| CPC H04L 63/0861 (2013.01) [G10L 17/04 (2013.01); G10L 17/06 (2013.01); G10L 17/22 (2013.01)] | 20 Claims |
|
1. A method comprising:
at a computing platform comprising at least one processor, a communication interface, and memory:
monitoring remote communications between a first user device and an enterprise user device, wherein monitoring the remote communications comprises gathering voice information corresponding to a first user;
identifying, based on the gathered voice information corresponding to the first user, a first voice sample;
configuring a user profile corresponding to the first user, wherein configuring the user profile comprises linking the first voice sample to the user profile;
training, based on the first voice sample, a voice authentication model, wherein training the voice authentication model configures the voice authentication model to output similarity scores indicating a degree of similarity between voice samples and the user profile;
monitoring a second user device, wherein monitoring the second user device comprises continuously gathering voice information corresponding to a second user;
identifying, based on the continuously gathered voice information corresponding to the second user, a second voice sample;
executing the voice authentication model using the second voice sample as input;
generating, based on executing the voice authentication model, a first similarity score corresponding to the second voice sample, wherein generating the first similarity score comprises comparing the first voice sample to the second voice sample to identify a similarity between the first voice sample and the second voice sample;
comparing the first similarity score to a threshold value;
identifying, based on the comparing, whether or not the first similarity score meets or exceeds the threshold value;
in response to identifying that the first similarity score meets or exceeds the threshold value, updating, based on the second voice sample, the user profile;
in response to identifying that the first similarity score does not meet or exceed the threshold value, initiating a security action, wherein initiating the security action comprises:
suspending access, for the second user device, to a network;
sending one or more commands to display a user interface to the enterprise user device; and
receiving, from the enterprise user device, user input, wherein the user input directs the computing platform to:
override the security action by reinstating access, for the second user device, to the network; and
initiate one or more additional security actions; and
refining the voice authentication model based on the second voice sample and the first similarity score.
|