If you have echo devices, it’s quite easy to set up announcements for them using echo routines (not ST routines). I have one friend who mostly uses Google home, but does have two echo dots used just for voice announcements.
We switched over to Google about a year ago, Amazon got ridiculous with its “by the way” stuff after every interaction… and then something on their end broke after working for eons it couldn’t complete simple commands like ‘turn off x light’ half the time.
We kept one Amazon device but I removed its access to ST… I did try using their voice for automations but I have to say Big Talker by far has the best , most pleasant voices to choose from. Echo is cold.
In Home Assistant if you use it with the Nabu Casa subscription, the Google text-to-voice languages are included and it can do that to any speaker what it supports. I know it works with Google Home devices and Sonos speakers as well.
I don’t use it too much, but I believe you can add other text-to-voice services as well.
All these services what SmartThings cuts will be a huge pile of money not spent. Plus the servers what they spare running the Groovy stuff…
You could also call the /services/tts endpoint which returns a URL that points to an MP3 file if your device commands supports passing an MP3 file directly to it. Or for that matter, you could generate a TTS or whatever sound file you wanted and store it somewhere then pass the stored file directly to your command.