$18M Saved in 7 Months: How Glassbox Transforms Fraud Detection
The company
A leading U.S. financial institution dedicated to providing millions of customers nationwide with seamless access to a robust digital banking platform and a full range of financial services — anytime, anywhere and on any device. With a strong commitment to customer security, the retail bank’s mission is to eliminate physical and procedural barriers while making everyday banking simple and secure.
The challenge
With the growing sophistication of online threats, including fraud, the bank began to notice more signs of suspicious activity targeting its customers. It needed a more effective way to detect and prevent these attacks, particularly those that could lead to account takeovers through phishing and other deceptive tactics.
Overview
The retail bank serves millions of clients across individual and business segments. Consumer offerings include checking and savings accounts, various loans, credit cards, wealth management and robust digital banking services. For businesses, the bank provides lending, treasury management, capital markets access and tailored solutions for specific industries.
As digital engagement accelerates, protecting users against sophisticated fraud becomes a top priority.
The Challenge
The retail bank was grappling with detecting and stopping fraudsters in real time. Traditional fraud detection tools were not able to catch remote access scams or identify when the DOM was being secretly altered. The bank needed a way to recognize patterns such as remote control activity and hidden manipulations of what users were viewing.
Two core issues included:
1. Visibility gaps: Existing fraud tools focused on backend data and transactions, leaving blind spots when it came to deceptive client-side activity. Fraudsters could change what the users were seeing on screen, without altering any transaction records. This made the attacks virtually invisible to standard monitoring.
2. Scalability of detection: With high volumes of user sessions occurring daily, manually reviewing session replays for evidence of such manipulation was slow and resource-intensive. Analysts spent hours on a single session, making it impossible to keep up with the volume of potential threats. This rendered traditional detection efforts ineffective and unscalable.
Without the tools to track subtle client-side behaviors in real time or flag attacks early, the bank’s fraud team faced a growing risk of undetected account takeovers and struggled to stop threats before damage occurred.
The Solution
To meet this challenge, the bank is using Glassbox’s AI-powered fraud detection suite, including real-time behavioral analytics and session replay, to monitor and analyze on-screen content the moment it changes. These tools enable the fraud team to instantly detect suspicious behaviors, such as DOM manipulation, extended pauses and erratic navigation as users react to altered account information.
With these insights, the fraud team acts fast, often coordinating with law enforcement to stop fraudulent activity in progress. In some cases, they intercept bad actors while the customer is still in a parking lot or inside a CVS, preventing losses before they occur.
Key benefits include:
Speed and Efficiency: Fraud analysis time drops from hours to seconds.
Pattern Recognition: Analysts spot emerging fraud trends before they become widespread.
Data-Driven Decision Making: Deep behavioral insights enable the fraud team to adapt and train AI models.
The Results
The impact was immediate and measurable:
$18 million saved in fraud losses within seven months
Real-time fraud detection, often identifying DOM manipulation attempts before traditional fraud systems trigger alerts
Analyst efficiency skyrocketed—what once took hours to review now takes seconds
Proactive fraud prevention—behavioral insights from Glassbox now feed into the retail bank’s fraud AI models for continuous improvement
“This has generated some of the best success stories we’ve heard in our department.”
Why It Matters
By partnering with Glassbox, the bank is transforming its fraud detection strategy. Analysts can now identify and stop sophisticated, client-side fraud quickly and efficiently.
With session replay capturing 100% of user actions, fraud teams can analyze activity both in real time and retroactively, enabling a behavior-first approach that scales with risk.
