Sigma v4: Socure's Groundbreaking Graph-Powered Identity Intelligence

Socure is excited to announce Sigma V4, an end-to-end fraud solution. For many years, customers used a combination of separate email, phone, IP, and device vendors. Socure has brought together all of these dimensions of a consumer’s identity into a single machine learning model, something that no one else in the market can provide today.

This ability to feature engineer more than 30,000 variables across every single dimension whether it be good, bad, high-risk, or low-risk, and apply billions of rows of known outcomes is a one-of-one solution.

Sigma v4 is the first fully integrated in-house developed solution that fuses data points from hundreds of global sources with our network data, velocity, device, and behavioral data into a single machine learning model that delivers results in less than 150 milliseconds with five nines uptime.

There are many ways that this new solution will drive tremendous value for our customers. Through Entity Profiler, we have a unique persistent view of a device. We can build an identity graph that tracks attacks at scale across 2,100 customers.

With Sigma v4, we’ve also integrated real-time anomaly detection. This allows us to quickly pinpoint if there are any unusual patterns coming through as part of an application.

We are seeing attackers with a multitude of stolen identities who are trying many different types of schemes. The Sigma Identity Fraud Suite and Sigma v4 are able to stay multiple steps ahead of these fraudsters by studying the patterns of how an identity evolves over time:]

• Do they have an iOS or an Android?
• Do they use English as their primary language?
• Do they use Turkish?

To stay ahead of these new types of attack vectors, we constantly study identities as well as fraud attacks across the entire graph to be able to determine if there are 2, 3, or 4 standard deviation shifts from what we’ve seen historically.