Audio will already crossfade between scenes automatically. If an audio source is not present in the destination, it will be faded out. If a source is not present in the origin scene, it will be faded in.
This ONLY happens during the transition period though, there is no carryover 'tail' for audio after the new scene has finished being transitioned to. You can submit that as an idea on the ideas.obsproject.com suggestion-box site though, independent audio crossfade duration on scene transitions.