PRODUCTION SCENARIO
An electronics retailer wants a spoken shopping assistant that replies in natural speech, lets shoppers cut in mid-sentence, and softens its tone when a caller sounds frustrated.
Which capability should the team build on?
Answering here is anonymous. Nothing is saved unless you sign in.
Show answer and explanation
Answer: The Gemini Live API, which streams audio with barge-in and affective dialog
The Gemini Live API handles low-latency two-way voice over a continuous audio stream, and Google lists barge-in, so a shopper can interrupt at any time, and affective dialog, which adapts response style and tone to the caller's expression. Text-to-Speech only voices a finished answer, Gemini 3.5 Transcribe returns text and holds no conversation, and a scripted call flow follows a fixed path whatever the caller sounds like.