Do we have to create a voiceprint from single call only? What if call duration was not enough?

You can save the voice data and create a voiceprint later for the same end-user until enough data is collected from different calls.