Walk into a modern distribution centre and you’ll notice the floor is filling up with autonomous mobile robots — AMRs threading quietly between racking systems, picking stations humming away. What you might not notice straightaway is the person next to them wearing a pair of smart glasses, following an AR overlay that tells them exactly where to walk, what to pick, and where to put it. No clipboard, no handheld scanner, no paper list. That’s AR picking, and it’s one of the more genuinely useful things XR has done for industry.
Warehouse operations have always been a grind of repetitive, precision-dependent work. Pick the wrong item, mis-scan a barcode, route a parcel to the wrong loading bay — and the downstream consequences multiply fast. Returns, re-ships, unhappy customers, fines from retail clients with strict SLA requirements. Any technology that reduces that error rate without slowing workers down is worth serious attention.
Here’s what’s actually being deployed, and what’s still finding its feet.
AR-guided picking: the case that’s proven
The best-established use case is vision picking — using AR glasses to display pick instructions in the worker’s field of view. Instead of reading a handheld terminal or a printed list, the worker sees an overlay showing the next pick location, the quantity needed, and confirmation prompts. In some implementations, the glasses use a camera feed to recognise shelving labels and guide the worker turn-by-turn, similar to navigation.
The results from trials and live deployments are consistently positive on error rate and throughput. DHL Supply Chain, one of the earlier large-scale adopters, reported error reductions of around 40% in trials compared with paper-based picking. That’s not an outlier — similar figures come up repeatedly in independent assessments, because the fundamental reason for picking errors (working from a screen or paper you have to reference separately while handling items) is genuinely removed.
Training time is another benefit that often surprises logistics managers. A new starter in a large warehouse typically needs days or weeks to learn the layout and processes well enough to hit productive picking rates. With AR guidance, that ramp-up time collapses. The glasses tell you where to go and what to do. You don’t need to memorise the warehouse — you just need to learn to work with the interface.
The hardware reality
For a long time, the hardware held this back. Early smart glasses were heavy, the battery life was inadequate for a full shift, and the optics weren’t good enough in high-ambient-light environments. That’s changed substantially.
Devices like the RealWear Navigator 520 (a rugged Android-based headset designed specifically for industrial use) and the Google Glass Enterprise Edition 2 (now largely discontinued but still deployed in some facilities) established that wearable AR could survive a warehouse environment. More recently, Honeywell and Zebra Technologies have integrated AR guidance into their enterprise device portfolios, meaning that distribution centres already running those ecosystems can adopt AR workflows without switching vendors wholesale.
Battery life is still a practical constraint — most devices need a swap or charge at a break point during a long shift — but it’s manageable.
The bigger constraint, to be honest, is software integration. Getting an AR picking system to talk reliably to your warehouse management system (WMS) is where most of the project complexity lives. If you’re on SAP Extended Warehouse Management or Manhattan Associates WMS, there are well-documented integration paths. On a bespoke legacy system, expect more effort.
Human-robot collaboration: the growing use case
The more interesting territory opening up in 2026 is XR as the interface layer between human workers and warehouse automation. As AMRs and goods-to-person systems become more common, human workers’ roles shift from doing the physical movement to handling exceptions — the items an AMR can’t pick, the unusual package dimensions, the customer queries that need a human decision.
XR overlays are increasingly being designed to support this collaboration. A worker might see, in their field of view, which robot is assigned to which task, what’s queued up at their workstation, and alerts when an exception needs handling. The goal is to keep the human informed without pulling their attention to a separate screen. It’s a practical problem that AR is, arguably, genuinely well-suited to solve.
Research out of Fraunhofer and several Scandinavian logistics operations suggests that workers in human-robot collaborative environments perform better and report lower stress when given XR dashboards compared with screen-based equivalents. Whether that generalises is still being tested, but the direction of travel is clear.
What to think about before adopting
If you’re evaluating AR for a warehouse operation, a few things are worth getting straight before you sign anything.
Start with a specific, measurable problem rather than a general aspiration to “use AR.” Picking error rates, training time, compliance checking — these are quantifiable. “Making the warehouse smarter” isn’t.
Run a proper pilot on a contained pick zone before you commit to floor-wide deployment. The integration work and the change management (workers who don’t want to wear a device on their head all shift, supervisors who don’t trust the system yet) are much easier to sort out at small scale.
And check what your WMS vendor actually supports. Not everything in the brochure is production-ready for your specific platform version.
Done right, though, this is one area of XR that genuinely earns its place in a business case — not because it’s impressive technology, but because it solves a real operational problem that paper and screens haven’t quite managed to crack.