Yandex B2B Tech Launches Brand Voice Lite: A Rapid Speech Synthesis Service for Corporate Communication

Yandex B2B Tech has announced the launch of Brand Voice Lite, the first service in Russia designed for the rapid synthesis of corporate voices across various customer communication channels, including voice bots, podcasts, and educational materials.

The service is powered by machine learning technology utilizing Yandex SpeechKit models. To create a unique digital voice, clients need to provide between 20 to 40 minutes of audio recordings from a speaker. An additional 20 minutes of recordings are required to integrate emotional elements, such as sadness, joy, or aggression.

From this data, a custom speech synthesis model is developed, with access granted exclusively to the client.

The entire process of producing a voice model takes seven days. The fine-tuning of the models is conducted using the client’s dataset in a secure data center, ensuring the privacy and security of the speaker’s personal information.

As per the service’s guidelines, client companies must obtain documented consent from the individual for the use of their voice in the synthesis.

*Interested in staying updated on significant AI news? Subscribe to our Telegram channel* BotHub AI News.

Source