Client Snapshot
U.S. Class II/III device manufacturer shipping high-mix sterile trays for cath lab and OR settings.
Challenge
Returns and field complaints were traced to missing/incorrect components inside sealed kits—an issue traditional weight checks and spot audits missed. The client needed in-pack verification without breaking sterility or slowing shipments.
Mineral City AI Solution
Using radiographic images, Mineral City AI’s model verifies component presence, orientation, and count inside sealed trays—catheters, dilators, guidewires, anchors, and IFU placement—against the gold-standard tray map for each SKU and revision.
- In-pack validation before case sealing and at outbound QA.
- Revision-aware models that track BOM and layout changes.
- Operator UI with pass/fail plus visual overlays that show exactly what triggered a flag.
- Traceability: Images and decisions stored per serial/UDI.
Auditability: Unit-level decisioning with image evidence for QA/QC and investigations.
Implementation
Began with three highest-volume trays; expanded to long-tail SKUs using fast model cloning and controlled reference captures. Integrated with WMS for unit-level trace.
Outcomes
- Material drop in packing escapes and returns tied to incomplete kits.
- Less rework—flags move upstream where fixes are fastest and cheapest.
- Surgical readiness improved—fewer day-of-procedure surprises in cath lab/OR.
- Proof for regulators & buyers—image-based evidence strengthens QA posture.
Why it matters (for investors & partners)
- Defensible differentiation for OEMs and sterile packers.
- Software-led margin expansion—quality gains without new hardware capex.
- Partner pathways with sterilization providers and kit assemblers.