Financial institutions are confronting a fraud problem that increasingly begins with seemingly legitimate activity. In its 2026 Risk Officer Report, Federal Reserve Financial Services said financial institutions were seeing growing fraud challenges across major payment channels, with impersonation, social engineering, and compromised credentials among the prominent threats.

The problem is changing how fraud prevention is approached. A password, one-time code, or other authentication check may establish that the correct credentials were presented, but it does not necessarily establish that the legitimate account holder remains in control throughout the session. This has increased interest in controls that can evaluate risk continuously rather than only at login.

One response to that problem has already attracted professional interest. In correspondence reviewed by this publication, PwC, one of the Big Four global professional-services networks, expressed interest in licensing or further developing a behavioral-biometrics fraud-prevention system created by accounting and finance professional Elisha Adeboye, identifying online banking, payment processing and customer authentication as possible applications. The system is designed to continue assessing a user even after access has already been granted.

Rather than relying only on what a user knows or possesses, the system examines how the person interacts with a device. Typing rhythm, key-press duration, cursor movement, scrolling, tapping, swipe pressure, and touchscreen gestures are used to establish a behavioral profile. When activity departs significantly from that profile, the system can require additional verification, restrict access, lock an account, or issue a security alert.

The distinction matters because many conventional controls concentrate on the point of entry. A fraudster using valid credentials may successfully pass an initial authentication check but behave differently from the legitimate account holder once inside. Continuous behavioral assessment creates another opportunity to identify that difference before suspicious activity becomes a completed financial loss.

Adeboye’s system also combines machine-learning anomaly detection with adaptive user profiles, allowing legitimate behavioral patterns to evolve. Its architecture is designed for use across a range of digital financial environments, where unusual behavior could trigger intervention before unauthorized activity progresses.

Behavioral biometrics itself is not new. Adeboye’s system brings behavioral profiling, continuous monitoring, anomaly detection, adaptive updating, and real-time intervention together within a single financial control framework. The objective is not to replace passwords or other authentication methods, but to add another layer capable of recognizing when apparently legitimate access begins to look abnormal.

His move into fraud prevention did not come out of nowhere. At Intuit, Adeboye worked on a cloud-based financial management platform for small businesses in the United States and Canada, with a focus on giving users a clearer view of their finances and identifying problems earlier.

Earlier coverage of Adeboye’s work also traces his experience through PwC and Cummins, where he encountered the same basic problem from a different angle: what happens when a financial issue is not identified soon enough. In audit and corporate reporting, the consequences can appear as misstated numbers or delayed corrections; in digital finance, they can appear as fraud. Seen together, the work points to a simple idea: the earlier a problem is detected, the easier it is to contain.

As financial institutions contend with compromised credentials, impersonation, and increasingly sophisticated digital fraud, the next generation of financial controls may depend not only on proving who entered an account, but on continuing to assess what happens after they get inside.