Voice model that lets AI agents listen, speak, and act at once
The Hook
“Hi Quique, returning your call this morning. You said you wanted to move a delivery.”
Problem Agitation
“Yeah, to Friday. And I also wanted to make a return.”
Product Reveal
“Sure. Let me get into that. I'll move the delivery first.”
Feature Teaser
“While that's updating, what order did you want to return?”
Call to Action
“The cycling jersey I bought last week. It's a bit too tight on me.”
Synthesizable in Remotion · key components of the audio-intent-timeline-tracer template.
A thin vertical hairline traverses horizontal tracks mapped to visual audio bars.
Use useCurrentFrame and interpolate to translate the playhead X-position from 0 to containerWidth across the composition duration.
A high-contrast black rounded pill container slides or toggles focus between sequential workflow status labels.
Compute current active state based on frame thresholds; apply spring-driven layout transitions on badge position.
Purpose. Demonstrate instant context retrieval from previous customer touchpoint.
Execution. Agent speaks while Extracting memory block is active.
Purpose. Introduce complexity via multi-task customer instruction.
Execution. Customer waveform pulses as user requests both date change and return.
Purpose. Show AI executing background jobs while maintaining active conversation.
Execution. State badge shifts to workflow triggering then API parsing while dialogue continues uninterrupted.