SecureAuth's Behavioral Biometrics is a new form of Multi-Factor Authentication that analyzes a user's keystroke and mouse movement behavior during login / page interaction to create a profile that is then compared to subsequent login attempts. Once a profile is trained, the user enters the required information into the fields, and SecureAuth IdP can determine whether this behavior is consistent with that of the mature profile.
To utilize Behavioral Biometrics, the Authentication API must be employed and the relevant endpoints called to by the application. The Authentication API enables customers to utilize SecureAuth IdP authentication and IdM Tools while maintaining the custom application's interface and preferred workflow.
Authentication API Behavioral Biometrics Workflow
The Accuracy Score is how close the current behavior resembles that of the trained profile, and the Confidence Score is how confident SecureAuth IdP is in the provided score. To increase accuracy and confidence, it is recommended to utilize numerous, static fields with six (6) or more characters per field. Having a profile that is rich in behavior produces the best results and raises the confidence level of the identity verification.
When a profile is trained, that behavior is what SecureAuth IdP expects; so if a field value changes (such as password, address, phone number, etc.), then reseting the profile is necessary to ensure that the correct values are being assessed. The PUT endpoint enables users to reset their entire, trained profile, or specific fields in their trained profile. Once a profile is reset, then the user undergoes the same training procedure as before until the new values are matured and the profile can once more be used in authentication.
Behavioral Biometrics Best Practices
To achieve the best Behavioral Biometrics results, SecureAuth recommends the following:
Use at least six (6) characters for text fields
After reseting a password, the profile (or profile field) should be reset to ensure accurate data collection
The more data that is captured in the application results in more accuracy and confidence in the scores
Profile training takes ten (10) sessions, so applications that are accessed more frequently yield faster training