Pulse
Notes from the field on clinical AI for the emergency chain: on-device intelligence, safer handovers, time returned to care, and what we learn making SONARA. One article a month.
-
Urgence médicale : l’IA doit d’abord faire ses preuves face au papier
Adoption is the wrong question. The right one is how to earn the trust paper already has, before offering what paper never can.
Read on Numerikare → -
Dans un métier où l’on prépare le pire, le papier reste d’une fiabilité redoutable
Why paper still wins on the worst days, and what any system replacing it has to prove first.
Read on La Libre → -
The EU AI Act, read from an ambulance
What the EU AI Act actually requires, on which dates, and where AI for emergency services lands in it. How SONARA was architected for the strictest reading, years before the deadlines.
Read → -
Why emergency AI must work offline
Cloud AI fails exactly where emergencies happen. Field notes on why SONARA runs entirely on-device, and what that changes for clinicians, patients and hospitals.
Read → -
The handover is where clinical data dies
The ambulance-to-ED handover is the most information-dense moment in emergency care, and the least recorded. How SONARA turns it into one structured, standards-based record.
Read → -
What we learned at HLTH Europe 2026
Four days at booth E130-3 in Amsterdam, dozens of conversations with hospital executives, EMS leaders and investors. Three priorities came up in almost every one.
Read → -
The zero-failure imperative
Emergency operations are asymmetric by nature. Connectivity is a luxury, not a strategic constant. Why real-time intelligence must reside with the operator, not the server farm.
Read →