Behavioral Biometrics for Fraud Manager

Platformized Behavioral Biometrics to Provide Defense in Depth

Strengthen your Fraud Prevention with a native Behavioral Biometrics solution that is built to provide signals for ‘All Cause’ Fraud Detection.

An integrated approach to Behavioral Biometrics

Fraudulent activities are increasingly sophisticated and difficult to detect using traditional fraud detection tools. The evolution of attacks for both Authorized and Unauthorized fraud involves social engineering of customers, where existing control layers based around authentication are less effective.

By understanding nefarious criminal behaviors, it is possible to provide a Behavioral Biometric capability that delivers risk signals for ‘All Cause’ fraud. Focusing on the right attributes gives tangible benefits no matter if the attack is an account takeover or a scam. Feeding these signals into a multi-signal risk engine using machine learning gives a new standard for unified protection.

Delivering Behavioral Biometric signals as a Platformized offering allows FIs to deploy them quicker and cheaper than integrating a separate point solution. This capability not only improves fraud detection accuracy but also provides Enhanced Authentication and regulatory compliance.

Understanding how Fraudsters operate drove our design

Outseer Behavioral Biometrics for Fraud Manager is a fraud risk signal that looks for nefarious signs within digital journeys. It analyzes user interactions to detect the anomalies in real time, being able to distinguish genuine users and genuine transactions from fraudulent.

With Outseer, an individual Behavioral Biometric profile is created. This profile is then used on future visits to analyze deviations from normal behavior. At the same time population level analysis identifies and correlates outliers with fraud typologies. New data points around bots/non-human, emulators, digital native traits and user focus are available to effectively look behind the device to better understand what is happening in the customer journey.

These behavioral risk signals are layered natively with traditional contextual fraud signals to deliver superior fraud detection while minimizing friction for legitimate users.

Better Overall ‘All Cause’ Fraud Detection

Identify suspicious behavior indicative of fraud in real time across a wider range of attack types.

Reduce False Positives by Correlating Your Signals

Improves your fraud detection false positives by giving additional insight into what is going on behind the device.

Enhanced Identification of Authorized fraud

Identify social engineering more effectively due to the changes in user behavior that occur when users are being guided.

Using the signals you need for the insights you want

Integration with Outseer Fraud Manager

The Behavioral Biometrics score is natively fed into the Outseer Risk Engine and used as a signal in the risk assessment of each transaction.

Behavioral Biometrics Data Collection

Collects signals from web and mobile browsers through a JavaScript that gathers keystrokes and mouse data from page interactions. For mobile devices, additional data can be collected from touch, orientation and movement through our Fraud Manager Mobile SDK.

Behavioral Profiling

Advanced machine learning algorithms detect deviations from established behavioral norms in real-time, flagging potentially fraudulent activities. The solution looks for nefarious activity at a population level and not just user profile level, allowing assessment on new users.

Population Level Analysis

By looking at how different features are distributed across the population it is possible to spot outliers which can be associated with fraudulent activity. This allows certain frauds to spotted even when there is no user profile yet.

Real-Time Anomaly Detection

Advanced machine learning algorithms detect deviations from established behavioral norms in real-time, detecting potentially fraudulent activities. The solution looks for nefarious activity at a population level and not just user profile level, allowing assessment on new users.

Behavioral Biometrics Score and Score Reasons

A unique score is produced for each activity or transaction from the data collected during the transaction. The score and indicators can be used in the Policy Manager and Case Manager applications.

Differentiating Fraud Types

By understanding what is happening on the device better clarity can be given in terms of distinguishing fraud types. Authorized and Unauthorized fraud can be classified to the distinct behaviors that happen in them.

How it works

Outseer Fraud Manager with Behavioral Biometrics

1) User Authenticates and Data is Collected

Behavioral Biometrics data is collected when the end user uses an application protected by Fraud Manager at logon and during interactions such as making a transaction. The behavior data is collected through a JavaScript component in the browser.

2) Risk Indicators Are Evaluated

Outseer Behavioral Biometrics monitors how the user enters their credentials and navigates through the session doing other interactions; it does not capture what is being typed, only how it is being done so data privacy is upheld.

A user behavioral profile is built with data including typing speed, mouse movements, device usage, and browsing habits. The profile is updated and refined on each user visit.

Advanced machine learning algorithms are used for real-time nefarious detection, to detect deviations from established behavioral norms and flag potentially fraudulent activities. This is done at both the profile and population level.

Outseer Fraud Manager also profiles the user's activity on over 100 additional risk indicators. These include signals such as device details, transactional patterns, signals from the Global Data Network and other custom risk factors.

3) Behavioral Biometrics Score Calculation

A unique Behavioral Biometrics score is produced for each user activity based on the data collected during the session and the comparison of anomalies to the user behavioral profile. This evaluates the level of threat posed from known nefarious behaviors and additionally indicates if the user was recognized as themself based on the comparison to past behaviors. This score is considered an additional signal that is considered in the overall fraud risk score, as it is complementary to the other existing signals.

4) Overall Risk Score Calculation

Additional risk signals are normalized and processed by the Outseer Risk Engine to compute an overall risk score for the user's activity. By integrating the Behavioral Biometrics score into the risk assessment, we enhance the accuracy and speed of detecting account takeover, scams and authorized push payment fraud.

5) Policy Evaluation

The total risk score is fed to the Outseer Policy Management application, which assesses if the activity violates any organization policies and thresholds, allowing for tailored decision-making and actions based on user behavior. There are two possible outcomes:

Proceed as normal: If the user activity doesn’t show anything suspicious or violate any policies or rules, the user is transparently authenticated and continues with their digital interaction without interruption.
Additional authentication needed: If the risk exceeds the threshold set in the policy application, the system can prompt additional assurance or step up authentication, mark the transaction for review later in the case management application, and/or block the activity.

6) Case Management and Analysis

The Behavioral Biometrics score and reasoning are accessible, providing additional context for investigations.