Basic Identity Use Case
In this page we explain how you can integrate your existing application with the Loccus AI voice identity service, in order to securely authenticate users with their voice.
Architecture
For the identity use case, we have two different flows that need to be integrated, the Enrollment Process and the Authentication Process.
Enrollment Process
This is the first flow for users, during which their voiceprint is generated for authenticating them later.
During this process, we need to gather audio from the user, and verify it, in order to compute a secure voiceprint.
Sample Enrollment Diagram
Enrollment Integration Steps
In order to start enrolling the users, you need to perform this three steps:
Record the voice
You need to have a record of the audio that you want to use for creating the user voiceprint. They can be of different quality and formats, please check the media file requirements.
Upload the audios and verify them
You can send the audios for verification using our API. Later you can perform message verification and authenticity verifications against them. Please check the guide for performing an authenticity verification and guide for performing a message verification.
We will return the verification scores and subscores, which provide insights of whether the voice spoke the specified message, is synthetic, replayed, etc.
Compute the voiceprint
Once the audios are verified, you can compute the voiceprint. Please check the guide for computing a voiceprint.
Authentication Process
After the user voiceprint is created, later audios can be verified against it, in order to verify that it’s the same person speaking.
This capability allows for biometric MFA implementation in different scenarios.
Sample Authentication Diagram
Authentication Integration Steps
In order to use this functionality, you need to perform this three steps:
Record the voice
You need to have a record of the audio that you want to verify. They can be of different quality and formats, please check the media file requirements.
Upload the audio and verify them
You can send the audio for verification using our API. Later you can perform message verification and identity verifications against them. Please check the guide for performing an authenticity verification and the guide for performing an identity verification. We will return the verification scores and subscores, which provide insights of whether the voice matches against the identity, or synthetic, replayed, etc.
Use the insights
Once you have all the relevant data, you can decide which actions to take based on your business logic. Please check the verification thresholds page for reference and recommendations on interpreting the results. This are some common immediate actions to take:
- If the voice is not real, warn the customer of the suspected fraud and don’t authorize the operation.
- If the voice doesn’t match against the customer identity, warn the customer of the suspected fraud and don’t authorize the operation.
- If the voice is real and matches against the customer identity, authorize the operation.
Administration
The Loccus Console provides a graphical interface that allows your team to effectively manage the Loccus Platform Service. API endpoints are also available for administration and platform monitoring
Security Team
Mitigate AI fraud risk, and gather more Thread Intelligence based on evolving TTPs of top actors. Monitor the fraudulent calls, identify key security improvement points and raise awareness of new types of cyber attacks.
Managers
Monitor fraud KPIs and leverage voice insights for internal and external reporting.
System Administrators
Easily configure and manage the Loccus Platform usage, including IAM and retention policies.