Physical AI / Operational Preparedness

When AI moves from recommending to acting.

Smiths Hospital PAI focuses on the operational conditions hospitals need to consider as AI-enabled systems begin to interact with the physical hospital environment.

Operational Limits

Where and under what conditions can the technology safely operate?

Dependency Risk

What happens when hospital operations become dependent upon the technology?

Critical Safety Gates

Which preparedness deficiencies should prevent progression despite a strong aggregate score?

Adoption Envelope

Under what conditions has the organization demonstrated sufficient preparedness for deployment?

Human Override

Can trained people safely intervene, escalate, and continue operations when technology fails?

Monitoring & Learning

How will performance, near misses, failures, and operational drift be identified and acted upon?

Emergency Department Physical AI Stress Test

Don't test preparedness only when everything is working.

Normal Operations

Routine demand and expected system conditions.

Crowding & Boarding

High census, constrained space, and prolonged throughput pressure.

High-Acuity / Resuscitation

Time-critical care, competing priorities, and rapid coordination.

Behavioral / Security Events

Unpredictable movement, access restrictions, and safety concerns.

Technology / Infrastructure Failure

Connectivity loss, downtime, degraded capability, and safe fallback.

Surge / MCI

Simultaneous demand, scarce resources, and emergency command conditions.

Public vs. proprietary.
Public materials explain the preparedness philosophy, domains, and broad stress scenarios. Detailed assessment questions, weighting, thresholds, scoring algorithms, dependency calculations, and stress-test protocols are not published.