Kardex commits to 99.8% uptime as vision moves beyond the bin

Kardex commits to a minimum 99.8% uptime target on its automated storage systems, and that number frames how the company deploys machine vision — not as a stand-alone feature, but as a component that must earn its place in the cycle time. Fred Fox Jr., president of Kardex Solutions Americas, laid out the position in an interview with Vision Systems Design, covering everything from projected picking interfaces to compressed-air sensor cleaning.

What does the picking interface actually do?

At an AutoStore carousel port, Kardex's intuitive picking assistant replaces the conventional monitor for basic tasks. An overhead projector and camera system present picking instructions directly at the workstation, while sensors track where the operator's hand moves. The operator interacts with projected controls instead of a touchscreen. "How can you increase quality? How can you increase speed, even reliability?" Fox said, framing the tool as a labor and usability answer for warehouse teams facing rising service expectations and shrinking training windows.

The system can also support barcode scanning and capture of serial identifiers such as an IMEI or a medical-device identifier, depending on business rules. Fox's caveat is a design rule in itself: if a vision system adds several seconds to every pick, a handheld scanner may be the more practical choice. Recognition performance alone does not justify a deployment; the evaluation has to include cycle-time impact and operator interaction.

How do cameras and sensors work together at the port?

Fox described a mix of 2D scanning, camera-based sensing and additional sensors, with 3D scanning under consideration. At a carousel port, sensors estimate bin fill by measuring the distance to the bin and to the product surface. Environmental conditions stay on the deployment checklist. Direct sun through a window at a particular time of day can affect performance, and Fox urged installers to ask: "Is this really a one-day thing…or is it actually showing and highlighting a larger issue?" Ambient light, dust, product presentation and workstation layout can all shift results even when the underlying equipment is identical — hence testing at the actual installation site.

Where does the intelligence sit?

The picking assistant integrates with Kardex's FulfillX warehouse execution system (WES). FulfillX manages inventory, orders and allocations; a local computer at the port runs machine-level logic, responding when a bin arrives, interpreting a scan or detecting an operator interaction. That split lets the vision-enabled device operate within the broader system and scale across multiple ports through configuration and site network integration. FulfillX installs on-premises today, with a cloud-based WES in pilot deployments. Kardex ships updates once or twice a year, with new functionality available under the licensing model.

Can vision-guided robotics handle unknown products?

AutoStore Carousel AI, developed with Berkshire Grey and AutoStore, captures an image of the presented bin and uses algorithms and machine learning to determine the grasp, rather than relying entirely on preloaded product dimensions. In a customer demonstration with aftermarket automotive parts, the system picked items never sent in advance for testing — including bags of washers and nuts and bolts — after the team created the SKUs, inventory and order in FulfillX. Each application still requires validation against its own product range, Fox noted.

What keeps a system at 99.8%?

Deployment is the start of a long-term operating relationship, not the end of a project. Kardex introduces service teams before go-live, runs regular customer-success reviews and schedules preventive maintenance. "Warehouses are dirty. They have cardboard. They have dust," Fox said. Two tools address this directly: a robotic vacuum cleans the AutoStore grid, and a sensor-cleaning station uses compressed air to keep robot sensors clear.

His operating philosophy compresses to a phrase: "Keep it simple, and I guess…another synonym there would be flexibility." He expects AI to advance through focused applications — comparing recorded product dimensions against observed ones to fix item-master data, for instance — rather than as a silver bullet. Vision, in his words, may be an "unsung hero": rarely the most visible part of the cell, but often essential to making it work.