AI in Warehousing: Where to start | Socius24

In April, MHI and Deloitte released their 2026 Annual Industry Report at Modex in Atlanta. It’s called Rewiring the Future: A Supply Chain Playbook for Innovation, and it contains one number to which all of us should be paying very close attention.

According to the report, 48% of supply chain leaders now consider AI’s disruptive impact on their operations to be significant or greater. A year ago, that number was 23%.

A quarter of respondents went even further, calling AI “transformational.” Robotics and automation came in second, with an increase of 16 points over 2025. And all of this underlines what many of us have known for a while, which is that the influence of AI in warehousing and logistics can no longer be considered theoretical.

What’s even more interesting is what the report says next. Leaders, it concludes, are “getting stuck on where to start and what it takes to scale.” They’ve read the case studies. They’ve sat through the conference keynotes. They’ve watched the vendor demos. They know AI matters to their operation. What they don’t know is where the heckity heck to start.

MHI’s CEO John Paxton summed it up rather succinctly. He warned operators about “running tomorrow’s operations on yesterday’s equipment and technology.” Which might be a rather uncomfortable read for most warehouses… mostly because he’s right. You can walk into plenty of UK distribution centres today and see exactly what he’s talking about. Paper pick lists, a WMS that was cutting edge in 2011, and a whole lot of workers holding things together using instinct, overtime and industrial amounts of rather nasty coffee.

Paxton’s suggestions to deal with all of this are where things start to feel like his report was written in a boardroom rather than a warehouse. He says supply chains “can no longer be optimized at the edges” and must be “rewired end-to-end.”

What a LOVELY idea.

FYI if this is the first time you’ve lived through this kind of significant upgrade, you’re probably going to need something along the lines of the following:

• A five-year capital programme with a steering committee
• A change management consultancy
• A PowerPoint deck that will most likely outlive the CFO who commissioned it
• Maybe, perhaps… some headache pills?

Most UK distribution operators just don’t have that kind of runway. They’ve got ‘this quarter’, a peak season that’s on its way, come hell or high water, and they’ve got a labour bill that’s gone up every year since 2020. Oh, and they’ve also got customers who want their orders to arrive faster than they did last time.

So back to the big question. Where do you start?

Our recommendation? Start with your Picking.

Picking sits upstream of almost everything else that happens in your warehouse. If it’s not optimal, the cleanest packing lane in the world won’t be able to rescue your throughput numbers. The best inventory accuracy programme in the sector isn’t going to save you, if your pickers are walking ten miles a shift to find the stock that you’re trying to ship. If you can fix the picking, the rest of your operation will get easier by default.

But don’t just listen to me… here are three real-life reasons why picking is your best first move.

The first of those is labour. UK warehouses lost roughly 125,000 EU workers post-Brexit and the worker gap hasn’t got any smaller since then. As well, wages have climbed. Agency rates during peak are just… punishing. Staff churn in some facilities runs between 30 and 60 percent a year, and the process that contributes to most of that cost is picking. It’s where your people spend their time, and it’s where every wasted step shows up on your payroll. If you can cut travel time then you’ll reduce your labour cost, directly and immediately. Simply put, there’s no easier way to reduce costs.

The second thing you need to look at is scope. Pick path optimisation doesn’t ask you to replace your WMS. It doesn’t demand a reskilling programme or a robotics business case. It runs on top of what you’ve already got. And that matters far more than most vendors are comfortable admitting. Every other AI conversation the MHI report focuses on talks about transformation. Transformation is expensive and it is slow. And frankly, it can be career-threatening if it goes wrong. What most operators can actually approve this year is a targeted upgrade to an existing process that’s going to offer a predictable payback window. Optimising your pick path fits into that box. End-to-end supply chain rewiring does not.

The third reason to start with picking is measurement. You’ll know within a fortnight if your pick path intervention is working or not:

• Distance walked per order
• Picks per hour
• Order accuracy
• Lines per labour hour

These are all numbers that your floor managers are already tracking on a whiteboard somewhere. When it comes to picking, the feedback loop is short enough to be both accurate and useful. Contrast that to most AI projects, which typically have benefits that are only going to be realised eighteen months from now. But to be fair, that lag doesn’t really matter, because those benefits are stored on a Gantt chart that nobody actually opens.

For those of you who are already running Blue Yonder Dispatcher WMS, this is where Optioryx Pulse fits in. Pulse connects to the move tasks that your Dispatcher WMS is already generating, and it optimises the sequence of those moves in real time, so that pickers walk the shortest sensible route, rather than the one the system happened to issue first. The numbers our customers are seeing are consistent: 25% faster picking, 50% less walking, 20% lower labour costs, 30% higher fill rates, and productivity improvements in the 15 to 30 percent range.

And the most important part of all of this is that none of that requires you to rip out a WMS that works. None of it asks your people to learn a new interface or means that your IT team have to sign off on a data migration. Optioryx Pulse is a targeted piece of intelligence that can be neatly dropped into the one single process that dictates your cost-per-case. It’s probably the sort of thing Paxton would call “optimising at the edges.” Which is fine by us, because the edges are where, in the real world at least, businesses actually see improvements.

The MHI report is right, though, about the scale of the shift. We’re at a tipping point. The operators who spend 2026 “evaluating” are still going to be evaluating in 2027. Whereas the ones who decide to move now can expect to see their numbers change within weeks rather than quarters.

Picking is where to start

It’s where your labour cost concentrates. It’s where your throughput is made or lost. And as luck would have it, it’s also the process where AI has the shortest route from pilot to payback. So, if you’re reading the MHI report and wondering where to start… our advice is to start with your pickers.

Book an obligation-free discovery call with Socius24 now, to see how effectively and efficiently Optioryx Pulse pick path optimisation works with your existing Dispatcher WMS.