How to create randomised audio subsystems in metasound?

Hi!

How should I build a metasound that randomises and plays one long wave sound (lets say it lasts 1 minute and is called sound A) and starts playing random short sounds from selection on top of that (lets call it sound array A) until that one minute has passed and then it randomises another long sound (lets say it randomises sound B) and starts playing random sounds on top that from its own selection (sound array B). So basically: How to create a metasound that plays one subsystem of sounds and then randomises another subsystem of sounds etc.?

I am trying to create a musical phraser that has these long sounds as a bottom voicings such as C major symphonic strings and shorter sounds would be short phrases/notes that matches in that C major.

Thank you if somebody could help me or link a tutorial.

Something like this?

And then input a Stop somewhere to kill the loop