The OP didn’t indicate whether the voice-over track is one continuous event or a series of discrete events, but I think it is relevant. Based on one review of Voice Assistant on their website, and the tutorial provided by VAAST, it seems to work like Excalibur’s and Vegasaur’s ducking (both of which I own and have tried for ducking). As far as I can tell, all these tools do ducking based on the presence of a series of discrete events in the voice-over track, and they expect gaps in the VO track where you want the music to go back to normal volume.
If the VO track consists of one continuous event, with the narrator pausing as needed, they don’t work like you would expect a ducking tool to work in a DAW. You have to split the VO track into discrete events containing the spoken parts, and discrete events containing the silent pauses, then delete the silent events to leave gaps. Or you could record the VO track by starting and stopping the recording as required. Either approach is somewhat tedious. (Although less so than creating a ducking volume envelope manually. To that extent, these are useful tools.)
Since all these tools have this limitation, I’m guessing that it must be because of a limitation in Vegas that doesn’t allow them to do the ducking based on volume changes within a VO event. If that is the case, I suspect that VST ducking plug-ins will not work in Vegas either.
Hopefully I am wrong about this and there really is a way to do auto- ducking with one continuous VO event. I just haven’t found a way. Comments?