The call transcription feature does not turn the audio into text inside LiveAgent. As soon as a call recording is created, the audio is sent to the AI provider you configured in your account - an OpenAI provider - which transcribes it with the Whisper speech-to-text model and returns the text together with the detected language. In other words, an external service processes the audio of your calls.
What this means for consent
Enabling the feature is your confirmation that you have the caller consent this requires. By checking "Enable call transcription" in Configuration > Call > Settings, you confirm that recording and transcribing your calls complies with the laws applicable in your country and in your customers' countries.
How the consent is usually obtained
- The standard way is an IVR welcome message, played to the caller when the call is initiated, announcing that the call is being recorded and will be processed by a 3rd-party AI company.
- Because transcription always follows a recording, the announcement you already use for call recording is the right place to extend - callers need to be informed about the AI processing as well, not only about the recording itself.
- If you do not want the audio of certain calls to leave LiveAgent, do not enable call recording for that phone number or call button - with no recording there is nothing to send for transcription.
Once the transcript is stored, it is protected the same way as your other personal data: the text is encrypted at rest with your account's own encryption key, access to it follows access to its ticket, and deleting the recording or purging the ticket removes the transcript with it.
Related articles