From Automation to System Design: Why Incremental Robots Are Not Enough
For more than two decades, automated i.v. compounding has delivered measurable improvements in accuracy, repeatability, and partial risk reduction. Robots work. That is no longer in question.
What the literature now makes equally clear—often between the lines—is something more structural: most automated compounding solutions were never designed as complete systems. They are machines inserted into workflows that remain fundamentally manual, fragmented, and serial.
The result is progress without transformation.
Automation has focused on replicating tasks, not redesigning the process
Historically, i.v. compounding automation has followed a conservative path:
reproducing manual aseptic procedures inside an enclosed device.
Robotic arms replace hands.
Scales replace visual checks.
Software replaces paper documentation.
This approach was logical in the early stages. It allowed automation to be accepted without challenging existing clinical and regulatory paradigms.
But the evidence now shows its limits. By preserving the structure of manual workflows, automation inherited their constraints:
serial manipulation of materials,
heavy dependence on pre- and post-processing by humans,
bottlenecks tied to single actuators,
aseptic robustness that still relies on perfect procedure execution.
Across vendors and “generations,” systems differ in execution—but not in architecture.
The same bottlenecks, again and again
When studies report performance drops with non-standardized preparations, or dramatic increases in operator time outside the robot, they are not describing implementation failures. They are exposing architectural ceilings.
Most current systems still rely on:
one or two anthropomorphic robotic arms,
serialized handling of all consumables,
limited internal buffering and storage,
manual finishing steps before delivery or administration.
No matter how refined, this model cannot scale in throughput, nor can it eliminate human exposure across the full lifecycle of the product.
The literature documents this consistently—even when conclusions remain cautiously positive.
Why adding features does not change the outcome
Over time, innovation has been incremental:
better vision systems,
tighter gravimetric tolerances,
improved software orchestration,
more sophisticated verification steps.
These improvements optimize within the same conceptual frame. They do not alter the fundamental dependency on serial motion, human intervention, and workflow fragmentation.
In other words, the industry has been optimizing machines—not systems.
A different premise: automation as an end-to-end system
A system-level approach starts from a different question:
What would i.v. compounding look like if it were designed from the ground up for automation—not adapted from manual practice?
This perspective leads to radically different priorities:
parallelism instead of serialization,
orchestration instead of task execution,
physical separation of humans from critical paths,
finished products that are truly ready for delivery and administration,
aseptic integrity enforced by design, not by procedure.
This is the gap the literature implicitly highlights—and the one legacy architectures cannot bridge.
Where Lyvra fits into this shift
Lyvra, developed by Tylent, was conceived from this system-level premise.
Rather than automating individual compounding steps, Lyvra approaches i.v. compounding as a continuous, end-to-end industrial process, spanning:
pre-processing,
compounding,
post-processing,
preparation for delivery and administration.
Key enabling principles—publicly disclosed and central to this approach—include:
magnetic levitation–based material handling, eliminating mechanical bottlenecks and enabling parallel movement,
driven orchestration of processes rather than task-by-task execution,
true parallel processing, decoupling throughput from single-arm limitations,
end-to-end automation, reducing or removing residual manual interfaces where risk migrates today.
The objective is not to make compounding faster in isolation, but to remove the structural causes of low throughput, residual exposure, and workflow fragility documented across the literature.
Learning from mature industrial automation
Other safety-critical industries reached a similar inflection point decades ago. Semiconductor manufacturing, sterile fill–finish, and advanced logistics only scaled when processes were redesigned around automation—not merely automated in place.
The same lesson applies here:
automation delivers its full value only when the system is re-imagined, not when legacy workflows are preserved inside a machine.
The takeaway
Automated i.v. compounding has proven its value.
What it has not yet fully achieved is systemic transformation.
As long as automation is constrained by serial architectures and manual boundaries, its benefits will remain bounded.
Lyvra represents a deliberate move beyond that model—toward i.v. compounding designed as an integrated, parallel, end-to-end system.
Not an incremental robot.
A different way of thinking about the problem itself.