Whether you say “OK Google”, “Hey Siri”, interact with Alexa, or an operator’s voice assistant, the “wake up word” already uses local AI, and customers now expect an instantaneous response or at least one within the Doherty Threshold of under 400 milliseconds. Such low latencies are already achieved today, with some processing happening closer to the Edge by servers installed by Google, Apple, or Amazon in operator networks. But a purely Cloud model can only let us go so far.
by Arnaud Bensaid, Chairman, prpl Foundation Marketing Committee VP Marketing, SoftAtHome