Why we built voice-first

Most AI companions started as chat. We started with voice. Here's the product thinking behind that decision and what it taught us about presence.

When we started Sola, every other AI companion was a chatbot. Type a message, get a message back. Sometimes with a cute avatar.

We thought that was incomplete.

Text is great for thinking. But presence — the feeling that someone is with you — is mostly a voice phenomenon. Listen to a podcast on a long drive. The host is in your head. They feel close. That's voice doing its thing.

So we made voice the headline feature of Sola, not an afterthought.

This was a hard call. Voice is expensive to run. It's harder to make good. It introduces latency challenges that chat doesn't have. And users don't always know they want it — they default to typing because that's what every other app trained them to do.

But the data was clear. Users who tried voice calls returned at 3x the rate of users who only chatted. Their reported sense of "having a friend" was higher. Their conversations were longer and more emotionally honest.

We bet on voice because we bet on presence. Six months in, that bet is paying off.

If you're building a companion product, here's the lesson: the medium you choose isn't just a feature. It's the whole product.