Sesame's voice AI sounds great. Will anyone keep using it?

Launched 28 May, 39 countries, four agents, built on Sesame's Conversational Speech Model — the one that genuinely sounds human and whose CSM-1B base it open-sourced. Credit: the naturalness is real, and letting conversation flow instead of stalling is a smart, hard-won detail.

Day-eight retention. Not downloads, not '39 countries', not how good it sounds in a launch video. Every voice assistant of the last decade nailed the demo and lost the habit. Talking to your phone is socially awkward in public, slower than typing for many tasks, and a behaviour people abandon fast. A small startup needs that habit to form on Apple's phone, against Apple's own Siri AI.

So we're not dunking on the tech — it's excellent. We're flagging that 'excellent voice' and 'product people use daily' have been different things for ten years straight. If Sesame breaks that, it's a real story. Check back on the retention curve, not the launch buzz.