Maybe fire the text-to-speech trigger before publishing the messages?
Maybe fire the text-to-speech trigger before publishing the messages?