ID is matched against the same data extracted from the machine readable zone (MRZ) on the back of the document.
With digital onboarding, applicants no longer need to be physically present for their documents to be authenticated or to match their face to the photo ID. Using an image capture app on a smartphone, the applicant simply uploads a photograph of their government-issued ID and a selfie.
How Does Digital Identity Verification Work?
The key to confirming physical document authenticity and facial comparison in a digital world is to rely on automation powered by machine learning and artificial intelligence. That means the document and selfie can be put through hundreds of validation checks in seconds versus an in-person glance by an untrained eye. The advanced image capture app guides the consumer on how to take high-quality ID document and selfie photos by eliminating blur, glare, and other issues, so there’s a better likelihood of a good user passing security checks on the first try.
- Face and orientation detection
- Edge detection and cropping
- Front to back data correlation
- Cross checks on thousands of global IDs
- Colorspace analysis
- Headshot integrity
- Selfie to ID photo match
- Selfie liveness detection
This automated, machine learning-driven approach is much more effective than a manual review because it spots forgeries and spoofing with greater accuracy and speed, while also boosting auto-decisioning by reducing false positives.
How Machine Learning Benefits Document Verification
Machine learning (ML) techniques depend on big datasets to gain high accuracy, and training data size is one of the main determinants of a model’s predictive power. Machine learning models rely on such data to train itself and learn continually, compounding performance. Socure’s turbo-charged machine learning classification is trained on 700 million “known good” and “known bad” identities and reinforced with third-party data from more than 400 sources. Here are the benefits that machine learning brings to document verification:
- Better forgery and spoofing detection
- Scalable to any degree
- No latency from volume spikes
- Highly predictive of fraud
- Data-driven decisioning
- Faster onboarding
Meeting User Expectations in a Digital Technology Landscape
Today’s consumers expect streamlined, low friction onboarding for immediate account access. As document verification becomes more mainstream, a requirement for head-turning video or multiple selfies creates a frustrating situation that threatens abandonment and negatively impacts your brand’s reputation. At a time when 84% of consumers say that a good onboarding experience is as important as a company’s products and services, businesses must ensure a modern approach to document verification to meet user expectations.
Low Friction and Automated Decisioning in Seconds
Socure’s Predictive DocV delivers a holistic, machine learning-powered decision in seconds with 98% accuracy using hundreds of multi-dimensional predictive signals on top of document authentication, liveness detection, and enhanced facial biometrics to identify more good customers and eliminate fraudsters in real time.
See Predictive DocV in action for yourself by scheduling a demo.
Brenda Gilpatrick is senior director of product marketing at Socure. She helps to lead go-to-market strategies for Predictive DocV. Previously, she was an independent consultant in the payments and fintech industry, working with companies of all sizes on marketing, technology, operations, and business development initiatives.
The What and the How of Document Verification
Before smartphones, high-speed internet, and a pandemic shifted the world to...
The Key Takeaway from Bitcoin 2022 Industry Day:...
The biggest event in crypto, Bitcoin 2022, kicked off yesterday in...