How XPO Logistics Uses Robots to Meet E-Commerce Demand
roboticsbusinessreview.com - Keith Shaw
Rising e-commerce demand and warehouse labor shortages are not just talking points made by mobile robotics companies as a way to sell more robots. For many in the supply chain, logistics and transportation sector, these are real issues that are being addressed through robotics and automation.
Robotics Business Review recently spoke with Mario Harik, the CIO at XPO Logistics, about how the company is deploying collaborative robots across their business lines to help generate value for its businesses, employees, and customers. It considers autonomous mobile robots, including goods-to-person systems, as collaborative, as well as smaller industrial robot arms. Harik spoke with us about the challenges facing the industry and how robotics and automation are addressing them.
Harik is responsible for the design and implementation of the company’s integrated technology infrastructure. With a technology budget of more than $550 million, Harik and the XPO Logistics team of 1,800 technology professionals address issues across the areas of transportation, warehousing, last-mile deliveries, and inventory management.
From weeks to days to hours
Q: From your perspective, what are the biggest challenges in the transportation and logistics industry these days?
Harik: It used to be that in supply chain, you had to move pallets of goods and you had two weeks to get these pallets to the destination. Today, you have to move units of goods, and you have to get them to that destination the same day or next day to an end consumer or customer. Supply chain leaders around the world are being pressured to deliver on that promise of speed down to the unit level and just-in-time inventory fulfillment, while keeping the costs equal, or even taking that cost down.
We believe that technology is the answer to make supply chains efficient, all the way from how we deploy technology in warehouses, to how we deploy technology that can automate and move transportation across the board.
Q: How are you set up to serve retailers that are facing rising e-commerce demand, which would then be transferred to you, their logistics and transportation provider?
Harik: In our case, whenever we fulfill e-commerce, or in some cases omnichannel facilities where we serve the store and the consumer, we would be the solution provider that can implement automation and warehouses to be able to drive on that promise of same and next day.
Q: How did you approach looking into robotics and automation technologies as a way to solve some of these challenges?
Harik: We have a fantastic solutioning team that can look at any customer problem and build a fully automated, end-to-end solution. But if I had to think about the secret sauce, what are we really good at? Number one is that we manage and build the proprietary software that can help manage the end-to-end [operation] of the warehouse, from where we manage inventory to how we manage the robotics implementation within the warehouse.
The second part of the secret sauce is being able to understand the customer’s needs, and customize a solution that fits their needs, where we can actually move down to a less-than-an-hour fulfillment, as an example.
So again, our expertise for customers is to design the solution for them, implement the software, implement all of the robotics implementation to fit their needs, and then deploy it so they can fulfill on those promises for their customers.
Q: With such a large company and different solutions for customers, we assume that you have a number of different robotics options that you’ve deployed. What types of robots are you using?
Harik: Depending on the form factor of the product, and depending on the customer, it can be very different.
First, we have collaborative robots, or cobots, which work with our pickers and colleagues in warehouses to help automate or speed up a portion of the fulfillment process. We think of a cobot as being an autonomous robot that works with the picker, as opposed to having the picker having to pull or push a cart to pick the product. The robot knows where to go – it shows the picker a photo of the product to pick, and even highlights the bin they need to take it from. Now the picker can take the product, put it inside the robot, and then the robot at the end of the picking mission can go autonomously back to the packaging station, effectively reducing walking time and making our pickers more and more efficient. We typically see about twice the productivity by implementing cobots that are going down picking aisles within the warehouse.
Q: Are your pickers then walking, or staying within the picking aisles and at the shelves?
Harik: We have two variations. One implementation of collaborative robots is where pickers are still walking down an aisle, but we limit what aisle they walk down. So this way, cobots are working on aggregating the orders. So you reduce the walking time by have them only working one segment of the picking area, and not having to walk back to packaging, which could be a long walk depending on how many SKUs you have.
The second type is goods-to-person, which we have implemented in more than one flavor. In this case, the picker is stationary, so the picker doesn’t have to work anywhere, they are at a packing station and the robot brings what we call mobile storage units, or MSUs, to the picker that has the product that they need to pick. We even have a system with lights on the bin where they need to pick it from, we show them the photo and they can pick the product and put it into the proper customer order, which eventually gets packaged in a box and goes to the customer.
This implementation has two flavors – one is where the robot brings shelves of product to the picker, but we also implemented one where the robots can climb a rack of shelves. So instead of the robot bringing a shelf to the picker, the robot goes up a rack of shelves and they pick one bin, and they take it down to the stationary picker.
With these types of implementations, we see four to five times the productivity, but more importantly for us is the safety of our workers in the warehouses. Because now you can imagine that our colleagues in warehouses don’t have to walk miles and miles every day. They don’t have to bend or lift product depending on the form factor of the product. It’s a much safer environment and a productive environment, and our colleagues enjoy much more from that perspective.
Choosing the right approach
Q: Do the different implementation options depend on the types of goods that are being picked, or is it more of a preference of the customers that you work with?
Harik: It’s a combination of both in many cases, and also depends on the physical footprint of the warehouse. You can imagine if you have a warehouse where you already have this massive picking area, but also have a very high number of SKUs, but these SKUs have what we call low-order velocity. The idea is that you have a high number of SKUs being sold, but some of them are only sold once a month or once a week. So depending on the customer, in that case you’re better off having a much wider picking area and having an autonomous cobot go down the aisles.
Mario Harik interview goods-to-person XPO Logistics
If you have a high frequency of order by product, goods-to-person becomes a much more viable option. So it’s a combination of customer requirements and physical footprint that you have as the driver of which solution you go with. One is more infrastructure heavy, the other one is more easier to roll out, but you still have to do more because it’s more spread out as opposed to the robots bringing the goods to the colleagues at the warehouse.
Q: Beyond the two types of collaborative robots, what other robotics or automation are you looking at?
Harik: The third type are robotic arms and industrial robot arms. If you think about the rise of e-commerce, there’s a lot of need for customizing products. For example, one of our customers is a large entertainment company where we do bands for them so people can go to their park. You can imagine engraving these bands and programming them. Activities that are more about customizing the experience need a robotic implementation. In many cases we use industrial robotic arms for tasks like putting marketing paperwork in boxes, or customizing or engraving certain products or programming devices. So here, we use industrial robotic arms that can do repetitive tasks and automate them for our customers.
Another application is pallet building and pallet dismantling. If you think about building a pallet by robot, it’s much better as opposed to having people carrying all these boxes – it’s an incredibly strenuous process. The types of robotic arms we use depends on the product type – if it’s a small product we’re customizing, we’ll use someone like a Universal Robotics. For larger form factor products, we would use a heavier industrial robotic arm that can carry heavier workloads.
Labor shortages, peak demands
Q: A lot of robot companies talk to us about how they are addressing labor shortages – do you face the same kind of labor shortages in your business?
Harik: Of course. Being the second largest supply chain company, we see different markets having different levels of shortages. But for us, the bigger thing is handling peak. When you think about e-commerce as an example on a Cyber Monday or a Black Friday, you could have four to five times the volume depending on the retailer compared to a normal day of deployment. Being able to handle those peak demands – you can imagine that this is what automation has to manage, or the peaks and valleys of demand, as opposed to not being able to do it.
The second piece is in markets where labor shortages are hard to account for. I think in the Netherlands, for example, they have almost a 0% unemployment rate.
Q: When you work with robotics companies, what are they getting right?
Harik: One of the good things for us is that we focus on the customer solutioning and the proprietary software layers, where we can customize the solution on behalf of our customers. A lot of these companies are doing a lot of things right, and we usually partner with the companies that are best in class in service, and how we can leverage what they’re really good at that fits the needs we have with our customers.
Over time, the cost of automation and robotics is going down, so from a cost perspective these solutions are becoming more and more effective. The second piece is leveraging machine learning and AI – it’s another area where these robots are becoming smarter over time and being more accurate over time, which is also very good.
They are also offering more API integration capabilities. So this way with our own proprietary software, we can easily manage the interfaces with these automation systems to automate the entire flow, and make warehouses more efficient.
Q: If you could talk to these robot companies and say, ‘Gee, I wish you could to this a little bit better,’ what would it be? Or what do you expect that they’ll get better at over the next few years?
Harik: I don’t have major feedback on this, because we work well with them and they’re pretty responsive on things we need from them, so there’s no broad category of things I would say, ‘I wish you can do this versus this.’ Overall, just keep on improving accuracy, speed, and reducing costs, so the solutions become more prevalent and more accessible.
Q: How do you work with new robotics companies or expanding relationships you have with robotics providers?
Harik: When I look at it in terms of expanding relationships with providers, with a lot of them already today we are a very big customer for them, given how many implementations we do in a given week or a given month. But overall, we’re always looking for new partners. I usually meet with robotics companies at least a couple of new ones every month. We’re always looking at who is progressing more in different parts of automation, and where we can leverage what they’re doing well, along with our solutions and capabilities for our customers.
