Many companies that sell surgical carts don't manufacture them. Their intellectual property lives in the software, the procedure, or the disposables; the capital equipment simply has to work so the rest of the system can generate revenue. That creates a decision most OEMs only make a few times: who builds it, and when do you bring them in.
GMI recently sat down with four members of the team closest to that question, Chellie Steinmetz (Project Manager), Spencer Strand (NPI Engineering Manager), Kyle Doyle (VP of Business Development), and Hans Dittmar (VP of Marketing), to discuss what it actually takes to launch a complex medical or surgical cart.
[Watch the full GMI: Behind the Build panel discussion]
Design intent is usually settled long before manufacturing gets a look. By then, tolerance stack-ups, part strategies, and supplier decisions are expensive to unwind.
Spencer's position is direct: the earlier GMI's launch engineers see a preliminary BOM, drawings, models, or a prototype, the more they can contribute across electrical, mechanical, hardware, and safety elements. That work starts with a design review and a first pass on the full design package, and in most cases, followed by one or more prototyping phases rather than a run straight into pilot. The reason is practical: you want to exercise the supply chain and the manufacturing process before ramp-up, so the ramp-up itself stands the best possible chance to succeed smoothly - the first time.
Alpha and prototype units built to clear a design gate are not the same thing as a product that can be built consistently, month after month. Closing that gap is the work.
Technical deliverables are only part of a launch. Chellie Steinmetz described the NPI team's job as scoping complexity, identifying risk, setting checkpoints, and then protecting the elements outside the design that can quietly become gating items later.
A design plan that reads "Stage 3 V&V" (Verification and Validation) is not a single activity. It's dozens of activities with multiple interdependencies and keeping them aligned is its own discipline. Engineering's preferred path only works if the supply chain can support it on the same timeline, and if testing can still happen when it needs to. GMI's cross-functional structure exists so those tradeoffs get evaluated from several perspectives before a decision is made.
Every project starts by defining success, a launch date, a cost target, a first market. Then the team sorts requirements into firm and flexible categories. When something slips, that distinction is what lets the team adjust instead of restart.
Weekly calls keep it honest. GMI uses them to flag what it's hearing and define the downstream impact before a decision in one department creates a problem in another.
Chellie's observation on derailments is worth repeating: most bottlenecks trace back to information not moving cleanly between groups, and decisions getting made without everyone aware of the implications. GMI's launch engineers are expected to surface context and downstream impact rather than execute instructions in isolation. That's less about communication style and more about protecting the launch date.
Surgical carts aren't built in the tens of thousands. Programs can run as few as 50 and up to a few hundred units a year, on complex assemblies potentially carrying thousands of BOM line items.
GMI is structured specifically for that reality. Orders are effectively unitized, every component goes through initial inspection, and every manufactured unit goes through final inspection, and each is packed and signed off individually. Practically, that means a customer can order one, five, a hundred or more. Units will be staged into production as needed, without setup-and-teardown penalties or minimum-run pressure from a fixed process flow.
Function is the most obvious bar to clear if it is well defined. Appearance is one people frequently don't see coming.
These carts sit in operating rooms and hallways. Kyle Doyle calls them the customer's billboard in the hospital, which is why they get inspected down to the gap between molded plastic panels. As Chellie Steinmetz put it: "The finished cart is an extension of their brand in the marketplace."
In practice, customers often expect cosmetic consistency beyond the minimum acceptance criteria on the drawing. The color has to match the chips. The panel gaps have to be clean and even, not on average - on every unit.
Low volume also rules out conventional economics on large exterior parts. You cannot justify a quarter-million-dollar tool to build a handful of units a month. That forces creative, non-traditional processes, and it requires suppliers capable of high-quality, low-volume work at commercially viable prices. Finding them is where a dedicated sourcing member on the NPI team earns its keep, vetting suppliers and matching capability to project requirements.
Kyle Doyle's answer was simple: "The value comes from the people and the processes, not one or the other."
The processes came first. They make each launch repeatable, and they give new engineers a structure to grow into. The people are what the processes can't produce. This team has launched a lot of complex programs, and they carry the lessons from all of them into whatever lands on the desk next.
Chellie added the mechanism behind it. Because the NPI team is comprised of dedicated members and runs multiple programs, lessons learned on one project Monday morning can be applied to another the same day. A short feedback loop is why the process keeps improving.
The full conversation covers additional ground, global manufacturing strategy, supply chain transparency, DfM economics, and the specific challenges of low-volume cart production.