Healthcare is ready. Robots are ready.
Humans are not.
Social acceptance is the bottleneck, not navigation.
Nurses lose hours to errands a robot could run.
They block halls and startle patients. They don’t belong yet.
One bad encounter and staff stop using it for good.
“The robot works, but people won’t use it. Trust is gone.”
Care OS
A behavioral intelligence layer for robots. Hardware-agnostic · Human-aware · Healthcare-ready.
Slows, yields and gives way like a considerate colleague.
Reads the room and adapts how it moves and signals.
On-device, respects clinical norms, integrates on your terms.
Nav2 gets the robot through the corridor.
Care OS gets it accepted.
How Care OS works
Raw sensors in, socially-aware motion out. In milliseconds.
Existing sensors, read as anonymous geometry.
A live read of people, space and intent.
Picks a behavior, with an explicit reason.
Caps speed, pauses, yields or re-routes.
Split-second decisions on the robot. No cloud.
No patient data leaves the building.
Plugs in only where you allow.
Evaluate on a real shift, in a day.
Reacts to what’s in front of it, one moment at a time. Safe, but stiff, and blind to what’s next.
Predicts how the scene will move and plans seconds ahead. Sharper with every mission on your floor.
Always shielded by the proven safety controller. It proposes; the rulebook keeps the brakes.
From reacting to what is, to anticipating what’s next.
Safety stays in charge. The learning never gets the brakes.
Where Care OS delivers value
Errands off the ward, so staff focus on patients.
Samples, supplies and meals, moved between departments.
Greets, guides, signals intent.
A light, friendly presence on the ward.
Designed with Swiss hospital staff, first validations begin 2026.
A 1-day on-site evaluation
30–60 min to bring a robot online on your floor.
2–4 h running real corridor scenarios with your team.
Staff-in-the-loop: watch how people actually react.
30-min review with KPIs, and optional follow-up.
Research × Innovation × Business Execution
