Decisions, straight
from the waveform.

Tazo is an audio-native decision model. It hears the conversation and decides what happens next, no transcript in between.