System prompts are not technically using any TTS engine. You are hearing the audio recordings of the system prompts (which was usually recorded with the same voice as the Genesys TTS engine but not always). If you want a different voice you can override them with recordings of your own. Your other option is to not to use any of the runtime data options in a flow. Use ToAudioTTS instead of most of the other ToAudioX functions (ToAudioBlank and ToAudio are fine). If you're using the audio sequence builder when you add a data option to play a variable always choose Text to speech as the playback mode; if you don't see it as a mode then use ToString to convert your value to a string then go back to ToAudioTTS.
------------------------------
Melissa Bailey
Genesys - Employees
------------------------------