The integration tax: why pick speed drops 12-15% between AutoStore and VLM systems.
If you manage a facility with both goods-to-person and VLM stations, you already know the friction. Operator A finishes a tote on the AutoStore portal, walks ten meters to the VLM bay, logs into a different interface, restarts to a different pick logic. Every shift, dozens of times. The muscle memory never fully transfers.
Recent OEM moves point at a unified picking interface that puts AutoStore and VLM workflows behind a single screen. One login, one pick sequence, one error-correction loop. That is real progress. But the question is not whether a shared UI exists, it is whether the physical layout underneath supports it.
A unified interface on two systems installed eighteen months apart, by different crews, with different conveyor tie-ins and different WMS handshake protocols, does not magically become one workflow. The screen reflects the layer of stack the eye sees fast. The part that makes integration meaningful is the part the operator never sees: cable routing, sequence logic, exception handling between platforms.
Interface unification is an architectural decision, not a software upgrade. It starts with how the systems are physically positioned relative to each other, and the operator’s natural pick path. The next bottleneck in multi-system warehouses is not the software layer, it is the physical seam between two systems that were never installed as one. Field perspective: when integrators bring us in to recover a stalled multi-vendor programme, the recovery work is almost always at the physical seam, not at the screen layer.