The bank is testing a new technology to combat card fraud. It consists in detecting atypical customer interactions with a smartphone. If suspicious activity is detected, the transaction can be blocked in the mobile app. Each smartphone user has his own typical habits, and the neural network will analyze them. In particular, these are the characteristics of tactile interaction with the screen, the angle at which the user holds the smartphone, etc. By analyzing these parameters, the neural network can determine whether the owner is holding the phone or not. If the bank detects that the smartphone is not used by the client, its activity in the mobile bank will be put under special control and in case of an attempt to transfer funds to uncharacteristic details the operation will be blocked.