aftercare.

aftercare

Inside a hospital, patients are surrounded by nurses, physicians, and monitors watching every vital sign. The moment they're discharged, that support vanishes — replaced by a folder of instructions and the hope that someone notices if something goes wrong.

We wanted to close that gap without pretending AI should replace a clinician.

That's Aftercare: a post-discharge follow-up system that checks in with patients by voice or text, in English or Spanish, flags concerning responses, and escalates them to a physician :- securely, and only when a human needs to step in.

The hard part wasn't making an AI hold a conversation. It was making one that knows when not to be trusted. So we built in hard limits: deterministic rules decide clinical severity, not the model. Every line in a physician briefing has to trace back to something the patient actually said. And every escalation passes through a cryptographic check before it ever reaches a doctor.

AI where it helps. Hard limits where it's risky. A real clinician always one step away.