Orchestration and Control Are Different Problems

By: Welsh Harris, Co-Founder & CTO at Ascendco Health

One of the most important architectural decisions we’ve made at Ascendco is one most people wouldn’t think about.

Our system decides what should happen next. It does not control the equipment that does it.

The instinct in healthcare automation is to build monolithic systems — one platform that schedules, prioritizes, tracks, and controls the physical equipment. Washers, sterilizers, automated storage, robotics, all under one roof. We went the other direction.

Ascendco maintains real-time operational truth: where every tray is, what every case needs, what’s at risk, what should move next. We recalculate priorities every 15 minutes against live surgical demand. We enforce safety guardrails at every transition point.

But we don’t move the tray. We don’t fire the sterilizer. We don’t route the robot.

Orchestration requires deep clinical context, schedule awareness, and cross-system visibility. Physical control requires millisecond-level safety, hardware-specific protocols, and mechanical reliability. Mixing them makes both worse.

The systems that will scale in healthcare automation are the ones that separate what should happen from how it physically happens. One layer reasons about the operation. Another layer moves the equipment. This is the architecture that lets hospitals introduce automation incrementally, without redesigning their entire operation.

Which is why the robotics question is usually asked backwards.

The question we keep hearing is “when will robots work in sterile processing?”

I think the right question is: when will the orchestration layer be ready to tell robots what to do?

A robot in a sterile processing department is useless without operational context. It needs to know what tray to move next, how urgent it is, where it goes, what the correct processing path is, what safety constraints apply. That’s not a robotics problem. That’s an orchestration problem.

We’ve spent years building exactly this layer. Not for the robots, but because humans need the same information. The system we run in production today already:

  • Reconciles live surgical demand with real-time inventory state every 15 minutes
  • Produces prioritized work tasks with routing intent, urgency, and handling requirements
  • Enforces IFU-driven safety guardrails at every workflow transition
  • Publishes operational truth that both humans and connected systems can act on

When a hospital is ready to introduce automation — automated storage, robotic transport, or something that doesn’t exist yet — the orchestration layer is already there.

You don’t wait for robots and then figure out orchestration. You build the orchestration layer now, run it with humans, and let automation plug in when it’s ready.

The future of healthcare automation isn’t a single system that does everything. It’s an ecosystem where orchestration intelligence and physical execution are cleanly separated, and can evolve independently.