The call already happens.
The signal doesn't get read.

Your platform already runs a spoken check-in. Somewhere in that audio is fatigue, stress, hydration, and energy data your transcript throws away. Sona-2 reads the same recording a second time, for what the words alone don't carry, and flags a trend before the next visit, not after.