Shipping AI to a Million Patients Without an A/B Test — Jared Joselowitz, Ufonia
Ufonia's Dora AI has completed 200,000 clinical calls and is contracted to reach 1 million patients.
“the model card won't save you. Um you can't claim like some model vendors said that they have 92% on some benchmark. Um it's not a defense at a post incident review.”
Ufonia's voice AI agent Dora has conducted 200,000 real clinical calls across 20 UK hospitals and is contracted to scale to 1 million patients within 2 years, now expanding to the US. The talk surfaces a critical deployment constraint unique to healthcare AI: no A/B testing (randomizing patients into worse variants is unethical and illegal), no rollback once a call is made, and benchmark scores offer zero legal or clinical defense post-incident. This frames a concrete alternative safety paradigm — rigorous pre-deployment evaluation — that is directly relevant to any team shipping AI in high-stakes domains.