How Avant Deploys Behavioral Analytics to Verify Its Customers

Monetizing Digital Intent January 2022 - Explore behavioral analytics' role in helping businesses prevent new account fraud

Stolen identities are available for as little as $8 on the dark web, making new account fraud more commonplace and usernames and passwords an ineffective authentication tool. In the Monetizing Digital Intent Tracker, Paul Zhang, chief technology officer at Avant, discusses how behavioral analytics can provide a potent defense against new account fraud.

Monetizing Digital Intent January 2022 - Explore behavioral analytics' role in helping businesses prevent new account fraud

An influx of stolen identity information has made its way onto the dark web in recent years, meaning new account fraud has become a pervasive threat. Bad actors deploy these purchased, stolen or synthetic identities to open accounts for nefarious purposes, including money laundering and obtaining fraudulent loans.

FinTechs have to be particularly vigilant about new account fraud, as such businesses have seen an uptick in new applicants during the pandemic. Consumers are looking for digital-first experiences, and fraudsters are exploiting this increased demand. Avant, a FinTech founded in 2012 with the aim of improving middle-income consumers’ borrowing experiences, was well prepared for this threat.

“The pandemic has accelerated the demand for our products, especially on the credit card side,” Paul Zhang, co-founder and chief technology officer at Avant, told PYMNTS in a recent interview. “The situation has ultimately accelerated the drive to digital-only experiences across the industry.”

Many tools are available to counter new account fraud, but few are as effective as behavioral analytics. Zhang explained that when combined with biometrics and other security measures in a multilayered system, behavioral analytics can help improve the customer experience while preventing fraud.

Leveraging Behavioral Analytics for Fraud Prevention

Protecting against new account fraud requires obtaining a big-picture view of fraud risk, Zhang said. No single trait can provide an accurate snapshot of an applicant’s risk profile, so a multivariable analysis is essential.

“We leverage in-house AI modeling to accurately model credit risk, fraud risk and verifications for our customers,” said Zhang. “We use a variety of user-provided information, behavioral metrics and credit metrics to feed our models, and then we leverage our in-house machine learning expertise to accurately model our customers and control fraud and security risk.”

Avant uses both its in-house risk engine and its vendors’ fraud analytics to assess these variables, segmenting customers by risk on this basis. Avant’s machine learning (ML) model adds its own risk assessment to drive down the potential for new account fraud to the maximum extent possible.

“We use behavioral metrics and analytics on a first-party and vendor basis to accurately model our customers and categorize fraud risk,” Zhang explained. “We are then able to use indicators from these data providers as well as some of our first-party captured data to segment the customers [and] allow our ML to perform its own segmentation.”

Avant does not leverage its behavioral analytics systems in a vacuum, however. The company has found that coupling these systems with other analytical tools provides not just fraud protection, but also a host of other benefits specifically for tailoring the customer experience.

Behavioral Analytics’ Multilayered Benefits

Any defensive system must cover the wide range of tactics that bad actors deploy in waging new account fraud. Behavioral analytics, though powerful, is just one tool among many for keeping customers safe.

“We use a variety of approaches, including biometrics and multifactor authentication, and we believe behavioral analytics is a part of our arsenal to fight online fraud, but it isn’t the only approach we take,” said Zhang.

The benefits of this system extend far outside of fraud prevention. Behavioral analytics and other tools can drill down on customer behaviors in normal interactions, enabling staff to improve the customer experience across the board.

“Our ML models for things like marketing response and collections take advantage of a variety of data points, including behavioral [ones], to better segment and model customer behavior,” Zhang explained. “This has yielded benefits across the business beyond just fraud prevention, and behavioral data points are a valuable source of data for the business.”

Preventing new account fraud and other types of malfeasance is a vital part of improving the overall customer experience, but businesses also need to enhance and fine-tune customers’ regular interactions. Behavioral analytics offers one crucial tool for making both happen.