Diversified Members Credit Union has picked Scienaptic AI to go after identity theft and synthetic identity fraud right where it often starts: the loan application. The announcement came from Scienaptic AI today.
What the Credit Union Is Building
The plan is a fraud detection and loan underwriting engine that works as one system rather than two departments passing notes. It pulls in data from multiple sources, then runs machine learning and anomaly detection algorithms over it. Anomaly detection simply means flagging applications that don’t behave the way real ones do. Together, those checks produce a strong signal about whether an application is legitimate.
Timing matters here. Scienaptic says early fraud detection is already producing significant savings in fraud losses for its existing clients.
Why Synthetic Identities Are a Headache
Synthetic identity fraud blends real and invented details into a person who doesn’t exist. That makes it harder to spot than a straightforward stolen identity. Catching it at origination, before money goes out the door, is far cheaper than chasing it afterward.
What the Leaders Said
Sati Smith, President and CEO of Diversified Members Credit Union, framed it as a member-protection issue. “Protecting our members’ financial well-being is at the core of what we do, especially in the face of rapidly evolving fraud tactics,” she said. She added that the technology gives the credit union data-driven tools to catch sophisticated threats early, while still delivering secure, instant decisions and a seamless, safe experience.
Courtney Mrsny, Chief Customer Officer at Scienaptic AI, pointed to the balancing act every modern credit union faces: staying secure without making members jump through hoops. She said the platform is built to give institutions like Diversified the AI-powered signals needed to detect fraud early and accurately, and to protect members “with absolute confidence.”
A Practical Takeaway
If you work at a lender, look at where your fraud checks and your underwriting actually meet. If they live in separate systems, ask how much suspicious activity slips through the gap, and whether you could flag it before approval instead of after.