I would have expected Bailey’s app to say ‘question mark’ to indicate that a question was being asked. While kind of clunky, wouldn’t it be more useful than the potential ambiguity of no indication at all? Checking on MacOS just now, my last question has minor, but noticeable inflection to indicate it’s a question. And using my iPhone, the (female)Siri voice is slightly clearer (but I can’t figure out how to switch a male voice, preferably with a British accent).
I know that MacOS has long had text-to-speech (and limited speech-to-text) abilities as part of Accessibility (and the separate Dictation & Speech) Preferences. I’ve only used it rarely, to amuse myself. iOS has similar capabilities in Settings>Accessibility>Spoken Content (not a surprise since they share core code). Apple’s technologies process things on the device itself rather than ‘phoning home’ the way Alexa, Google, Samsung, and other voice activated systems do to interpret what you’re saying.