Hospital Robots and Elevators: Lessons From a 2026 Florida Deploy

What did HCA Florida JFK add in September 2026?

According to a September 30, 2026 announcement from Relay Robotics, HCA Florida JFK Hospital in Atlantis, Florida brought two autonomous delivery robots online, named Dash and Daisy. The release positions them as support for routine, time-sensitive trips such as lab specimens, pharmacy orders, telemetry gear, and medications in areas that pneumatic tubes do not reach.

The same announcement stresses integration work across elevators, authorized secure doorways, and life-safety systems, with no physical tracks and no major construction. That detail matters because many hospital leaders picture a robot that only rolls on one floor until they see what multi-floor delivery actually requires.

Relay also notes that hospitals run around the clock, which means deliveries compete with visitors, shift changes, and cleaning carts in the same corridors. A Florida Atlantic coast site like Atlantis sits inside a dense health system footprint, so the lesson travels: your robot program must be designed for live traffic, not for an empty mock ward.

Why do elevators dominate the conversation?

In an acute-care hospital, the hard part is rarely drawing a map on a single level. Items that matter clinically often move between floors, which means the robot must call elevators, ride with staff and visitors, and recover gracefully when a car is busy or a route is blocked.

Relay Robotics describes coordination with building and technology systems so robots can move independently between floors and pass through badge-controlled doors. That is the same class of work you would scope for any serious internal logistics program, whether the hardware brand is Relay or another platform.

If your facilities team has never documented which elevators allow carts, which doors must stay latched for infection control, and which corridors are off limits during shift change, those answers need to be on paper before you invite a robot vendor onsite.

Relay's CEO quote in the release calls out elevators, secure doors, fire systems, and hospital IT in the same breath. That is your integration punch list. Each item has an owner on campus, and robots fail quietly when those owners were not in the same planning meeting.

Which deliveries are realistic on day one?

The HCA Florida JFK materials highlight specimen runs, pharmacy drops, equipment moves, and tube-free medication or lab transport. Those jobs are repetitive, time sensitive, and easy to measure, which makes them safer pilots than open-ended errands.

Emergency departments often surface early because specimens and supplies cannot wait on a single courier. Relay's public FAQ notes that ED teams benefit when another resource can keep point-to-point trips moving while clinicians stay at the bedside.

Start with routes that have clear pickup and drop-off owners on each unit. When both ends know how to load, track, and receive a compartment, adoption sticks. When ownership is fuzzy, the robot becomes everyone's backup plan and nobody's priority.

Authorized users load secure, lockable compartments and select a destination, according to Relay's description of how the robots operate. That workflow sounds simple, yet it implies training for every unit that will send or receive work, plus a policy for controlled items that cannot ride in an unlocked cart.

How do robots navigate busy corridors?

Relay states that its robots use LiDAR, sensors, and computer vision to move through hospital environments. Navigation tech is only half the battle; the other half is social. Staff must know how to step aside, when to pause a trip, and how to report a near miss without disabling the fleet.

Because the HCA Florida JFK deployment is infrastructure light, there are no floor tracks to guide wheels. That freedom helps older buildings, but it also means every obstacle is handled in software. Narrow entries, temporary construction walls, and double-wide beds in hallways should be logged during mapping so they do not become daily exceptions.

Think of navigation as a living document. When your facilities team moves a waiting area or closes a connector wing, the robot map needs the same change ticket nurses already use for patient flow.

What timeline should planners expect?

Relay Robotics states that a typical hospital deployment, including integration hardware, floor mapping, and staff training, takes about 8 to 12 weeks, with onsite installation commonly lasting 1 to 2 weeks. Treat that as a planning band, not a promise your campus will hit on week eight.

Mapping during live operations is slower than mapping an empty warehouse. You will stage work around patient traffic, fire drills, and infection-control rules. Clinical education belongs in the same calendar as technical cutover so nurses are not surprised the first time a robot arrives at a unit door.

Service Robot Co. often phases hospital programs the same way: prove one or two routes, document escalations, then widen the map once elevators, doors, and help-desk workflows behave predictably.

Relay also reports more than 1.5 million deliveries completed worldwide across its installed base, a scale marker that matters when you ask how mature hospital delivery robotics has become. Scale does not remove your obligation to pilot locally, but it shows the category is past novelty stage in many U.S. hospitals.

How should reliability be judged?

Relay's release cites more than 99 percent delivery reliability for its robots and a 99.8 percent delivery success rate across more than 30 major U.S. hospitals, with some sites exceeding 1,000 deliveries per month. Those figures come from the vendor's public materials as of September 2026, so your team should still define what success means locally.

Track failed trips, door timeouts, elevator waits longer than your threshold, and how often staff override the robot with a manual run. A robot that is technically autonomous but routinely abandoned mid-route is not saving anyone time.

Pair reliability metrics with infection-control habits. The same announcement highlights contact-free operation and surfaces that are easy to sanitize, which is table stakes when compartments carry specimens or medications.

Build a simple dashboard for executives: completed trips, average wait at elevators, manual overrides, and time saved on the pilot route. Without those four numbers, reliability debates stay subjective.

Where does a neutral integrator fit?

Relay Robotics built Dash and Daisy for this hospital, yet most health systems will evaluate multiple robot lines over a decade. Elevator interfaces, nurse call workflows, and access-control policies outlive any single hardware purchase.

Service Robot Co. stays OEM-neutral: we compare robots against your floor plan and IT constraints, then finance, deploy, train, and service under RoboCare with a nationwide engineer network. We are not Relay's dealer, and we do not present any manufacturer announcement as our own deployment.

Use public case studies like HCA Florida JFK as a checklist. Ask your shortlist how they handle secure doors, how they test with fire alarms, and who owns the runbook when an elevator bank goes out of service. The answers tell you whether you are buying robots or buying an integration program.

Financing and service should ride on the same contract as integration. Hospitals that buy hardware first and discover elevator work later often stall in capital committees. Bundling select, deploy, and RoboCare keeps the elevator conversation in phase one where it belongs.

What should leaders ask before signing?

Request a written integration scope that lists every elevator bank, door type, and life-safety touchpoint. If the vendor cannot name them, they have not walked your campus.

Clarify cybersecurity review for robot management software, badge data, and audit logs. Hospital IT teams will ask, and it is cheaper to answer early.

Plan staffing impacts honestly. Relay emphasizes that robots supplement employees; your labor model should show which trips disappear from human couriers and which new tasks appear, such as loading compartments and clearing exceptions.

• Document tube-covered zones versus tube-free zones before route design.
• Run a pilot route with nursing, lab, and pharmacy sign-off on both ends.
• Define who receives alerts when a robot stalls at a secure door.
• Schedule refresher training after the first month, when workarounds creep in.