Text to speech - How to fix the first word being cut off

This is a pretty old thread now, but I'm experiencing it in fall 2023.  I believe the issue to be this:

Many bluetooth headsets/speakers will go into a type of power saving mode when there is no sound being transmitted, such as when a Storyline slide ends.  Sometimes this happens as quickly as 1/2 a second of silence, depending on the device.

When the next slide begins, the headset won't be triggered to "on" until it senses sound.  With a human narrated audio file there is likely to be a tiny bit of "room tone" or other sound that will precede the voice kicking in, even by a fraction of a second.  Most editors will always have a slight bit of dead air at the start of a narration file. This will trigger the headset to turn on and by the time the voice starts the first word, we'll hear it. But, with a generated voice the audio is perfectly clean, and thus the voice is the first sound the headset picks up.  This leads to the first word being frequently clipped.

I'm testing this out at my company now.  Most users are listening to e-trainings on Poly headsets and this problem is happening a lot.  I'm experimenting with background sounds, like music, to keep the headset on.  I don't really want music playing on everything so the next step is to experiment with some kind of sound that will trigger the headset but humans won't actually hear.  Maybe that tone that only teenagers can hear or something?