What an AMR actually is — and what it isn’t
Autonomous mobile robots, or AMRs, are self-navigating units that move materials around a hospital without fixed tracks, magnetic strips, or pre-programmed paths. They use a combination of LIDAR, vision sensors, and onboard mapping software to navigate dynamically — meaning they can route around people, equipment, and obstacles, and they can be redirected to new tasks in real time.
AMRs are not AGVs. Automated guided vehicles follow fixed paths and require dedicated infrastructure. AMRs are not service robots — they are not the units rolling into patient rooms with telepresence screens or delivering trays directly to a bedside. AMRs are the workhorses of the back-of-house: they move bulk materials between fixed points on a campus, freeing staff for the work that requires a human.
Where AMRs work in hospitals
Five use cases account for the majority of AMR deployments in U.S. hospitals.
- Linen and laundry transport. Moving soiled linen from units to the central collection point and clean linen from receiving to floor closets. High-volume, predictable, and labor-intensive.
- Meal tray delivery and pickup. Moving food service carts between the kitchen and nursing units on a scheduled cadence. AMRs do not handle individual trays at the bedside; they move bulk carts.
- Pharmacy delivery. Moving medication carts, IV bags, and bulk supplies between central pharmacy and unit-based pharmacies or automated dispensing cabinets.
- Lab specimen transport. Moving specimens from collection points to the lab on time-sensitive runs. AMR climate control and security features matter here.
- Supply and waste runs. Replenishing PAR-level supplies on units and moving regulated medical waste to disposal points.
Where AMRs don’t work (yet)
Honest scoping prevents bad pilots. AMRs are not a fit for direct patient interaction, complex unstructured environments like construction or partial-renovation zones, or workflows that require frequent ad-hoc redirection by clinical staff without dispatch coordination. They also struggle in facilities with very tight corridors, heavy elevator dependencies that have not been retrofitted for robot access, or building infrastructure that cannot be modified to support charging stations and dispatch points.
The single most predictable pilot failure: scoping AMRs into a workflow that fundamentally requires human judgment — coordinating a complex discharge, escorting a confused patient, or making a clinical decision about which supply goes where. AMRs are a transport solution. They do their best work on the routes where the job is consistent, the endpoints are fixed, and the value is in volume.
How to scope an AMR pilot that succeeds
1. Pick a route, not a department
The strongest pilots are scoped around a specific transport route — say, the soiled linen route from the medical-surgical units to the central collection point — rather than a department. Routes have clear endpoints, measurable volumes, and discrete outcomes.
2. Establish the baseline before any robot moves
Measure current state for at least four weeks before the pilot starts. How many runs per day? How long does each take? What’s the labor cost per run? What’s the failure rate (delayed runs, missed pickups)? Without this data, the post-pilot ROI calculation is guesswork.
3. Map the building honestly
Elevators, automatic doors, fire doors, network coverage, charging station locations — these are where pilots stall. The pre-deployment site survey should produce a written inventory of every infrastructure dependency, with a remediation cost and timeline for each.
4. Pre-design the dispatch and exception logic
What happens when a robot encounters an unexpected obstacle? Who gets paged? How does a unit clerk request an ad-hoc run? Who reviews exception reports? The pilot succeeds or fails on these procedural details, not on the hardware.
5. Set the SLA before the launch
Uptime target. Mean-time-to-respond. Mean-time-to-restore. Run-completion-rate. Throughput per shift. Define what success looks like in writing, and define what failure triggers escalation.
The vendor decision
There are credible OEMs producing AMRs for hospital environments today, and the gap between top-tier and second-tier hardware has narrowed. The bigger decision is not which robot — it is who operates the fleet. Hospitals that procure AMRs as standalone equipment, on the assumption that biomedical engineering or supply chain will run them in-house, consistently underestimate the operating burden. Vendor-neutral operating partners can run any OEM the hospital chooses, which preserves OEM optionality while delivering the operating capability immediately.
What good results look like in a mature program
- AMR run completion rate: 97 percent or higher.
- Labor redeployment: 50–70 percent of the routes’ prior labor hours, with the displaced staff redeployed to higher-value patient-adjacent work.
- Cost per run: 30–50 percent below the prior labor-based cost, depending on baseline wages and run density.
- Throughput improvement: clean linen on units within 15 minutes of request; meal trays on time within a 5-minute window; lab specimens at the lab within posted turnaround targets.
Where to start
Start with the highest-volume, most predictable route in the hospital. Run the baseline. Define the SLA. Scope the pilot for eight to twelve weeks. Measure the outcomes. Then scale across routes, across units, and eventually across sites — but only once the operating model behind the fleet is proven.
