/voice
How your product sounds.
Voice is not an output format. It's a design material — with texture, tempo, and personality. Used well, it creates intimacy no screen can match.
As AI makes voice interfaces genuinely viable, the question is no longer whether to add voice — it's what kind of voice you want. The wrong answer sounds like every other AI product. The right answer sounds like you.
We design conversation — the structure of how an AI speaks, the choices it makes under pressure, the pauses, the tone when it doesn't know. These are creative decisions that compound into character.
We work with tools like ElevenLabs to give AI voices texture and emotional range beyond the default. And we design the inputs too — voice interfaces where speaking feels faster and more natural than typing.
Audio output — whether that's a generated voice, a sonic brand, or ambient sound design — is part of the experience. We treat it as such.
What this covers:
- 01Conversation design—The structure and language of how an AI speaks. Flows, fallbacks, tone, persona. The script and the improvisation.
- 02Voice interfaces—Input design for spoken interaction. Where voice is faster, more natural, or more appropriate than a screen.
- 03AI voice production—Character, range, and texture for synthesised voices. Built with ElevenLabs and equivalent tools.
- 04Sonic identity—The audio layer of a brand. Sound design, earcons, ambient audio. What your product sounds like when it's working well.