Speech recognition

Make identity checks more pleasant for your clients, handle more calls, and lower the risk of insurance fraud.

Make things more comfortable for your clients when you verify their identity over the phone, and save them from having to remember passwords and access codes. Doing so just requires their voice, which is as unique as a fingerprint. The voice-biometrics algorithm listens and compares more than a thousand distinct speech characteristics, making the entire verification process significantly more reliable, faster, and more pleasant.

Recognizing the client and insurance fraud

Our automatic voice-recognition system uses deep-learning neural networks (DNN), which can function in one of two ways:

  • With the technologically simpler analysis of so-called vocal passwords, you’ll be able to compare predefined sentences with a stored voice sample from the client. This doesn’t depend on a particular form of password, which can be the same for everyone, and the system can even remind the client of a forgotten password. The discussion with the operator begins only after the verification process, when the client has been clearly identified.
  • In continuous voice analysis (free speech), the client is immediately connected to the operator, and identity verification is carried out in real time during the conversation. The operator is informed of the result during the call. This makes everything easier and more natural for the client.

The system can’t be deceived

Voice biometrics recognizes even the slightest differences, so you can still verify your client’s identity in the event of a change in his or her speech due to, among other things, illness or a deliberate attempt to alter the sound of their voice. This makes it easy to prevent the misuse of voice recordings.

The high reliability of this form of verification can thus be used, for example, to speed up and facilitate updates and changes to insurance policies without requiring your clients to come to a branch. At the same time, it’s a suitable solution for recognizing problematic clients and detecting attempts at fraud, such as when a claim is reported. In addition to identifying the customer, the system also perceives nervousness as well as other slight speech anomalies.

97 %

The average success rate of voice biometrics systems is 97%.

Get more info today, start implementing tomorrow.

Thank you

We will contact you as soon as possible.

Oto Gücklhorn

Account Manager