Human-supervised AI brings proven dementia care to scale
Key points
Key takeaways from EU Startups interview with NewDays CEO Babak Parviz (October ️2026(:
Origin story. Both founders share family histories — their fathers suffered from Alzheimer's or dementia;; Parviz previously helped build Google Glass in Google X,, spent several years as VP at Amazon leading the company into healthcare, and was a venture partner at Madrona investing in AI, healthcare and deep tech.
Scale of the crisis. Among people over ️65, ️11% have dementia and a further ️22% have mild cognitive impairment — about one in three needs cognitive treatment,, and no pharmaceutical intervention exists yet that cures or stops these diseases.
The hidden cost. Compared with medically similar peers, MCI adds about $18,000 per person per year to the US healthcare system,, and dementia up to $30,000 — costs that, with the world ageing through ️2050, could bankrupt healthcare systems if unaddressed.
Proven, but never scalable. NewDays built on randomised controlled trials: effective interventions existed but required years of one-on-one, highly personalised work with trained clinicians, so none of them ever scaled.
Human-supervised AI, at last. Generative AI offloads the repetitive work while an expert human clinician stays in charge and accountable — "the combination of the two is what allows us to scale," said Parviz;; this was not feasible even 3 years ago,, and the company has no plan to ever cycle out the human.
Rolloutand regulatory friction. NewDays operates state-by-state in the United States — six states covering just over a third of the population — with partnerships spanning neurologists, caregiving organizations and senior living companies;, demand already reaches from Latin America, Europe and Asia. He argues state-by-state medical licensing adds burden with zero patient benefit and should be removed.; some regulation should stay or even get stronger.
The endgame. No cure for Alzheimer's exists yet,, so biological research must continue — but millions cannot wait for a hypothetical cure years or decades away; they need help today. Parviz expects all medicine to eventually be practised this way: AI assisting — never replacing — an accountable human clinician.
Read more: EU Startups
