Purpose-Built AI Platform Performs Deep Analysis of Consumers’ Online and Offline Footprints to Validate Authenticity of Digital Identities with 95% Accuracy
NEW YORK, July 19, 2018 -- Socure, a leading provider of predictive analytics for digital identity verification, today announced Aida (Authentic Identity Agent), the first and only purpose-built bot for establishing trust in online transactions. Named in honor of Ada Lovelace, the world’s first computer scientist, Aida uses artificial intelligence to process billions of multi-dimensional online and offline data points per second to validate the authenticity of digital identities in real-time.
According to the Javelin Strategy & Research 2017 Identity Fraud Study, due to the increasing adoption in the US of EMV (chip) cards and terminals in point of sale environments, fraudsters have increasingly shifted to fraudulently opening accounts and become better at evading detection.
As the number of data breaches has skyrocketed in recent years, personally identifiable data has flooded the black market, making it easier for fraudsters to impersonate real identities or to create fake ones using real data. The reliability of identity data has become a key concern across industries, especially within the financial sector, which often depends on manual review to confirm the identity of new applicants.
AI-based Decisioning for Identity
By providing a multidimensional view of consumers based on applying self-training, predictive analytics models to hundreds of online and offline data sources, Aida enables financial services organizations to automatically approve more digital transactions than previously possible without performing manual reviews. It can reduce fraud for online new account opening by up to 90 percent.
Levels of identity
Using the Aida-powered Socure ID+ platform, financial services organizations have achieved the following significant business results:
1. Top 10 Credit Issuer reduced fraud by 85%, saving more than $50M in fraud losses
2. Top 5 Bank reduced their dependency on knowledge-based authentication (KBA) by 70%, which substantially increased
auto- acceptance rates and improved the customer experience
3. Leading Digital Bank completely eliminated KBA, while increasing annual revenues by $5M
Today at 1:00 pm Eastern Daylight Time, Sunil Madhu, Chief Strategy Officer for Socure will join PYMNTS CEO Karen Webster for a live Digital Discussion on using robots like Aida to fix digital identity verification problems.
“Socure is solving the single most difficult problem in identity verification – validating a person that’s never done business with an organization before,” said Sunil Madhu, Chief Strategy Officer for Socure. “Using traditional approaches for vetting the identity of new customers in a mobile and digital world has been a miserable failure. Aida can assess in real-time and with unprecedented levels of reliability, whether a digital identity is authentic, synthetic or has been stolen by performing beyond-human analysis at machine speed. Aida essentially lives every minute of every day to verify identities and fight fraud.”
Aida learns customer identity from their digital footprints to calculate risk and correlation scores, which empowers businesses to dramatically increase online transaction acceptance rates as well as reduce manual reviews and fraud. In the future it could provide consumers with a portable “Socure Verified” identity that can be used with participating financial institutions and ecommerce merchants.
The Brains in the Machine
Aida combines artificial intelligence, unsupervised machine learning and clustering algorithms to perform a continuous loop of:
- Ingesting, normalizing and evaluating data from hundreds of online and offline data sources including credit bureaus, email history, phone records, IP addresses, social networks and more
- Automatically generating fully explainable and transparent machine learning models in hours, while continuously training and improving them. Aida performs the work of human data scientists in software, at scale, to achieve accuracy levels far beyond traditional human rules-based approaches
- Performing predictive analytics on real-time transactions to assess and assign a risk score to identities, which is used to determine whether a request should be auto-accepted or flagged for manual review by a fraud analyst
Topics: Machine Learning