There’s also a goofy way that works for some use cases: you just set up another device near the echo that has the ability to play either a pre-recorded MP3 or has an intercom type feature and then you have that device play its voice file which Echo then responds to just like a person in the room saying the same thing.
Quite a few people do this with nestcam just because it has a good audio intercom (but not good ST integration). If you’re looking specifically for integration with smart things, you can do it with an android device and the smart app called LANnouncer.
So while it’s hokey, it works for a lot of situations.