Today, MetaVoice becomes Familiar.
Voice AI is growing quickly, but most interactions still feel like using a walkie-talkie.
Current voice agents work best when conversations are turn-based and predictable. They struggle with interruptions, overlapping speech, corrections, pauses, and background voices. They also reduce speech to a transcript, losing how something was said.
These are not edge cases. They are part of normal human conversation.
The root cause is architectural. Systems designed for sequential text exchanges have been retrofitted for real-time spoken conversation.
As a result, most voice AI adoption today is concentrated in narrow, scripted calls where the conversation can be kept predictable.
Why Familiar
For voice AI to move beyond those scripted interactions, the AI has to adapt to the way people already speak. Not the other way around.
Talking to AI should feel familiar, like talking to another person. That is the standard we named the company after.
At the core of Familiar is a full-duplex speech model built for conversation itself. It listens while it speaks, understands what was said and how it was said, and keeps up as the conversation changes.
Where we’re starting
We’re starting with inbound and outbound revenue calls. Teams building voice agents with today’s systems told us that more than 40% of people hang up in the first 30 seconds. Every one of those hang-ups is a lost booking, an unpaid balance, or a lead that calls a competitor instead.
Familiar is already live and taking real calls across outbound lead qualification, appointment booking and debt collection. The model responds in about 350ms and runs inside your own VPC.
Talk to Mia and Leo, our AI personalities for revenue calls. Or start a 30-day pilot.
Welcome to Familiar.