 We're going to move to the panel session if we can go to the polls that we have out. So one of our panelists was stuck in a victim of traffic coming into Boston, so we'll be a smaller panel, which means you can ask more questions. Before we get ready with the polling questions, you can see Slido. If you guys have used this before, just take your phone. You can use the QR code or type in slido.com and the password is CTL. So I just have a couple of ones, and while you're doing this, Bill, do you want to introduce yourself? Yeah, Bill Stenger, Connaught in North America, going on for almost two decades it's hard to believe in automation and supply chain. My career spanned across service and operations, parcel, airports, and the last seven and a half years purely warehousing. So pleasure to be here and thanks for letting me join. All right. And so while we're waiting for the polls to go, let's ask just some quick questions. So Bill, I'll go with you so that Mohan can catch his breath from his talk. What do you see as the promising trends in automation that you're seeing right now? What do you think is happening and will be happening in the next couple of years? I think Mohan did a pretty good job summarizing everything. So we're done? We can call the panel? A longer break? I think it's a beautiful thing about having a Walmart, having the capital and the size of the organization, the speed in which they can go and try, try fast, fail fast, and really innovate. And we're fortunate to be a partner with Walmart as well. And I would say that the trends going on right now is consumer behavior is dictating what the supply chain will look and how logistics will look. That's without question. And additionally, the labor topic's not going anywhere. 9 million open jobs, presently challenging the supply chain today. It's only going to get worse. So it's not that we have to automate to replace people. We have to automate because the customer demand is dictating so, right? With all that automation, we have to be mindful also of the importance of the interconnectivity, the intelligence. There's so much data being generated by these systems. It's impossible for a single human being to, let's say, to process all this information. So it's not just a physical automation, but I agree with you the digital software analytics and BI behind it is, let's say, the recipe for success. It's funny. You brought that up on the difference between physical and digital. But a lot of digital automation I see is almost for desk work, for like RPA, you know, processing. Are you looking at a different type of digital automation or is that more the processes that someone do for the back office? So there are two. One is, of course, the automation of the regular tasks with physical and digital as well. But the next layer of automation that I see on the digital side is about the planning. I mean, historically, you know, there are lots of gut feels that go into a lot of the decision making that happens in the business, but replacing that with AI based intelligence based solutions in terms of, you know, how much to deploy, where to deploy, when to deploy, when to flow the product, how to flow the product, all of that. I don't think we have, or at least we are transforming, but you know, that is one big opportunity for us to go incorporate intelligence into them so that it not only replaces human back, but it actually makes much better decisions than what a human would do. Do you think they'll replace the human or will there always be a human in the loop? I said they will always be a human. It's not never, it's never going to be an autonomous model for, because of business priorities change. And you always have to have humans, you know, monitoring the decisions. But the depth of intelligence that will go in is should be much more than what it is. And that's what we're really focused on. I'll go one step beyond that. I think that, you know, with the challenging labor force situation, we're not seeing employees being retained as long. And there's a lot of knowledge being lost along the way or lack of experience. And what I see the tools, what they're doing is empowering these, let's say, less experienced operations with intelligence and making informed decisions or suggestions so that they're not afraid to make a decision. I think that very often operations, particularly new operators, are afraid to make the wrong decision for fear of backlash or potentially losing a job. I think these tools, these BI tools in conjunction with ERP systems, making informed decisions on how to run the operation more effectively is where things are going. I just want to add Chris, which is on the physical automation side. I think, and I probably should have probably mentioned this before, but the associate feedback has been through the roof in terms of, you know, we are seeing that in data on a daily basis. But reduced turnover, our NPS scores have beaten our plans. Then what we really anticipated was deployment of automation. The associates love the job. They really want to come to work in an automated building. That surprised you? Usually you hear from labor, there's a lot of resistance to automation, fear of losing the job. So that's the change management part of it, which is, you know, we have to hand hold the associate to be able to upskill them. But one of the things that may or that might be critical in this is being able to get the associates trained to take on the new job smoothly that then helps the associate not fear automation as opposed to embrace it. And that is something that we're doing in a big way through dedicated paid training programs and education for our associates through our Live Better You program, where we actually pay for their tuition. We train them on the right skills so that they can now still continue working in the same building but not have to walk for like nine, 10 miles every day. So can you tell me more about the training programs you've done? Because that's a major upskilling from walking nine miles a day to actually monitoring automation and robots. Absolutely. So Walmart has launched this program called Live Better You, which is a fully paid tuition scholarship for our associates to get trained on specific areas could be cell operators, maintenance operators, systems technicians. And of course, it depends on what specific role that you're targeting. We would actually provide tailored training for the associates fully paid by Walmart. Therefore, you know, the associates would not feel that, okay, I cannot do this new job in this building with automation as opposed to now I have a much more ergonomically favorable job than having to kind of walk nine miles or work in like subzero temperatures. I mean, that to me has been the biggest game changer for us in from an associate base. And the overall, it also means that we are not replacing people or we're reducing the headcount. If anything, we're going to be same or even more associates in the future than what we currently have because the business is growing at a rapid pace and that will more than compensate for the automation. So one thing I'm just curious with the automation and you brought this up Bill, do you find, how do you empower them to make those decisions to problem solve? Because that's what you need them for as opposed to if you're just a picker you're following orders, it's a mindset. It's not just an upskill of individual skills. How do you let them know that they're empowered to solve problems? And is that a problem or does that happen naturally? I don't see that as a problem. And in fact, I think it is part of the culture where at the end of the day, the associates understand that their job is to serve the customer. And how we do it is what we're really debating about. And once that we have aligned on the fact that this is what you're ultimately what you do is manifesting for our customer. I think the adoption rates have been significantly better than what we anticipated. And I don't think we really felt that kind of resistance, Chris, in terms of our associates taking on these jobs. And I mean, like I said, I think if anything, they are actually wanting to do these jobs than opposed to going back and working in a traditional manual building. So that to me has been a big revelation. Well, how are you seeing this at other companies? What are you seeing other companies do? Because I assume you have more customers than just Walmart. Yeah, I think Walmart is again, the resources and the knowledge and the people you have. You're sitting at different echelons, so to speak. There are a lot of smaller regional retailers, startups and things like that, that really are looking for the experience and the know-how because they don't have it in-house. And I think that what we're seeing is that they're looking for partners that can tie everything together. There's not a silver bullet, if you may, for the supply chain and automation. But I think that it's about finding the right strategy and the right partners to go on the journey with you and to own it together. And I think that is something that is becoming much more prevalent now because people don't feel like they used to. I think automation was a very scary topic. And I think that automation now is not, it's the definition of automation robotics, you could argue all day long about what that really means, but I think automation has been around for hundreds of years, 100 years with the manufacturing and conveyor and things like that. It's the know-how and powering them with things like ASRS with AMRs and robotics when they don't necessarily have the resources in finding the software and let's say BI leaders in their space to drive that ball forward more rapidly. So in your other companies, you haven't seen what has been the biggest hindrance to adoption of automation. It sounds like for Mohan, for Walmart, it was not the labor getting them engaged, the workforce getting them into it. I think it's the fear of the unknown. I think that the people making these decisions is supply chain leaders that are putting their careers on the line to fire behind a new automation partner. It takes first off years to develop them as we know, it takes years to develop the right solution. And then when it's all said and done, you've designed something and the business may have fundamentally changes like COVID did for all of us. I think that there's the fear of adoption because there is a significant capital investment in the early days. I think now you're seeing startups and new technologies that you can fail fast and much more affordably, the supply chains can, take these decisions much more rapidly without having to go to the board for the $20, $50, $100 million in the capital. You seem to be that automation used to be heavy investment and you lose all flexibility, but you're really efficient, but that's changed over time. Mohan, how has that changed? Is there still a trade off between flexibility and efficiency? There is, and that is one thing that we constantly evaluate. And it's not an easy problem because there are, and it also kind of is a parameter that in the decision making of which automation you wanna go with. Flexibility is one of the big factors to say if the business priorities change, which is oftentimes it changes, what I see as a plan for three years out will not be the same next year. In the wake of changing business priorities, can your automation still be the right choice? I think it is an evaluation that we will have to do regardless and make sure that it can support changing mix, it can support changing volumes, it can support different shopping behavior from the customers. So we trade that off and then we all still make sure that will the capital investment make sense? And are we able to pump the right kind of volumes through these assets to be able to lower the overall cost? Because that is a big factor in the decision on automation, which is do I have enough volumes to justify that automation? So a lot of the stress would be on the pre-work to make sure that it is robust enough and you are accounting for all business changes that can potentially come up. 100% agree with you. And I think that you can have the vision and again the vision changes, but I think the goal is to really identify technologies that are modular and scalable and that you can kind of bend with the cost or demand and you've nailed the head there. I think that's what's really the main drivers towards automation today is partners who can be flexible with the business and not to say like you're stuck with this, you're gonna depreciate over 10 years and you're just gonna build something new. I want to add one example just to kind of bring this to life, right? So think about our e-commerce fulfillment capability where we are an ASR system that is bringing products in to us to customers. So which is the core capability of an automation system. Now, if the business changes, priorities or things like that, we still want to find alternate use cases for the same technology and that is what we evaluate and we have some backups to say, okay, can I use this for supporting the stores, store distribution? And what more do I need or what changes do I need in this automation to support that? Now you have two or three use cases for the same automation technologies. Therefore, if one doesn't pan out, we still have options to utilize the asset and still add value to the business. And that is something that we constantly, as my group and strategy, that's literally what we do day in and day out to make sure that any installation that we do stands the test of time. Okay, so let's look at the poll really quick and of course it's totally spread out. It's almost a normal distribution, right? With average of five. So I asked how, what is your current level with you? And so we have a wide range here of people, some of them and kind of skewed to the lower side. So Bill, what advice would you give a company that is low, that is below say four on that scale of automation to dip their toe in that automation? I would say the one thing that is a common problem across all leading supply chains is the inability to communicate across channels. I think you have a C-suite vision. I think you've got a supply chain executive vision, an operator in a retail. And I think that very often you'll have silos. So I always tell our client the number one thing is that we've got to have a buy-in across the entire organization of where we're trying to go with this project or this idea. You would be surprised how very rarely that happens. So- A dish some dirt, give us some examples. No, no, no, no, absolutely not. Career boom, career boom. They're supposed to all agree with each other. How are you limiting move? Yeah, exactly. And we're promoting supply chains across every industry you can imagine. But I think that in the past, I mean, I've gone through this experience even recently, it's that you build something with a vision and where you're trying to go with it. And then if you don't run that site as was intended, it has detrimental impacts to the overall success of the project, which in turn affects the conversation on the IRR with your leadership. So the communication is critical to the success of any project, in my opinion. Bring up the word silo. What I thought was really interesting from your video on the making of the pallets where it improves the store. That does not help the warehouse that helps the store. And so if you're siloed for efficiency measures, you're actually taking somewhat of a hit to improve the efficiency somewhere else. Was that a hard sell? No, not at all. You think about this, right? One of the things that we really wanted to, it's about just aligning on the fundamental principles. And for us, reducing complexity at the store is at the core of what we want to do. If it means that I need to complicate my upstream supply chain, but I can actually have millions of associates that are in the store focused on servicing our customers. Why? Why are you simplifying there? Is that because there's so many more of them or because? Because the role of the associates in the store, the way we see it is about servicing the customers and not moving products and stocking shelves. So we would want to take those hours away and translate those hours into talking to customers and then making them feel at home in the store servicing their needs. And that to me is something that is the core of our existence as well, which is let focus the time on the customer. And we would, this palletization, by the way, the other thing is we are, while we think it is complexity upstream, but with the automation that we're bringing in, it is not. And we are able to build pallets in, I'd say 15 minutes, 10 minutes, right? And the pallets are much more intelligently layered. So you could have a layer that goes to an aisle, the next layer that goes to a different aisle. Imagine humans trying to figure this out and trying to build this pallet manually. But you always had that one expert who could build a perfect pallet. Was he resistant to the machine? Probably the one expert, but the rest of the 99% of the associates wanted the solution. And the tall pallets, and then the ability to go nine feet high and having a step to be able to do that. I mean, those things, you cannot even really accomplish with humans and have them really focus on the bugs and making sure that the system uptime is high, making sure that they provide real-time feedback to improve the productivity of the automation. So to your earlier question on empowering associates, that's where the role of an associate comes in, which is they are still adding value to the business. They are continuing to make sure that these automations are working at their levels of productivity and their uptime is 99 plus percent and so on. So that to me is where the value comes from. You tie back in the GS1 piece here. I mean, in the topic of pallet build, we instituted a program in the last few years where fit for automation, where we're not collecting length, width and height dimensions. We're collecting 60, 70 attributes specific to a product so that we can get the densest pallet build possible, but also be mindful of stability, rigidity and things like that. So we don't have shrink happening in mid transport. So these are huge topics. I think that we as Canop have put emphasis much further up the supply chain and something as simple as master data that's affecting automation because we all know automation's only gonna grow from here, right, based on labor. So you brought up the D word, the data. So everyone's master data file is usually the worst shape of anything but automation relies on that. And well, I know the next session we'll talk more about data, but how have you seen the automation that can only be as good as the data you feed it? So has this changed the way you've seen in other companies where the quality of their master data or is that an Achilles heel for these systems? I like the second open. I think it is absolutely. You can talk about bad companies. It's absolutely an Achilles heel. I think everyone in this room can say that if anyone tells me their data is good, then I would, I'd have to call you out on it. But I think that it's only as good as the operators are running it. And I think what we have to do is we have to empower the operators with tools that are easy to interact with to collect the proper data. I'm a big proponent of putting that data in a data lake and sharing that across your supply chain because the name being grocery, I'd say I probably 70% of my skews are consistent across the supply chain, right? And if I can ensure that I've got a site where I'm capturing really accurate relevant data points, I can benefit the entire supply chain network, yeah? So let me go back to something that Mohan that you said, making it simple. I love that the idea to make everything simpler for the associates at the store. Can you share with us any numbers as far as what percentage of their time is now spent not on back office things? Has that, I assume that's reduced? It has reduced. I mean, I cannot publicly share the numbers, but what we have seen is the unloading time, the need to re-sort these boxes to aisles. And I mean, if you walk the store, you'll see that the amount of time it takes to unload, to re-sort, to stock and so on, versus the opposite of just taking a parrot all the way to the shelf. We have seen significant improvements and part of the reason why we're investing tens of billions of dollars on the upstream supply chain because we are actually are starting to see the value both in terms of our associate experience and customer experience as well. And are you seeing the same thing at other companies? I think I think you've, I mean, again, Walmart some reasons it nicely. I think you guys got a pretty good figure on that post. I would agree. Okay. So looking at the poll, I asked for one word, describes your experience in implementing automation. This is actually weird. Usually you show the answers and people are like lemmings. They log on to things. This has so many different things. Some resistance, complexity, exciting bias. What have you seen, Bill, in your, what do you think the biggest hindrance that companies you work with that is preventing them from going further? I think that the fear of complexity and these systems are complex, don't get me wrong. But it's our job in our industry to basically make that complexity much more simplistic to deploy. And I think that's where the intelligence and the software layer of these systems is really moving exponentially forward. We also see that you can deploy, let's say scale down versus the solutions. You can familiarize supply chains with technology in advance of going, doing the big project with, with pilot sites. And I think that that has also been something a way to mitigate that fear of the sense of complexity and reduce the resistance across the organization. And without question, you know, automation is not, it's gotten more expensive. The inflation has affected every one of us. And that has been a topic and discussed, you know, the return on investment and where automation fits in the equation. But the overarching message from the supply chain leaders to me is that I don't have a choice. I have to do something or I'm gonna erode my customer base. Their service levels aren't changing. They demand it more conveniently. They demand it faster. They demand it, you know, at the best possible price. And supply chains have to take, let's say, the steps and invest in automation to retain their client base. And so if we look at the whole chain of operations for the supply chain, where's the last frontier for automation? Is it picking of each's? It sounds like I thought pallet building would take a lot longer. I'm actually so automated pallet building I thought would take, wouldn't have been as complete at this point. Where do you see the last frontier? What is not being automated that you think will be? I think when it comes to distribution, every aspect of it, I think we would want to infuse some level of automation. It may not be 100% automated because there will be products that still need manual handling. In transportation, I think, you know, we are obviously launching drones in different parts of the country. We are launching autonomous vehicles that could do point to point. And then eventually, you know, we would want to do long hauls also through some autonomous vehicles as well. Like to me, those are those, from a transportation perspective, there are some areas that probably are a little out in the future. But when it comes to distribution, I do see that we will continue to disrupt ourselves as we go into the future. And there, I mean, there are different value propositions based on the complexity of the task that you do. If it is about picking each's, the value of automation is pretty significant. If it is a pure pallet flow building, you know, the value may not be that much. So that's typically how we would even prioritize our investments in automation. And there is almost a pecking order that we would go with. But eventually, with the scale and size that we would have, I think we will end up automating a majority of the supply chain and having our volumes. I mean, one of the metrics that we use is what is the percentage of the volume that actually is flowing through physical automation. I would love to also see a metric that says what percentage of the volume is where the planning and execution is also automated from a systems perspective. And then you have those metrics and those have to go north all the way up to like 90%. And that's when I think truly we would be disrupting ourselves and really having an impact on cost itself. Bill, what do you think? What's the last front to you? I do the last mile autonomous delivery. I think I do see that be as a significant hurdle to overcome. I mean, there's been obviously massive, as I'd say, advancements in the space. But I think that's a big one for me. A few years ago, the robotic each picking was a topic. The problem was it existed. It was division technology behind it that was really the hindrance, I would say. And we've got a partnership with a company that has advanced that division piece where it's a self-learning vision camera that every time we deploy it, it's not happened to learn from scratch. It actually takes all the learning. So it doesn't have your specific data inside the robot. It just knows the type of packaging, it's handling the shape, the dimensions, the weights and things like that. So every time we deploy a new robot, it's immediately hitting the ground running. And we're seeing now that robot, even in grocery, 60 to 70% of items can be picked by a robot now for an e-grocery basket. And in traditional supply chain and pharmacy, near obviously 100%. And that's a general merchandise, 85 to 90% can be picked by an age-picking robot now. So I was surprised how fast it advanced in the last five years. And one thing I would also want to add, one is just the pure value of automation at a transactional level. But different companies are at different scales. So I mean, we have the fortune of the scale that we have where we can justify automations with the volumes that we would flow. But now you go down the spectrum of companies, I think one is you also need to make sure that you have sufficient volumes to sweat those assets to be able to justify automation. So then the last frontier is not necessarily just what I just said. It could even be each picking is manually still better because the fixed cost of automation may not justify a pencil for the variable cost benefits. That's a really good point. But scale is a double-edged sword, right? Because it's hard for a company of your size to incrementally add something. When you make a step, it's a big step. And so one of the things I want to go back to the pallet building because I'm fascinated by it. To make that work, you need to know this format of each store. Is each store format and Walmart identical? Or do you have that so they need to know which store they're going to? Because imagine the layout changes or modify season to season. So how is that coordination? Is that a hurdle or is that something that's easy to do? So we do need not only the format of the store, we also need to know the aisle composition as well. And eventually we would want to get to a four-foot section. So if I had to go stock a four-foot section, can I create a very dedicated load? It could be a layer on a pallet that just goes to that four-foot section. And I mean, it's not like we have everything buttoned up or ready to go, but we are continuously evolving. But if that information is fed in, comes back to the point that you made on data, if that information is fed in, the automation has the ability to play this Tetris game to build the most optimal three-dimensional pallet so that can be taken to an island and stocked. So capabilities exist. We are obviously feeding in the more data, the more accurate data we feed in, the better we can get. So it's just not from here. And it's unfortunate that a planogram is not always available. So I think a planogram, in my opinion, has been the best way to maximize the density of the pallet build. But additionally, if you don't have that information, particularly I'd say in the grocery retail space when you have multiple banners under you, I think that it's very hard. So you start grouping by commodity as opposed to grouping by planogram. It does have some negative, I'd say detrimental impact to the overall density of the pallet, but grouping by commodity and then applying certain rules on a number of aisles of travel distance built on an average pallet. So back to your comments earlier, there's a trade-off. You gotta figure out, okay, what am I trying to achieve? Maximum density in my truckload to the store? Am I trying to minimize the labor impact and maximize the customer experience? These are all things that have to be discussed across the entire supply chain organization on the way forward. I'm gonna go to ask some questions that are coming in. Thank you for putting these in. You can also upvote the questions as they come in. I won't always listen to your preferences. That's why I'm the moderator, but you can actually signal which ones you find more interesting. Let me ask a question to you, Mohan. How do you prevent people from blindly trusting the recommendations of any automation or orchestration any decision support system has this challenge? You have people who follow blindly or totally ignore it. Was that a challenge? And how did you train people to be able to make those kind of decisions? I mean, for us, it just comes back to really putting the facts out there. And the other thing that we rely heavily on is pilots. It's sometimes it's really difficult to predict to estimate no matter what math and what science you use and all of that. You still have challenges in being able to pinpoint the true productivity impact of an automation solution because we don't know how it actually manifests inside an FC or a DC. We don't know the pallet, the cases come in the way they're supposed to come in, the dimensions are different. So we have all of these challenges. So what we have done very well over the last five, six years is we've piloted all the technologies that you've seen. We have had pilot instances in different parts of the country, much smaller scale, but enough for us to be able to have a much greater level of confidence. And that was when we then infused billions of dollars in kind of scaling that automation up across the supply chain. So to me that helped really bust all the myths and really bring folks to the table to say, okay, now it's not any subjective decision anymore. We have the facts, we have a live example. Now let's make a decision on do we want to invest in automating that particular building? The question, if you can give this, if you can't just say so. What percentage of your warehouses have this level of automation that you showed in your videos? What percentage? I can't- That's a hard number to say, because they're not equal. Exactly, it depends on its units, number of buildings, number of SKUs. It's very different in terms of how you look at it. But I can say this, over the next five years, I'd say we have a majority of our network, perishable, ambient, e-commerce, would be flowing through automation. If that helps, I mean, we are on the journey. Different automations are at different stages of their evolution, but I think we have a very clear plan of execution over the next five years with proven pilots as well. So we are confident that a significant portion of our supply chain would flow through automation, at least for case order filling, each picking and pallet building, as well as other market performance centers in the back of the stores as well. So that's how far I can go. Okay, that's fine, that's fine. Another question for you. How do you balance transportation and fleet optimization with dynamic DC sourcing? It's a total cost equation. So depending on whether it's e-commerce versus stores, I mean, one thing we do on the stores is what we call stores realignment. So we constantly, so we have 42 distribution centers, 4,600 stores, roughly, call it 100 stores per DC, but it's not always 100 for every DC. We constantly keep changing it, just based on the changing dynamics of the demand. That's one. That balances the cost of transportation to the distribution. In e-commerce, what we have is this, we call this multi-channel sourcing engine. MCSC is what we call it. It basically, every time you place an order, in milliseconds, it actually determines what's the best way to source the order from, which means which FC does it have to be fulfilled from, which carrier or which transportation model we have to use to ship the product and when we can commit the order for delivery to the customer. So when you open up an item, it actually shows it's available for next day delivery and that computation is happening in milliseconds and the number of possibilities runs into billions. So the sophistication of the engines that we build is so critical and hold this new technologies of generative AI and AI will play a big role in that, but that engine is making the trade-offs of better fulfillment costs but longer transportation or shorter transportation but worse fulfillment costs, ultimately making sure that the customer's SLA is met. So that is a non-negotiable and then we work back to determine the best sourcing model for that. Make sense. So Bill, a question for this and tell me if you can answer this. Because the question is, how are you selecting and are testing for automation vendors? You're from the other side. How are you seeing the, I guess it's an RFI or an RFP process. How are you seeing that changing and can you give us some insights and how that's being done? Maybe I can make a request. I think that what you described is actually the right strategy, which is a typically, depending upon the level of investment, these are multi-year journeys together where you are identifying the right partner to work with. Very often you get these phone calls or these emails saying, hey, we'd really like to get you guys to propose something. Here's our master data. Here's our peak and average order profile for last year. I need a price in two weeks and love to hear from you and hope you participate. That doesn't work and I'd say 90% of the time Kenop would decline your request. We have a personal reputation. I think the retailer is the one they want to keep intact as well. And I would say please take the proper time and do your due diligence and don't rush this. And as much pressure you might be getting from senior leadership, I think that you've got to really pay this the attention that's due. For me, a successful RFI or a period of time to analyze the right solution is probably a four to six month journey. Pretty common, I'd say. But if you start talking about things like the next gen program for Walmart, I mean that was a four or five year journey to figure out what was the right solution with many different people, with different ideas on how to solve the problem. So give it time, don't rush it and get it right the first time because I think that's what will lead to building the confidence and reduce the sphere of complexity and risk in an organization. Anything you want to add about the bidding or vendor selection process? No, I think he's nailed it. I think it's so important to go slow to go fast, especially when you're deploying these big vending machines in million square foot buildings that you don't have an option to walk away. So doing the right kind of due diligence is extremely important. And it's painstaking, but it is worth all the effort for you to be able to save millions of dollars on the other side. So I fully agree with it. And I think there are standard RFP templates and we go into a great level of detail, item profile, order profile, package profile, the seasonality that he talked about, the unit capacity, case capacity, the server versus the automated storage. So we really go into that level of detail to be able to plan the buildings out. And like I said, all of this needs to ultimately still be, despite doing all of that, we still want it on a pilot before we take a decision. We asked a question about pilots and I'm curious from both of your perspectives on this, can you give any sense of the percentage of pilots that don't go to scale? Is it 10, is it every other, every pilot successful? Every pilot is not successful. That is for sure. I had to pin a percentage, but we've tried a lot of technologies over the last, I mean, I've been involved for 10 years and I've seen, I was telling Bill earlier at breakfast today, I've seen so many technologies come and go that, but now what I'm seeing is a concerted effort across every network within our business with a powered by proven automation. So it almost feels very different now than what I've seen in the last 10 years. So I'm very confident and excited about what we will do with it. Very difficult to put a percentage to it, but there's always instances where we, there are several instances where we actually fail and we moved on. Is that something that the company has to be comfortable with culturally, because a lot of times no one likes to lose, because the pilot is someone's baby. And so if it doesn't go, then you're killing its baby. Is that something that you see in other companies, Bill? Yeah, you have to be willing to take on that risk. And I think that you, yes, one person's feelings might get hurt if a pilot fails, but there's a lot more at stake. But they might get fired. They might get fired. A lot more at stake. It may have a hard time finding a job in the supply chain. It depends on the project ground. But I would say that the pilots, we liked them, but we also hate them, because to your point, if a pilot, especially when you're doing something for the first time, things that you could not have foreseen, it's always the thing that you don't anticipate that goes wrong and that you have to work through that. And there's a certain level of visibility at the C-suite, right? That if this pilot doesn't do some work, it will dictate the strategy. And I think that we love to do pilots because you want a new opportunity to solve a problem, but it's also the risk of the pilot not achieving what you set up to accomplish together. Do you sense there's a greater challenge for failure across the industry, not just your company? Yeah, I do actually. I think they're scaling it down. I mean, if you look at- Sounds big of a bet. I mean, if you look at what we set out to accomplish, I guess when we started it with takeoff technologies, the MFC program, the MFC program was a way to- Can you explain what MFC is? I'm sorry. Micro fulfillment or the market fulfillment center that Walmart has, we wanted to establish a way of going to market rapidly to service the e-grocery market. E-grocery adoption was much slower, was being far uppaced by, as I say, general merchandise retail. And nobody wanted to go and build a large scale grocery DC because to your point earlier, the demand wasn't there at the time. Throw gasoline on the fire after the pandemic, but the MFC served as an enabler to prove that look, you could serve the last mile, you could make e-grocery profitable for the retailer. And I think that's where I started to see the shift. It was around around the same time with some of the pilots at Walmart I saw doing, we're doing that. I think probably in the last three or four years, really the pilot conversation is really, I'd say, sped up across the industry for me. Got it. So a bunch of questions here, we're not gonna be able to get to them all, but Mohan, is there something you'd like to ask your suppliers to do to reduce the complexity down at that store level? Are there things that they can do differently? Yes, I think one of the things is we have, work with suppliers to create the right format of the product that comes in. That is sometimes, sometimes there are products that are not automation friendly. And like I said, you wanna make sure that every item to the extent possible flows through the automation technology, which means that we may have to work with suppliers to be able to do that. The other instance is, we have drop shippers, or what we call direct store deliveries. Maybe that's the right word. So where we, the supply is directly fulfill the orders to the stores. I mean, with automation, we would want to be able to really utilize the efficiency that we have created to be able to bring a lot of the volume in house, which, because the more volume that we can control, the better availability for the customer. And if I really think about that, keep the cost aside. If I'm able to make the product available for a customer who walks in or wants to pick up or who wants to deliver online to his home, if that is being solved, we can really solve the rest of it pretty easily. And so taking control of the supply chain, I think is a key enabler of that. And automation is an enabler of that. And therefore, you see that cascading effect of the fundamental thing is, can we automate, create productivity, create capacity? And then we can do a lot of these other programs with automation. Can I make one request to the suppliers as well? Sure, go for it. The load carriers, you know, the pandemic was, it was the number of whitewood that was entered into automation versus the Chap'n Pico type of pallets or that. If we could really be mindful of the impact downstream to your customers, the load carriers that you're shipping to their distribution and fulfillment networks, how that impacts their operation, that would be a huge accomplishment. You're calling the load carriers, the trucking? The pallet itself. The pallets, okay. Yes, so the pallets have become really a problem in the last few years for us, I know, particularly. And I'd say I would love to have someone understand what the impact is. Well, you might be cutting costs or maybe there's availability of a quality pallet. Think about how that's affecting your retail partners downstream in their distribution environment, for sure. On it. Another question kind of from left field. What needs to be changed or added to high school curriculums to better prepare the upscaled workforce required to support working in automation? I'll go first here. I would love to see, I would love to see the retailers, the automation companies to get more involved with helping develop the curriculum that pertains to, I'd say maintenance, electrical, mechanical technicians, the four, I'd say, those types of jobs, software development, robotic engineering. I mean, there's a plethora. I would love to see if you are passionate about, I'd say advancing the curriculum in your school system and your district. Like there are so many companies scattered across this country that I would love to get involved and participate. You just gotta reach out and ask. I don't think you're gonna necessarily find the level of productivity, the level of engagement on the high school level right now, unfortunately, but I know that we are really active in the collegiate level, trying to, and I'm sure you guys are as well, to try and fill these impossible positions to get right now. Any opinions on high school? Yeah, I mean, I have a high school kid, so definitely have a ton of opinions. And my youngest son, nine year old, he'll give me feedback on Walmart deliveries. He does benchmarking for me with Amazon and he'll tell you, you should have done this better, you should have done that better. So they are a lot more knowledgeable than what I thought they will be and we constantly get feedback. They understand e-commerce, they understand parcel shipping, they understand last mile delivery, and it's a bit surprising for me to be honest. But I think the awareness is great, but to Bill's point, if there is an opportunity to create programs that are tailored towards some of these capabilities, being able to understand what shipping e-commerce is all about, automation is all about, it really prepares them well as they get into their undergrads and beyond and really positions them for success. And I think I've seen examples of that happening in schools with other companies and something that we would definitely want to do as well. And of course, we also want to work with the universities in a big way to be able to really engage and tap into their capabilities to be able to disrupt our business. Any universities in particular? I think MIT probably will be one of them, I guess. So we only have a minute or two left. Let me ask the last question. AI, is it hype or is it reality and where is it on that spectrum? I'm gonna go first, you want to? I'll let you go after me, I'll open it up and let you know it's a tee-up so you can go home run. I think it's not hype, it's real. It's a digital twin, it's too often the vocabulary. It's like what we're describing, is it really AI? I would say that I see that the generative AI does play a place, particularly on the customer service as it relates to how we support the uptime of our automated systems. The systems are deploying a form of AI that they're collecting millions of data points every second to make informed decisions or to empower operations to have visibility. So I think AI is also making suggestions not only on the automation level but also on the operational level. We used to talk about labor management. I think there's more of a focus now for us as labor optimization. Being able to identify through the ERP system, when orders are coming in, what type of orders be able to make suggestions to the operations, to ship operators from receiving to outbound, things like that. AI is very much in play there. And again, it's advancing so rapidly that talk to me in 12 months, I'll tell you whatever. Bring it back. Bring it back. Go on. I agree with all that in the bill said. I'd add that I have a PhD in operations research and I discussed this multiple times, but we obviously fall in love with what we've studied. And what I see is the traditional methods still being deployed in a lot of the engines that we develop. And I have to force myself to say, not mixed integer linear programs, no column generations, no Lagrangian relaxations, and all of those kinds of techniques. Go adopt the new technologies. You're speaking my language. Exactly. So go adopt the new technologies because those are the ones that are going to continue to improve over time. And we have to almost force ourselves to have an AI first approach, prove that AI doesn't work, and then go back to your traditional techniques. So I think that's something that I'm trying to kind of push even within my organization as well. But the applications are like enormous. I've been kind of thinking about it myself in the imagine like a distribution center where you're able to feed all the data and then run a generative AI model on top of that. And you're able to say, tell me this associate or this particular function on Wednesday what the productivity of that particular is or what were my bottlenecks? How many mischiefs did I have on Wednesday, morning, 10 a.m. to noon? So all of these, you make it a lot more interactive and you open it up for all the associates inside a building. And the power of that in terms of the actions that they can take to improve their work is phenomenal. That's just a very small microcosm of what a typical AI application could be. Now you expand that to the overall supply chain and the use cases are significant. So I think it is definitely a big, it is coming. I do believe strongly that there is value and this will probably be another, call it a disruption that we will see in the coming days. Yeah, I always challenge with what AI is because some people, if you use math in a computer, it's AI as opposed to a generative. So there's that big spectrum of what it actually is and it depends by different people. But I want to thank the panel. Thank you for all your questions.