Name: Yard Reimagined: Why Supply Chain Leaders Are Investing In AI-Powered Yard Management
Uploaded: 2025-06-13T17:01:09.067Z
Duration: 50 min 12 s
Description: Yard Reimagined: Why Supply Chain Leaders Are Investing In AI-Powered Yard Management

Transcript for "Yard Reimagined: Why Supply Chain Leaders Are Investing In AI-Powered Yard Management": Welcome, everyone. I'm Tyler Nickel, director of product marketing here at FourKites, and I'm thrilled you can join us. AI is transforming our industry at break net speed. And if you're like most companies, you've probably already dipped your toes into supply chain AI. But here's the million dollar question. Are you focusing on the areas that really move the needle? Today's focus, the AI applications that are quietly revolutionizing supply chains from the ground up. By reimagining how these facilities, schedules, and yard operations are handled, companies are uncovering massive efficiencies. And we're seeing hundreds of thousands in cost savings across operations and labor. And this is where AI meets the real world impact. I'm joined by Carrie Wiginton, our manager of solution advisors here at FourKites. For those unfamiliar with us, FourKites is is a supply chain intelligence company that helps businesses spot and eliminate waste end to end through our intelligent control tower, our digital workforce, and our connected applications. We're also joined by Colin Masson, director of research focused on AI at ARC Advisory Group. ARC specializes in industry focused research and advisory services, offering deep insights since technology trends, supply chain strategies, and industrial automation. While its complimentary publication, Logistics Viewpoints, provides expert analysis on logistics and supply chain innovations, helping professionals navigate emerging challenges and opportunities. After we dig into these topics, we'll save time for your questions, so start thinking about what you'd like to ask our experts. Let's jump right in. From a supply chain perspective, what makes the facility, yard, and schedule such unique areas compared to warehousing and transportation, and why have they remained so resistant to technological advancement? Colin, let's start with you. Thanks. Thanks, Tyler. Really, we're seeing very rapid much more rapid than I thought adoption of, AI, in supply chain much more broadly. But as you highlight here, there's some areas, where the opportunity has been somewhat constrained by lack of the data foundations that you need. Right? You don't have in yards always the the same degree of sensors, but the opportunity we do have lots of cameras. So there's some things we'll jump into a little bit later on. But, and and it's often the forgotten part of the supply chain. Right? We we kind of look at inventory optimization, but kind of product in the yards or assets and containers in yards are often a a blind spot for many when they're thinking about their overall end to end supply chain visibility. Yeah. Colin, if I just jump in, that was a great point. And from my perspective, what I see with companies that I work with, you're right. Really, it's often used the yard within the facilities and primarily when we're talking about drop and hook operations at companies and that could be at a manufacturing site or a distribution center. It's also often looked at as a temporary staging area for inventory, and these trailers are out there. And traditionally, it's easily managed, let's be honest, by, spreadsheets and walkie talkies. But now with the, with the focus on, expense reductions, with labor shortages out there, also with increasing on time in full or ODF fines that, companies are receiving, we're seeing a increased focus on improving their labor productivity, and that would include anywhere in the company, but now they're looking at how can we increase labor productivity in the yard. And then, obviously, how can we increase increase that, the product or inventory velocity throughout the supply chain. And we've historically seen that, inventory can get stuck in the yard. We have trailers being lost out in the yard, or we have trailers out there, and people just don't know what is what inventory or what product, whether it's, inbound raw materials or outbound finished goods. What is in these trailers? So, we're starting to see, from my perspective, more and more interest from companies wanting to, use advanced technology to help, drive those benefit areas of increased productivity, as well as, reduce expenses and improve customer service overall. Absolutely. And especially when you see companies investing a ton of money into warehousing and supply chain, that hand off point kinda gets lost in the mix. So, definitely a a huge area for investment. And, Colin, I'm I'm curious. This last question, I think, is really good for us to start digging a little bit more into AI specifically, because this is a real world place where, actions can actually be driven by, AI rather than having it live on a server or have it do analysis. There's actual real world operations that are impacted. So, you know, just as, like, a a ground, for our audience, who may be new to AI applications and supply chain. Can you explain the key differences between the emerging AI technologies and traditional automation? Especially how do these technologies apply to the yard and scheduling operations? Yes. I think that's probably wise to start with that foundation. And I think you've got a a slide or two from me on on kinda how, ARC Advisory Group, I think we put this definition out of for industrial AI back in 2023. So shortly after the hype began around CHET GPT 3.5, which I think, if I remember now, is November 22. Right? It's only it's not that long ago, but we could already see, a lot of confusion, in the language that was being used by marketing teams, by, you know, some of our our hyperscalers, obviously, investing heavily in large language models and and what has been framed as generative AI or Gen AI for short. But we've actually been using AI and ML techniques. I would argue that in many ways, the industrial sector was ahead of almost every other, sector in its adoption of, other AI and ML techniques. And we could see confusion emerging because everyone was calling everything AI, Gen AI. So we kind of put out this framework, people, processes, and technology and and how they're going AI is gonna impact all of those, at scale. But we wanted to make sure we were thinking about the industrial AI toolset, not just Gen AI, which is primarily large language models, really good at natural language processing. And the slide here you're gonna see is my my my kinda, fallback slide just when I need to get into a discussion with, with someone about what are all those tools and and techniques. And you'll see we've got things that have been around and used for decades for, you know, machine learning, for anomaly detection in areas like predictive maintenance or for quality, particularly in, discrete manufacturing. You see lots of application of computer vision to recognize shapes and, and activities, in the case of video or to to read, drawings, for example. So there's, lots of applications of different technologies, which in many ways and and the next slide. Thanks, Tyler. Kind of, does two things. One, it kinda shows that broad range of AI and ML tools that are being used. And we also kind of researched a little bit, like, where is Gen AI really being, applied. And I think we would agree that it's having a major impact, not to be underestimated. So it has, I think, woken up everyone up to the potential of AI across much broader set of use cases. And in some ways, Gen AI, I think you'll see one down there, has become almost the new user interface. So it may not actually being be the AI technology that's solving the underlying problem or providing the the predictions or inferences. That may be one of those more predictable and accurate, techniques. But it you've you've probably heard the term GenUI. Right? It's because you're starting to interact with assistance or agents, and collaborating with those. And almost every UI that I see now has some component of that, Gen AI, for the large for its understanding of language and its ability to navigate complex systems and simplify them. So I think it's got that importance. The other thing I wanna point out about this slide is you'll see there's agent technology in there. There's other techniques like causal AI that that is much more root cause based so you can trace why it made a recommendation, which is another kind of serious limitation of, Gen AI. And then you'll see neurosymbolic AI, which kind of blends, math capability with some of the scale capabilities that we get from GenAI. So that's kind of maybe the next wave where we're gonna get much more accurate and predictable forms of what is currently labeled, GenAI. And last but not least, in case you're wondering, this is 600 industrial sector companies. This is not financial services or or retail. This is, companies that make and move stuff. Right? That, that that, this survey, was based on, late, just a few months ago. Right? Late at the end of, 2024. That's great points, Colin. Just from my perspective, where I'm seeing interest in AI from companies out there that we're talking to, really threefold. One is, I think the most prevalent we're hearing today is computer vision, and I think later on, I'll give specific examples of where, these different AI technologies are being applied, but, computer vision is one. The one would be, what Colin was just talking about, that generative AI, with the large language models. And then thirdly, definitely is of interest a lot, especially to drive productivity, is the agent based. I think some people may talk, think of it as agentic AI, but the, the agents that can automate, or intelligently automate specific workflows and tasks. Definitely. I think it's really important to realize that even though things like chat GPT and the large language models have really taken the the stage that these other types of AI are, really being developed and really valuable, especially as it applies to physical operations. So, definitely excited to hear more about the, computer vision and some of the other types of AI that we're seeing out in the facilities space. So, it's actually a really good segue into our next question, which is, you know, what makes AI particularly well suited for solving yard and scheduling challenges compared to previous technology approaches? So what capabilities do these newer technologies bring that weren't possible before? Alright. Let me jump in. I think I've got a couple more slides just to make, a a couple of points. But I think we we kinda almost teed off with this. Right? The yard is often forgotten, in terms of it doesn't usually have as many sensors like the factory or the, the warehouse, which is typically laden with, automation technology. And so having other ways of being able to to go and drive productivity with other forms of AI has really changed the game, I think. And part of that is really what we've mentioned a couple of times now is the fact that we have lots of other data sources that have been untapped, by enterprise systems or by control systems and automation systems, lots of data that is there but not really used to to drive productivity, whether that's, your camera feeds, right, in the yard, or it's actually locked away in other enterprise systems. And so, when we start talking about tapping into that data, there's a couple of things that arise. One of the challenges straight away is, alright, it's in lots of different systems, some of them from different vendors. Right? So you're you're dealing with that interoperability challenge, but we've got new ways of addressing that, which I think we'll get into, particularly agent technology. One of its key drivers is its its ability to to go and get that data behind the scenes, if you will, monitor that data, and then move it when it becomes significant or important, to, other systems. The other things that I think are really in the next slide, Tyler, is I think the types of data, right, that, we can now tap into, because of all the different AI tools we have in that industrial AI toolbox. Computer vision, is not new, but it has benefited from all the investment in, AI infrastructure and then blended with, GenUI, for example, it's a lot more accessible. And so now we can take almost, real time camera feeds, and we can, make sense of that. And we can use that to drive, specific, workflows in a lots of different systems. And the way we're increasingly doing that is with this new wave of AI agents. So you can have specialized agents that are just doing very specific tasks or functions, and then use those to kinda drive a multi agent, multi agent and very often human in loop, right, collaboration to drive new levels of productivity in areas that have where where you don't have all the sensors and all the data conveniently in one place. Yeah. Great points, Colin. I just wanted to follow-up to what you just said of some specific examples of of how these AI agents or sorry, the AI technology are actually, being researched and slowly being adopted out there in the field. off, starting with that computer vision, you know, how are people using that in yard operations. And this applies both to what I would call both drop and hook operations where you actually have trailers sitting in parking spots or in areas of the yard, both for inbound and outbound, or what I would call live load type of, facilities where there is no yard. Everything is, when carriers show up, they go directly to the dock doors of either the manufacturing site or the distribution center. But two points on computer vision. One is using computer vision to drive gate automation. So you can eliminate the need of having manned gates, which is obvious obviously an expense to companies to to pay people to be out there and, check-in and check out drivers and trailers. And so that's one way to be able to read, those trucks and trailers that they're showing up, make sense of what it shows, of what it's reading, match those to anticipated shipments that are expected, and, automatically raise and lower the gate arm to allow entry or exit, for that facility. So huge savings and productivity improvement there. for computer vision, as far as a use case would be automating, the yard inventory. Again, for those facilities that have drop and hook operations where they are parking trailers temporarily. And when I say temporarily, it could be a couple minutes. It could be, days or weeks, especially if you're looking at, potentially ocean containers versus, regular over the road trailers. But being able to use cameras to monitor the yard and update the inventory of what trailers are out there, how long they've been there, if they move and, match that up to, the data being used by the enterprise that shows, alright, what is in that trailer. So that's one use case for computer vision. Also, when we talk about, when Colin's talking about those agents, or agentic AI that are assigned to do specific workflows or tasks, we're seeing those being used, in examples of potentially reaching out to carriers on a automated fashion, to get updated estimated time of arrivals into a facility, for planned, shipments on any given day, especially if you're using, what FourKites has a real time transportation visibility platform. If you're using that to track shipments coming into a facility using GPS technology, and for some reason, it's not working on a specific shipment, you can reach out to the carrier using these type of agents and get real time updates and then read that response if the carrier responds via email, being able to read that response using generative AI and actually update the appropriate systems and notify the appropriate parties of when that shipment actually is arising or arriving. And then another use case, that's pretty interesting is, again, using those, one to many different, agents that are assigned to specific task to actually monitor what is what inventories are in those trailers in the yard. Also, monitor the inventory levels of specific products at that facility. Again, whether it's manufacturing or warehouse and be able to do stuff like if if, the agent notices that inventory is running low on a specific item, to be able to to create or generate automated move tasks to a spotter, that says you need to go pick up trailer one two three, that has this specific product on it and move it, immediately up to the dock doors for unloading because that product is needed because it's running low on inventory. So a great use case there where these agents are automating the process, but actually monitoring, what the inventories, are and trying to avoid either plant shutdowns because the inventories are low or, you know, increase that fill rate for outbound customer orders. And then lastly, generative AI and using larger large language models. What we're seeing there is, you know, basically giving the facility operators and, people at the facilities the ability to, in their own language, ask questions to get immediate responses to the questions they have, versus relying on static reports or having to build on reports to find out, you know, what do we have out in the yard today. So those are just, some specific yard based use cases that we're seeing, that apply these these new AI technologies. And and, Carrie, can I perhaps just add a a thought that that I should have stressed more, but you kinda you brought it out kind of the the real time transportation visibility platform, and all the data that you have brought together, is critical when it comes to, solving some of these AI problems or capitalizing on the AI opportunity? We're seeing the fact that you've been tackling that that data problem, and you kinda saw industrial data fabric at the core of, of my industrial AI model at the beginning because it still comes back to that data and data quality that you can act on. So don't think AI can solve the problem without having, access to that data. And you have a critical differentiator. I think there is four guides in terms of the the the data that you have access to and can make available to help, AI do what it does best. Right? Absolutely. I I think that's a a great point, Colin. I really wanna kind of, you know, emphasize that where if you're using something that's bolted onto a WMS, you really don't have the context of real world real world operations at your fingertips. You really don't know, you know, the the ETAs of those inbound materials, so you can't form yard operations. So when you have something that's almost like the connective fabric between your TMS, your WMS, your OMS, your ERP, all these different systems, you're able to take so much context and create really personalized action plans for facility scheduling and facility operations that kinda help with optimizing your operations in the way that you always imagined they could be. So, definitely a really good point to make where your yard lives in a real time world and having it disconnected by either digitizing the clipboard or, you know, using a bolt on system that only lives within one of your systems. That definitely does your business a disservice by not having it interoperate with the rest of your platform. So great points. It was really, interesting to hear those use cases as well, Carrie. So let's move on to our our next question here, and it really gets into, those use cases and how customers are actually experiencing value with them. So we're seeing really impressive results from early AI adopters at some facilities. What specific use cases are delivering the most immediate ROI, and what metrics should operation leaders start tracking to see if they're getting this ROI? Kelly, did you want me to jump in perhaps and just Yes. Talk about it more generally? Right? I think I think you'll be able to give much more specific, examples for, yard management, but, I I just wanted to give some thoughts on, we have done what we call a a pace set to report. Also around about 600, companies respond to that. But where we dug into 25 use cases across industrial AI, half a dozen of those were kind of supplied today and half a dozen in in, production, some in workforce management, some in sales and service, and some in product product design and and development. So we kinda tried to look across the business. And I can tell you that by far the most used at this stage, because they have the most proven ROI and predictable ROI, are a couple in sales and service. As you can imagine, they've had assistance and and been trying to, to automate their call centers and act as the frontline, before you engage a person to talk to. You talk probably are talking with, an AI assistant of some shape or form to either try and answer it or or narrow down the problem before it gets handed over to, a a customer expert, a subject matter expert, technical support, whatever is the right way to go. But in the factory, it's mostly predictive maintenance, and, quality, I mentioned earlier on, are the top top two in there. And certainly happy enough with the returns there that when we looked at the leaders and laggards, the leaders sorry. I'd we called it a pacesetter report, but you can imagine people don't dislike to be called the laggard, but we could identify a 40% gap, between leaders who had already solved their data problem and, felt they had enough digitization and quality in that data, to really start pursuing a lot of the other AI use cases. But to be frank, that is, you know, less than 10% of, our respondents really felt they were at that stage where they could now use AI for, let's call it, autonomous operations. So whether it could close the loop or, you know, pretty much just have a human in loop to sign off. So let's not get too far ahead of ourselves, But what I will tell you is, AI is seen as the most impactful technology, and it's not even close next to robots and drones and everything else. It is, I think it was about 60% in the next on our list of technologies that were gonna be important over the next five years, was down at 30%. Right? And so that's how big that recognition of the ROI and impact they're gonna get. Now is it is everyone prepared to talk about what they're getting in each individual use case? Well, as I say, in some areas like predictive maintenance and quality, yes, it's measurable. You can dig out numbers on those. Not so clear on some areas like generative design and and, some of the more emerging applications. But just the proliferation of, AI agents across a wide range of solutions in supply chain. You're gonna talk more about, the yard management, ROI, Carrie, but, obviously, I talk with lots of supply chain, warehouse management, and planning and scheduling, and and they all have agents, that they're delivering and and seeing early results. So I'm interested to hear from you, Kerry, on I think they must be huge in in yard management because it's been underserved for so long. Yep. Great point, Colin. So, just some real quick use use cases, that are driving ROI from what we're seeing. And again, they relate back to the the AI technologies that I talked about earlier, those three main, types of AI. The one really gets into computer vision and potentially agent workflow, to drive that gate automation, again, which increases the speed, the velocity of the product both in and out of those facilities. So driving that supply chain and, obviously, ultimately, driving that fill rate, to the end customers. We're seeing metrics, some some companies, that are have been through pilots and actually have it in production now. One company was actually telling me they've seen the average check-in time at their facility for a new, driver truck and trailer showing up going from fifteen minutes down to two minutes by using AI to help drive that gate automation. Meaning, you don't need the gate attendant anymore. You don't need the driver to, make a phone call or anything like that. So that's one use use case. Again, from a computer vision standpoint to help drive automated yard inventory, I can't get into specific numbers, but, obviously, by using, computer vision to automate yard inventory and updating exactly what trailers are in the yard and where they are. You're looking at, productivity improvements. Obviously, you're not paying, personnel to go out there with their clipboards and manually do those inventory or those trailer checks once to many times a day, like, many companies are out there doing still. And also, you're saving on fuel because those people, if they're not walking, they might be driving the spotter trucks around. So, cutting down on fuel consumption at the yard. Again, those specific numbers, feel comfortable sharing there, but it is a big driver, in productivity improvement. And thirdly, by using those agent based AI, or agentic AI, we are seeing reduction in detention charges, whether that's what I call driver based, so over the road drivers, if you have those live load and unload operations, as well as, what I call container or ocean based detention, where it's, measured in days versus the driver based detention is measured in minutes and hours. But having those agents monitor when was the check-in time of that, asset, how long has it been sitting there, and then, you know, notify the appropriate parties, or or even automate, scheduled task in the yard to actually get those trailers either loaded or unloaded and out the door from a case of a a driver detention, but also the same thing for ocean based detention where you're measuring, how many days can you actually keep that container until it needs to be returned to the container yard or the port, and actually, monitoring the time left and then, you know, automatically scheduling task, for those spotters to actually move those those containers up to the dock doors and get them either, loaded or unloaded and out the door. So you're you're avoiding those very costly detention fees that everyone probably listening know listening now, is very, very familiar with that you get those bills, either once a month or once a week from the carriers. So you owe, a lot of money. So not only, automating those tasks and avoiding those detention fees, but also being able to provide that, you know, compile that data, that can be used when you go in to speak with those carriers on a monthly basis. And, I'm not gonna touch on the, what I call appointment or doc scheduling, benefits yet because I think we're gonna talk about that. Yeah. Absolutely. Good good segue for us, Carrie, because when we think about scheduling, there's so many factors that are out of shipper and receiver's hands. And I think that AI has a really important role in starting to rein in on some of that variability. So, you know, what practical advice would you offer about assessing readiness and planning implementation of these types of capabilities? So I tend to to base a lot of my guidance on research and talking with, obviously, hundreds of companies in the space, and they're looking for the the best practices. But I think the tried and tested people, processes, and technology, and you don't and you have to look at those together. And we talked a little bit about kind of autonomous operations, and I still think that's going to be, to me, that's not something that's gonna happen anytime soon. Like, I don't see many lights out factories as they like to call them. I think there's still gonna be lots of, human in loop, with, AI technologies for some time to come. So there's a few characteristics I see, with leaders. They are typically they've put AI into their, governance programs along with data governance and security and compliance, all those things you have to do, regardless. But now AI is part of that, governance program, and so they have, you have to go through approval mechanisms and what type of AI and how you apply it. But, also, it gets prioritized in terms of what use cases, prove the use case before you scale out. So, usually, I'm seeing that those leaders have a center of excellence around AI, which is multidisciplinary. Alright? So it's IT, OT, and sorry if I use those acronym. OT is operational technology, to me. So that's typically the people that do look at factory technologies and warehouse technologies, on the operation side of the business. And data scientists, have specific needs that they looking for for the data quality and how to build these, more specific, AI solutions. So I see a center of excellence. I see a structured way of evaluating the technologies. I think the other thing for the FOMO question, right, is, that I think most industrial companies I talk to are actually have more AI than they realize as long as they go back to that you know, we had AI before Gen AI. So when they start to look at it, they realize they're actually further ahead than they thought they were, and they are right to be conservative about when to apply, Gen AI and what it does best. And I think we talked a lot about that. Yes. Gen AI is great for large language models, for simplifying interactions with what have typically been complex systems and for helping us orchestrate agents. But in many ways, we've agents aren't new either. We've had forms of agent, and they don't all have to be AI enabled. In fact, many agents are really just workflow interoperability agents that may not be doing a lot of AI. How you're orchestrating them may be using some form of AI. So I think a lot of what you need to do is think about your governance model for AI. Don't look at it in isolation. Tie it back to what you're trying to do about data governance, how you're managing your people, how you're managing automation, what you're trying to achieve from a productivity point of view. And the other point on fear of missing out is I don't really, at this stage, see in factory and supply chain where we have long standing skills gaps. I see this more as helping us close the skills gap and increasing productivity rather than we're putting people out of jobs. There are other disciplines where maybe that is, you know, if you're a developer and a coder, maybe you're worried a bit more than than if you're in the factory and supply chain domain where we have a tremendous shortage of of, skills, and this is just helping us close a little bit that that gap, and increase productivity. Those are great points, Colin. I just wanted to follow-up just I guess, if anybody out there is even thinking about evaluating just a couple of I won't call them red flags, but if you experience or if you're experiencing any of these, obviously, you may wanna evaluate some AI capabilities for the yard. One would be if you're seeing increasing detention, fines and, charges at your facilities. Again, these aren't at your suppliers or at your customers, but if you're seeing them at your facilities, you might wanna evaluate that. Also, if you're noticing long lines outside your facilities of drivers waiting to get in, that's an usually another indicator that you could, use some help for, say, gate automation. And then there is, as we all know, there is a lot of tribal knowledge, when it comes to yard operations. So if you do notice hiccups or slowdowns when somebody goes on vacation or if there's sick leave involved and you notice that, it might be worth evaluating, these agents because, again, they they do increase productivity. They work twenty four hours a day, seven days a week. So those are just a couple areas that, if you're making notes, and you see these at your facilities, it might be worth look, evaluating different areas. I think a lot of people maybe haven't even thought about checking in how their yard and gate and schedules are are performing. So maybe this will inspire some people to go to go take a look. And I I think that it brings up a a great question about what's coming up here for a lot of companies. So looking ahead three to five years, how do you see AI technologies evolving for yard and scheduling operations? You know, what should those forward looking organizations be preparing for now? Well, it's really simple. One is gonna be, words evolving is, again, it's gonna be driving that productivity improvement, automating those manual tasks, and based on what I say, automating those manual tasks, you know, those non value added activities, looking to automate those, as well as, more of that intelligent scheduling of the yard. So when I talked earlier about monitoring inventory systems and where the critical what are the critical items need to to maintain production or maintain, customer fill rates is being able to monitor, you know, what's coming into the yard and making the decision of where we need to dispense of that product. Does it immediately need to go to the dock door to be unloaded, or can it go out to a parking spot in the yard and actually sit there until it's actually needed? So making those intelligent decisions as it's, as it, leans towards either inventory levels or production schedules and making sure those are humming and and going according to plan versus causing disruptions, or losing visibility of that inventory once it gets into the yard. So, really, productivity improvement and helping increase or maintain, the planned operations is where I see it going. Yeah. I think just a closing thought. I'm really excited about agents. I don't like the term agentic AI because it infers. It's a you know, like GenAI, it's another type of AI, whereas, really, it's about getting productivity out of the full industrial AI toolbox. Right? But I am excited about agents, and you should be too. But you should also be looking at interoperability of agents across all of your different systems. And we've kinda touched a little bit on that. The good news is, back in November 2024. Anthropic kinda released a model context protocol for agents to kinda be able to understand what other agents are doing and the context of the the data and how AI, can kinda share that information across different systems. That was very, recently followed by, support from many different, companies. It was kinda led by let me get it right. On this one, it was led by Google. Right? They kinda released an open source capability that extended model context protocol with a to a, agent to agent, standards. So now the agents know how to talk to each other, not just how to expose through the the model, to AI agents. And then, actually, just a couple of days ago, Amazon Web Services added on to that with some capabilities for actually building those agents themselves to run on lots of different platforms, both cloud and and edge. But the good news is there's this wave of momentum. A to a, for example, is supported by all the major system integrators, the majority of of, application enterprise application, vendors. And so there is this real conscious attempt to make sure that agents are not going to be just FourKites agents, and you can only use FourKites agents. You'll be able to use FourKites agents interoperably with SAP's agents, for example, or with, with, a warehouse vendors, if it's not SAP. And that's important when you're thinking about the future. And that means these agents are going to get more and more productive. They're gonna be able to, you're gonna be able to do multi agent orchestration. And importantly, as we've talked about, I think assume that at least for most of the early stages of what you're doing, there will be human in loop. But as you get that confidence with those agents, then some of those agents may start to do thing things almost autonomously, and you will just need human in loop for, for some of the exceptions. But a lot of the use cases, humans can still do and look across multiple systems when you don't have all the right agents and access, to all of the right data. But I think thinking about agent to agent interoperability is is critical right now and making sure that you identify the the open standards that, that that vendors are committing to, I think, is, is the best we can do to plan for the future. Definitely. It's really exciting times right now, and I I just I can't wait to see, you know, in three years where where we're at, let alone five years. It seems like the pace of change, accelerates every year. So, really quick before we jump into our q and a, I wanted to make sure that we invited every single person on the call today to our FourKites Summit in August here in Chicago. If today's conversation interested you and you're really looking forward to understanding more about how AI isn't just impacting the Yarden facility, but your end to end supply chain, we're bringing all the brightest minds in the industry together. We're gonna have some great hands on activities. We're gonna see this technology live, and you'll hear some great stories from our, presenters who have actually brought this type of technology to life within their own operations. If you need a little bit more, persuasion, here are 10 great reasons why you can't miss FourKites Summit. Not only is Chicago beautiful this time of year. There's a lot of great learning and networking opportunities. Make sure that you guys reserve your spot. Scan this QR code now. If you use the promo card yard, you'll get 10% off your registration. Great opportunity. Make sure you guys are there. I can't wait to see each of you guys there. I wanted to open it up for questions. We are ready to answer those. If you have any, drop them into the chat, and we will, answer them as they come in. Awesome. Alright. So let me take a look and see what we have here. one we got is we are a midsize company with about eight distribution centers. How do you typically recommend prioritizing which facilities to start when implementing AI yard solutions? We actually just lost Cary, so I'll cover this one for him. When you look at your your operations and say you have eight different yards, you don't wanna start with the smallest one because that one may actually not help you kind of suss out those problems that you'll experience at scale. But, obviously, starting with your largest yard may be, you can you can start using those processes that you templatize with, more mid sized yards. So, one of our largest users, they have about 50 sites. They picked one that was right in the middle of their pack. They understood how they integrated with their communications process on-site, their tech stack. They were able to kind of break down some of those IT barriers that present themselves during implementation, kinda get a really good game plan, implemented. They trained the trainer, and with that, they're able to scale pretty quickly across their network. So, really good question. Definitely one that's really relevant when you start thinking about how this can be applied to your real world use cases. Short on time here. So let me go ahead and knock this last one out. I'm curious about the learning curve. When you implement something like computer vision at the gate, how long does a template typically take for drivers to adapt to the new processes? We're worried about pushback. I know that you you probably work closely with drivers, and many of them are very resistant to technology, but this is one that actually reduces their time on-site. You know, they wanna get back out there back on the road pulling freight. This reduces check-in time significantly. They don't have to call their dispatch to find load information. They don't have to worry about in dealing with a gate guard. They're in and out of the site really quickly. Great way for them to increase their loaded miles, which is how they make money. So, we haven't seen any problems getting drivers on board here. So great question. Definitely have seen resistance from drive drivers for other types of technology. This is actually one that they they really like. Yeah. Can I add quickly there, Tyler? Because, certainly, you know more about how people are adopting computer vision and yard management. But I think I would just I always like to relate it to our everyday lives and having just spent the last few weeks traveling from airport to airport. And every time I go there, you know, my face gets scanned and I'm recognized and off I go. So I think, that's another factor. I think everyone's getting more and more used to the fact that, let's not talk about the dubious uses. Right? But at least in our everyday lives, we're kinda used to the fact that cameras can actually recognize the person. If you're in gaming or or you're you're used to using Teams, you're kinda used to the fact that there's a lot of intelligence now that that we inherently trust. Right? And so I'm not surprised to hear you say that, really, you're not seeing much resistance when it comes to this, particularly with the productivity benefits, the streamlining of the workflow. Absolutely. Yeah. Definitely. Great great anecdote. I mean, we're we're seeing that just in our own social lives. So, the the drivers are not gonna be surprised when they see computer vision or computer cameras at the at the gate. They're they're interacting with them everywhere. So that is all the time that we have today. Thank you so much to everybody who joined us. The questions were fantastic. I know we didn't get to all of them. Don't worry. Someone from our team will email you over the next few days to answer anything that we missed. And if today's conversation got you thinking about your own yard and scheduling challenges, just hit reply when you get that follow-up email or reach out to your forecast representative or just hit us up at forecast dot com. We'd love to dig into what's happening with your specific facilities and explore how these solutions might help. So thank you so much for joining us. Carrie, Colin, great discussion. Everyone, enjoy the rest of your day, and happy Friday the thirteenth.