Hospital emergency departments face a constant, high-stakes balancing act. Ambulances arrive in waves. Patients walk in. Vehicles circle for drop-off. Staff try to keep flow moving while patients wait — sometimes in the very spaces designed to help them. Much of this activity happens outside the doors, in ambulance bays, drop-off zones, and entrance areas that are difficult to monitor in real time.
Most hospitals manage these spaces with a combination of cameras, radio calls, and the experienced intuition of charge nurses and operations teams. That intuition is valuable, but it has limits. Staff often know when things feel "busier than usual," yet they lack precise, timestamped data on dwell times, queue formation, or emerging surges until the pressure is already visible inside.
At Iron Hornet, we've spent years building an AI vision platform that solves exactly these kinds of operational flow problems — originally in high-throughput truck yards and secure gate environments. We're now exploring how those same proven capabilities can be applied to Emergency Department exteriors.
A platform, not a pitch
This is not a new product we invented for hospitals. It is our existing, production-tested platform being adapted for a new environment. The core technology — scene intelligence, pattern-of-life learning, automated event detection, real-time alerting, and a conversational interface — already runs 24/7 in demanding logistics settings. What's new is the application.
We define zones (ambulance bay, ED entrance, drop-off area, parking approach) and let the system learn what "normal" looks like in each one, hour by hour and day by day. It automatically detects and timestamps meaningful events: an ambulance arriving and completing offload, a queue forming at triage, a vehicle lingering too long in the drop-off zone, or an unusual surge in walk-in arrivals. Alerts are contextual — what's normal during shift change may trigger attention at 3 a.m.
Perhaps most useful for busy frontline teams is the ability to simply ask questions in plain language:
The system answers with live, data-grounded responses drawn from the cameras and historical patterns — no dashboard hunting required.
Why we're approaching this carefully
We built this platform to solve real operational friction in complex outdoor environments. The underlying challenges — managing flow, measuring dwell time, detecting queues and surges — look remarkably similar to what we hear about at hospital ED entrances. But we're not assuming the fit is perfect.
This document and our current conversations are deliberately exploratory.
We want to understand where things actually break down outside ED doors, what information would be most valuable to charge nurses and operations leaders, and whether our approach to pattern-of-life intelligence and zone-based monitoring creates meaningful value in a healthcare setting. Your answers will shape what we build and how we configure it.
Privacy and practicality by design
Because this is healthcare, we designed the system with privacy as a foundational principle, not an add-on.
Operational activity, not individuals.
The system monitors vehicles, queues, flow, and zone utilization. No facial recognition, no capture of private vehicle license plates.
Minimal retention.
No storage of raw video beyond what's needed for event classification. Role-based access and a full audit trail on every query.
Practical to deploy.
Works with standard IP cameras (existing ones may be compatible), requires only edge compute hardware, and is up and running with minimal disruption after a short calibration period.
Healthcare-grade by design.
HIPAA-aware, audit-logged, and architected so the system has no path to identify any individual. We treat patient and staff privacy as a hard constraint, not a setting.
The bigger picture
What excites us about bringing this capability to hospitals is the same thing that drives our work in maritime and logistics: turning fragmented observation into coherent operational intelligence. When staff can see not just what's happening, but how it compares to normal, and can query that intelligence conversationally, they gain the ability to anticipate bottlenecks instead of constantly reacting to them.
We're not coming to Emergency Department leadership with a finished solution looking for a problem. We're bringing a set of battle-tested capabilities and a genuine question: which of these — if any — would meaningfully improve flow, reduce uncertainty, and support the people who keep EDs running?
The platform is ready when the problem is confirmed. What we want most right now is the conversation that tells us whether we're looking at the right problems at all.
We'd welcome the chance to learn from hospital operations and technology teams who are living these challenges every day.
We're here to help teams see it more clearly.
— The Iron Hornet Team