To the best of my knowledge, the only consideration is whether there is an open stream from or to a device speaker or mic at the time of backgrounding; audio session is not a factor.
If so, would there be a way to save the speech to an audio file and link it to a local notification?
I can imagine that this _might_ be possible, but it’s probably going to be a complex implementation that isn’t going to make use of many of the conveniences of the original code. I can’t give pointers on how to do it, sorry.