AI and the Autonomous Supply Chain | Duncan Angove, Blue Yonder

Clips
News
More from AI and the Autonomous Supply Chain

Duncan Angove

CEO

Blue Yonder

Blueprint for the AI supply chain: Blue Yonder reimagines cognitive automation

AI supply chains are no longer a future ambition — they’re a present-tense reality, and Blue Yonder has stepped forward with a clear vision to shape what comes next.At ICON 2025 in May, Blue Yonder introduced its most expansive solution suite yet — the culmination of a $2 billion investment in next-generation cognitive applications and AI agents. The launch marks a strategic push to unify decision-making, break down silos and enable machine-speed precision across the global supply chain, according to Duncan Angove (pictured), chief executive officer of Blue Yonder

play_circle_outline Highlights from Blue Yonder's ICON 2025: Unveiling Next-Generation AI-Powered Cognitive Solutions and Key Announcements

play_circle_outline Enhancing Supply Chain Decision-Making: The Role of Pragmatic Agents and Generative AI in Overcoming Industry Challenges

play_circle_outline Need for unified supply chain systems instead of fragmented point solutions.

play_circle_outline Emphasizing Adaptability: Leveraging the SADA Loop for Enhanced Supply Chain Performance Beyond Traditional Efficiency Metrics

play_circle_outline Transforming Industries: $2 Billion Investment in Cognitive Solutions Enhances Product Development and Supply Chain Efficiency for Customers

play_circle_outline Blue Yonder Network enables collaboration across enterprises and optimizes end-to-end processes.

play_circle_outline Future vision for businesses to embrace end-to-end collaboration for machine speed and precision.

Info
Transcript

Duncan Angove, Blue Yonder

Duncan Angove

CEO Blue Yonder

In this episode of the AI & the Autonomous Supply Chain series, theCUBE’s Dave Vellante and George Gilbert sit down with Blue Yonder CEO Duncan Angove for a discussion about ICON 2025. Angove highlights the company’s multi-billion-dollar replatforming effort and how it's delivering on cognitive supply chain solutions designed to move “beyond boundaries.”

Angove outlines Blue Yonder’s pragmatic approach to generative AI and agentic systems, emphasizing precision, real-time orchestration and human-agent collaboration. He explains how new cognitive applic... Read more

explore Keep Exploring

What was the overall assessment of the recent show and what notable developments were announced during it? add

What insights can you provide about the role of generative AI in improving supply chains and the recent developments related to its application? add

What challenges exist in the current approach to supply chain management and the software that supports it? add

What is a method for measuring agility in an organization compared to measuring efficiency? add

What investment and technology upgrades were made in the planning applications, and how did they relate to the emergence of Generative AI? add

What is the significance of the Blue Yonder Network in the development of cognitive applications for supply chain management? add

What factors influence how organizations in the automotive sector distinguish their operational strategies from those of their competitors? add

bolt Powered by CUBE AI

Duncan Angove, Blue Yonder

search

Dave Vellante

>> Hi, everyone, welcome to this special CUBE conversation with Duncan Angove and George Gilbert. This is our debrief of Blue Yonder ICON, the company's customer conference. George and Duncan, great to see you guys again. You both were at the conference and I wasn't, so I'm really excited for this conversation.

Duncan Angove

>> Great, awesome to see you guys again.

Dave Vellante

>> So Duncan, let's set the stage. And George, I'd love for your independent analyst opinion as well, but ICON 2025 and the vision ahead. Give us Duncan, ICON in a nutshell. Coming off the conference, what were the headline moments, customer reactions, learnings that you would think best capture the experience?

Duncan Angove

>> Yeah, it was a phenomenal show. I mean, the analysts there, the partners there, the customers, even in the employee there said it was the best one they'd ever been to. We had our largest solution launch ever. Two years ago when we held this ICON event in Vegas, I announced in Shockwave and laid out that we were going to spend $2 billion, invest in product, and this was the moment to now go do it given the confluence of what was happening in the business world with supply chains and also the emergence of AI and all of that. Right? So, and we've been working on it for two years and this year it was sort of like, and just announce it, now we're actually shipping it. And by the way, and customers are adopting it, we now do six go lives a day around the world across industries and countries and basically unleashing massive value for all of these organizations. So, we had some major announcements. I mean, one of them was we announced our next generation cognitive solutions. These are all written on the Blue Yonder platform, they're built on the Blue Yonder network, and obviously powered by the Snowflakes AI data cloud and Azure. Right? And we can talk more about these later, but they are incredibly powerful next gen solutions, they're very AI forward in the use that they deliver, and they're more predictive and intelligent and quick and all of that. Right? So we announced our cognitive solutions. And then we announced our agents and we announced five of them and we can maybe get into more of them later. But we took a very pragmatic approach to how we built our agents. In a world where everyone's shipping agents and there's a lot of sort of called AI or agent washing. We've been very pragmatic, we've worked with customers, we've been very responsible in thinking about the role that generative AI plays in improving supply chains. Supply chains are absolutely critical as it relates to precision of decisioning, and generative AI obviously has had issue over the last two years of hallucinating and not being the most accurate. So, we've been very thoughtful about where we've applied it to specific problems in the supply chain to basically introduce machine speed and precision in the way supply chains run. And then we also launched our new brand and our new tagline, which is, Move Beyond Boundaries. And so super exciting, and the feedback we had from customers was just incredible.

Dave Vellante

>> George, as an independent analyst, what were your takeaways?

George Gilbert

>> If I were to distill it, it's what Blue Yonder seems to be doing is Amazon for the rest of us. Amazon redefined supply chain management so that a relatively small number of people could manage a supply chain with 400 million items and align their activities from forecasting demand a couple years out to what needs to be picked, packed, and shipped for an individual order. And Duncan, maybe you can elaborate on this, but what I saw was a big step toward the ability for companies to align not only their strategic assumptions that look out several years, but starting to be able to realign their activities, their sort of tactical activities, so that everything is in service of one objective. You can unify the measures for different departments or functions and different time scales, all so that they roll up towards one goal. Maybe you can elaborate on how you're moving towards that.

Duncan Angove

>> Yeah, actually it's a great observation, George. And in a world where everyone's talking about AI, we kind of lose this idea that supply chains and the software that supports them should be more joined up. It's one of the only enterprise software categories that is still very fragmented. And the way the market buys is with point solutions, they're just looking at TMS or they're just looking at planning or just WMS. They don't think about the end-to-end. And at the end of the day, it's a supply chain and that drives a lot of why you think you're optimizing one thing, you're actually sub-optimizing the end-to-end supply chain. And by the way, this point solution architecture with all its data silos, is the downside of it is amplified by local incentives, local KPIs, lack of cross-functional collaboration inside organizations, let alone across multiple partners in a supply chain and multiple tiers in a supply chain. And it's something we talked about three years ago as the central thesis around our investment, was that these things should be designed to work together to enable cross-functional and cross-enterprise orchestration and optimization and be built on a common platform and a common data model. And that's what we talked about three years ago. And I knew back then, and I talked to both of you about this, about the massive change management exercise ahead of us in terms of educating the market and extolling the benefits of actually having stuff that enables this cross-functional orchestration. And ironically, to your point, George, it was the big takeaway in the midst of all this announcements about cognitive and agents and AI, that was the big kind of a-ha moment for customers. Wow, I should unifying my decisioning, I should be working to enable cross-functional and cross-enterprise collaboration, and I should be thinking end to end. And that was sort of the big a-ha around companies thinking about, how do I take my on average 28 different supply chain point solutions and unify them on one architecture so I get to a global optimization versus local optimizations?

Dave Vellante

>> So you guys are doing some of the hard work, Duncan, we've dug into your engineering team. A lot of people, we see a lot of agent washing, we see a lot claims of agentic OSs. Within the supply chain, however, you guys actually can I think, take designs on that. So I want to go to the hard stuff. You've got multiple domain agents going live. You've got agents that can respond to changes and things like public policy, like tariff agents, despite this morning's announcement. How do you make sure that there's not automated finger-pointing in effect, in global logistics? Like when something goes wrong, how do you know where accountability lies? How are you helping solve that problem?

Duncan Angove

>> Well, certainly, I mean, some of it is about being end-to-end and having that end-to-end flow instrumented. It's a lot harder when these are all point solutions and they're not designed to work together to enable this orchestration. So that makes it really, really hard. I think from an agentic perspective, having these agents, making sure these agents don't just reinforce the silos is a really, really critical thing. Because what you don't want to build as an agent that just optimizes this one thing and all it did was replace a human with an agentic silo. Right? So we made sure we built our agents so they actually look across the supply chain. I read a really interesting paper a couple of weeks ago and it was talking about the benefits of human teamwork, and that when humans work as teams, they end up with better kind of outcomes, better decisions. Where it was the genesis of why people did design thinking, this idea that you can bring different people from different backgrounds and you get this sort of eclectic decision-making and people from different domains and knowledge sets and you actually end up with a better answer. Right? And one of the interesting things about agents is they can actually bring the benefits of teamwork and multidisciplinary thinking for decision-making. So imagine if you are a warehouse operator. As you're thinking about decisions you're making in the warehouse, you can now have a transport ops agent, a store shelf ops agent, an inventory ops agent, as part of that decisioning process. And what they're bringing to the table is that kind of benefit of collaboration and saying, "Hey listen, if we expedite those orders and we don't do this with this pallet, it's actually going to drive more transportation costs or it's going to drive more labor cost," or something like that and actually, you agentically deliver the benefits of human collaboration, which I thought was a really powerful idea.

Dave Vellante

>> What's your vision for how these agents work with humans? I saw the CEO of Anthropic the other day, he said some outrageous stat, like 50% of white-collar jobs are going to go away within the next two to five years. What's your thinking as to how agents and humans work together?

Duncan Angove

>> I think when, and there's two ways to answer that, Dave. I think when new technology comes along, the first thing we always do is we kind of do what we've always done a little bit better. Right? So the example I always give is that when the movie motion camera was invented, we filmed theater with it, right? And we're probably at the filming theater stage right now. I think it's a very lazy analogy when we talk about agents as sort of replacements for humans or some people have classified them as digital workers and all of that. Right?I think it's a lazy, lazy thing to think about it. Agentic technology is going to have a much more profound impact than that. An example I'll give you is in the car, the modern car of today, even the combustion engine, you don't have a choke anymore, right? So it's not about modernizing the choke and turning it into some kind of sexy touch button, you know what I mean? You just don't need it, the technology has automated that away. So we're obviously going to see a shift in the type of work that gets done because it's going to be made agentic. So there will be, and when you look at the work that workers do, it's going to be elevated and be more strategic because the mundane repetitive stuff is going to be automated for you. Right? And it's not just one-for-one, you're going to get 100 agents for every individual if you like it. So we built something called the shelf ops agent, and it basically deals with using natural language in a large language model. We actually built a fine-tuned supply chain model called the large Planogram model, it's informed by 30 years of intellectual property and hundreds of millions of Planograms that we've built. And it basically automates the editing of a Planogram a shelf. Right? So you want to take this item out, add this item. You want to change the assortment to appeal to a different demographic, or there's a promotion or allocating more shelf space to that category and it manages all of this. And today, retailers don't change these because it's incredibly laborious. Loads and loads of tabs and right-click menus and it just takes a long, long time to change a Planogram, so none of them have actually, they don't respond to changes and opportunities quickly as they should. And they end up with Planograms that aren't tailored to individual stores because managing at store level again would just be too manual and onerous. So we built an agent that basically automates all of that. So that space planner now, instead of doing all this mundane work around trying to figure out how to add this new item, is now spending time thinking about, how do I strategically grow my category? Should I be accepting this trade promotion offer or this new product introduction from the CPG brand? And thinking through things like that around how you strategically grow the category. So when you think about what that user interface looks like from a software perspective, you need the choke anymore. You don't need all the functionality related to laying out the shelf and doing all of that. The UI you need is more around data analytics and having a data analyst that lives in your data cloud that can help answer complex questions. I mean, it's more strategic, the user interface changes as does the work.

Dave Vellante

>> And George, you and I have talked about, just referred to what I call paving the cow path. We're trying to avoid paving the cow path. And George, in many respects, that's what we've done with software for a number of years.

George Gilbert

>> Yeah. Let me pick up on something you mentioned, Duncan, about how difficult it was to redo this, like the Planograms in the past. And you talked at ICON about this new notion of see, analyze, decide, act. I want to tie that back to the notion of adaptability and agility, which is in the past all the different local efficiency measures rolled up into something like return on invested capital, but there's this new measure or this new metric that you talked about, this see, analyze, decide, act, that is about realigning activities to changed assumptions. And is there a new metric that all these activities roll up to? How do you measure that so that, in other words, this notion of agility is something that you organize around instead of a notion of efficiency?

Duncan Angove

>> Yeah, George, gosh, I mean, we could... As you know, I studied economics at school, right? So I mean, this is something that we could spend a lot of time talking about. But yeah, we introduced the SADA loop, and I think I said at the conference, it was based on Colonel Boyd's kind of OODA loop from the military, about trying to figure out why certain pilots in the same planes won dogfights. And it was their ability to see, analyze, decide, and act before the enemy could. Right? And we took that kid of concept and applied it to supply chains, but we said we need to do it at machine speed and precision. And if you look at through the course of human history, we've always introduced technology which makes things go dramatically faster, but not just faster, but fundamentally changes everything. Right? And it's the same thing for precision and how do you take SADA and use agentic technology and everything else that we've built to basically make this thing dramatically faster? And it's your point on the metrics are interesting, because some of the metrics we've got will make no sense anymore. And I'll give you, I live in Atlanta, I go to London all the time. That for me is an eight-hour flight. So it's not just about making an eight-hour flight, six hours. It's like, what if that eight-hour flight is five minutes? It's no longer about flying to London faster, it changes everything, right? So I think that's where we're headed, and it will have a profound impact on not just software but the world in general. And there's something called the Solow's Paradox, as you know, and the professor basically said, talking about IT, "Productivity is all around us except in the numbers." And there are debates around, did the productivity measures we have really capture the benefits of technology? For example, I can get digital photos on here straight away instead of going to a film shop, it didn't capture that. But I think AI, we're going to see profound impacts on productivity and life and the wiring of the world and all of that. And we're probably going to need a new set of metrics in terms of how we think about all of this.

George Gilbert

>> Do you have customers who are grasping the change in the competitive access around this new measure? I mean, this is sort of what Amazon made possible was that they could plan and replan at a two-week timescale, where other people might've been replanning at an annual cycle or a quarterly cycle. Are you seeing customers now grasping what's possible and beginning to organize around it?

Duncan Angove

>> There'll be a few examples, but in general, George, no. We're still early, right? I said to you guys earlier, this is the internet in '96, and we're putting catalogs on a webpage. Right? We didn't know Uber was possible yet. And so, no, I think where we're seeing the metric side of things is what I talked about earlier, is if you're a warehouse operator, you're worried about number of cases you've shipped, cost per case, labor, transportation, it's your freight spend. Stores are generally focused on availability, waste, those types of things. And all of those metrics drives local behavior but sometimes sub-optimized it. If you said that the cost between warehousing and transportation was an index of 100, 75 of that is freight spend. So, why aren't you making decisions in the warehouse that optimize freight spend versus labor? You could have metrics in the warehouse where it is centered by number of units shipped, so I'm actually not going to ship to small stores that just get each's, I'm just going to ship the stores that want pallets. So again, you can see this everywhere across the supply chain where local KPIs and incentives and how you compensate people drives suboptimal behavior, right? And it's exacerbated by data silos. Right? So we're seeing customers now look at metrics that span it. Let's look at cost to serve and the entire flow all the way to the shelf. Right? And some thought-leading CPG brands and grocers we're working with, we're actually working backwards from the shelf. Everything intersects with the shelf, it's where the supply chain meets customer intent. Right? So that's probably where we're going see the beginning of this and it ties back to this idea of thinking end-to-end, thinking about flow, thinking about all of that. But metrics around how you think about the velocity that the supply chain runs at, I mean, we're still working on that.

Dave Vellante

>> Well, so this may be early, but I'm going to ask anyway because I think you guys talked about this at ICON. I want to get to the business impact. You did a multi-billion dollar replatform bet, and my understanding is that cognitive solutions, you referenced that earlier, that's your suite of AI offerings that seem to be fairly advanced. And that allowed you to rewrite, I don't know, dozens of planning apps. So do you have any data from early pilots, Duncan, whether it's cycle time compression, inventory turns, margin lift? Any insights you can share with our audience?

Duncan Angove

>> Yeah, Dave, they're staggering, right? And we did, we invested $2 billion, and just in the planning space we took 28 different applications on different data silos from whether it's demand planning or it's supply planning or it's allocation, it's replenishment, it's category management. And we rewrote all of those with an AI first mindset on one platform and one data platform, the AI data. And that allowed us to do a bunch of really incredibly powerful things. First of all, they're called cognitive because they are. Our cognitive platform is agent first, right? So we are very, very fortunate is that we were rewriting applications at the same time Gen AI emerged. So our agents weren't a bolt-on after the fact, they were actually architected into the application. So you take the example of the Planogram one I talked about, there was like 50% of the functionality you don't need in the new release. Right? So instead of spending on things that humans used to have to do manually, now that functionality went into allowing you to run the business strategically and we automated it with Gen AI. So in some cases, the cognitive application, the agent is the app and they're very agent first, and they've got massive amounts of intelligence fused into them. Over the last 12 months, we did nine trillion supply chain intelligence operations in the data cloud. This is ML, it's all this, nine trillion. We do more than Google searches per day by orders of magnitude. Right? So we've always been running AI and intelligence at scale. And by the way, some of these things around digital intelligence were invented in supply chain. Right? Linear programming came out of logistics. So this is sort of the home of where this has always been and applying it to problems in the real world. So we've always been doing that at scale. Our cognitive apps take it to the next level by bringing kind of agentic user experiences into it, combined with ML and solvers and all of that. But the big thing is, is it unifies these decisions. So instead of making decisions all over the place, you're making one decision concurrently where you optimize all of the objectives and constraints and resources simultaneously, and it drives massive benefits. The customers that are deployed cognitive between business benefits and IT benefits have seen a 12X increase in ROI. And by the way, they're coming from a very mature Blue Yonder application already. Right? So, . I mean, just from an IT perspective, here's a really trivial one. So when you look at customers, on average they have 28 different supply chain point solutions that run their business. Okay? So think about all the batch jobs that go between those different data silos and they're always fighting their batch window because they have to get orders down to the DC or the plant at a certain time in the early hours of the morning, and they're always fighting to find time in their operating clock. Well, I was speaking to one customer, they have 25,000 batch jobs and all it's doing is moving and transforming data and all that latency and all that cost and all that complexity. Right? So when you unify everything on a common platform, you don't have any of that integration debt, it's gone. So just the simplicity, the speed. When one app updates something, every app sees it instantaneously. Right? So just from an IT perspective, the benefit of getting rid of all of that integration and all of that latency. And what ends up happening is you can find three to four hours in your batch window, which makes you more agile and more responsive. So that's just one example, not in the license sprawl you have and all the inefficiency from procurement perspective across 28 different apps. So, just that's a simple IT one and there's loads of ones on the business side around concurrent decisioning.

Dave Vellante

>> Yeah, which would probably dwarf the IT savings. That's an amazing stat you gave.

George Gilbert

>> Duncan, maybe you mentioned one thing, you talked about optimizing across functions within an organization. But you guys also announced the acquisition of, I think you're calling it the Blue Yonder Network. Talk about perhaps how that makes it possible to collaborate beyond the scope of a single enterprise. And does that interoperability also perhaps make it possible for companies that have still multiple vendors within one organization to start mapping their capabilities into this sort of one planning and execution umbrella?

Duncan Angove

>> Yeah, and I think, George, I know you love knowledge graphs, but I mean, some of that is about the knowledge graph, which maybe we can talk about in a moment. But certainly, all of our cognitive applications are built on the Blue Yonder Network. Right? And I've always said, if you were going to start from first principles and build supply chain apps, you would build them on a network. Right? The idea that every company you go to has a network of suppliers and carriers and those suppliers have their suppliers, and all of these process flows go across that entire network, right? The inventory flows and then the financial flows. I mean, everything is built on a network. So the software that we built to support supply chain should also be built on that network and understand it, and it should be instrumented in real time so when things happen in the network, we know about it instantly, and we're to concurrently solve for it immediately and then communicate with the other stakeholders and orchestrate not just things inside one company but across multiple companies. Right? So, that's what the Blue Yonder Network brings to us, and it's incredibly strategic, particularly in a world of sort of agents. I've always said agents can reason and understand and they can propose decisions, but at some point they actually have to act, they have to touch the physical world. We're in the supply chain business, we make and move stuff. And at some point you need physical, you need bridges from digital intelligence to the physical world, either to see something happening physically. A truck is late, a supplier is short, I'm out of stock, and you need to be able to act, right? Change the root for that truck, rebuild that load, find an alternate supplier. You need these bridges to the real world, which is where having an end-to-end solution that's instrumented all of our execution systems like TMS and planning, TMS and WMS and OMS, and then all the trading partners that we have on our network. Right? So it's a critical component of how you think about orchestrating all of this agentically.

Dave Vellante

>> We love long-form Duncan, when we get a great guest like you on, we try to keep you as long as possible. So we've got a few minutes left, George, why don't we get into the data and the cloud strategy and the knowledge graph piece of it.

George Gilbert

>> Maybe, yeah, Duncan, along those lines. Everyone's talking about agents, it's like the shiny new toy, but not as many people are talking about the plumbing and platforming that's necessary to make agents effective. Perhaps you can elaborate on, again, going back to the knowledge graph, explain why just having a bunch of connectors to legacy systems is not enough to make agents effective in seeing and acting across silos. Help us understand why so much work has to be done at the platform level before that becomes useful at the agent level.

Duncan Angove

>> No, it is a good question, George, and you've always been a believer in knowledge graphs, right? But this is the moment for them and this technology. There could be a thesis where someone says, "All I need is agents now, and this agent's going to reach across these different data silos and everything's going to be agentic," and that's never going to perform. And there are huge amounts of issues in semantic understanding across these different data silos. Right? And that's basically bringing all of these different data silos together into one unified data model like I talked about, and then putting a knowledge graph on top of it is actually the unlock for first of all insights, but most importantly, agents. Right? And what's really missing from today's data is the semantic layer, the understanding of what the data actually means. Right? It's the essential truths of how your business and supply chain operates, your entities, your dependencies, your rules. And then it records it in a way that systems and AI can access, understand, and act on. Right? It turns sort of not tacit knowledge into explicit intelligence. So the first thing it does is it dramatically speeds up analytics. Right? So in the example I gave at ICON, if you remember was, say someone has a super complex question about what customer orders are impacted by something from a certain factory in a country that had an earthquake and what does it mean? And in the normal world, that would take a couple of analysts two to three days to actually figure it out. Okay? In a world where you have a knowledge graph, it does graph differential and it answers that question in five to 15 seconds, because the power of that knowledge graph and connecting all of this data together with this understanding, right? But more importantly, it also enables faster AI modeling and it unlocks all of this data and this logic for AI agents, right? Basically, it turns all these disconnected facts into kind of a connected brain and AI agents leverage the knowledge graph to basically think, adapt, and act on their own. But step one is unify the data. Step two is put a knowledge graph on it that basically allows the AI to understand it and act on it. And then step three is what I talked about earlier, which enables the SADA loop, is you need to connect it to the APIs that run your business, whether it's a solver, which is always going to be more compute efficient than having gen AI code Python for you, or it connects to a physical API to change a pit, change a truck or something like that. Right? That's the plumbing that you need to make these agents work.

Dave Vellante

>> Let's talk about your competitive edge. It's a crowded market, I'm curious as to what is kind of un-copyable. You've got incumbents like SAP, Oracle's got its SCM, others claim end-to-end orchestration. What, Duncan, would you say are the capabilities that Blue Yonder has that you can deliver that others can't replicate without kind of ripping and replacing their architecture?

Duncan Angove

>> There's always been a best of breed market, and the value proposition of ERP, which is, it's just good enough but it's all integrated, doesn't work here. Right? It never has because again, it comes back to that $2 billion in benefit on there, and that's why you don't see 3PLs using ERPs to run their supply chains. They operate in two to 4% margins and they have to be world-class at what they do. The benefit we give companies is you get the benefits of all of that world-class leading functionality, by the way, protected by 535 patents on file or granted, but it's all integrated and designed to work together and it's unified on a common data model. So that's what you get. We're very fortunate in this market is that we have a very, very complete portfolio. In the planning side, we either compete with best of breed planning vendors, but they don't have execution or we compete with companies on the WMS side that don't have planning. Right? So we bring it all together. And in this world where we're talking about agentic and speed, you have to have planning and execution connected, right? If they're disconnected, you're just not going to be moving with machine speed and precision. So, that's really our competitive edge. And I would say we're an authentically scale company in AI and have been, in a world where you said it earlier, everyone is agent washing, right? We do nine trillion operations a year in the data cloud with... I mean, we've been doing it for over a decade. Right? So real applied understanding and how you leverage predictive solvers and generative AI to solve real problems. That's sort of our edge.

Dave Vellante

>> I actually think your biggest competition is inertia, because you have functional silos and those are rewarded, you've got P&Ls and it's unintended. But your biggest challenge in my view is, how do you get customers to really rethink the incentive systems across the network and optimize the network? Because just optimizing one part of the network or one part of the organization is not the way to go, it's the whole house. That to me is your big opportunity.

Duncan Angove

>> I would agree, Dave, and that's what I said earlier. I mean, three years ago, I said this is our change management exercise. Right? The analysts come in and it's like we're just evaluating this. SI runs a process and it's just this, and that in an age agentic world is just going to create a ton of friction. You're not going to move quickly enough, you're not going to be precise enough. So companies that aren't thinking more joined up to enable a new age agentic world, I just think are really going to struggle.

Dave Vellante

>> Yeah, I think, George, we got time for one more question. I'll give you the last shot here at Duncan.

George Gilbert

>> What, maybe when we look out or when we look back in a couple years, will have differentiated your customers? How will they be competing more effectively relative to those who are still stuck on the silos? That seems to me the most fundamental transformation. This whole replatforming allows a new form of, a new competitive sort of differentiation for your customers. How will they measure themselves differently from their more siloed competitors?

Duncan Angove

>> I think our organizations that, and by they way I said this to you guys before, but automotive as a vertical has always been ahead of a lot of other verticals that adopt stuff. They were the first kind of guys to implement automation on the factory floor. Right? They were the first to basically implement just in time and then quality. And interestingly enough, they're the first to actually think end-to-end. They're all adopting network, they're all adopting end-to-end from planning to WMS to TMS to OMS to factory. That's exactly how they're thinking because they know it has to be, they have to operate at machine speed and precision, particularly in a world where you've got massive changes in digital preference from consumers and all of the instability that's impacting them from a supply chain perspective. So they just fundamentally know they have to build an architecture that's all about machine speed and precision. And we're going to see that thinking particularly, that was my most encouraging thing coming out of ICON. That thinking is now permeating other verticals where people are stepping back and saying, "Hold on a second. What we used to do in the past isn't going to be good enough anymore." Right? Some of them even think they can take agents and use that as bandages on all of these point solutions, that's not going to work either. Right? You need to step back and fundamentally think about things end-to-end, which includes your trading partners. And then, how does a agentic work sit on top of that, right? Versus again, I come back to just saying, "Hey, we can use an agent that can do what this person used to do." It's lazy and it's not going to be enough.

Dave Vellante

>> Duncan, I know you got to go and I want to thank you for your time. I just wanted to remind the audience, we have been running now with Blue Yonder for several months, the AI and the autonomous supply chain series. So you go to thecube.net and see that. Duncan, we had you at the New York Stock Exchange during NRF, we've dug into your product, we've dug into your AI and gen AI expertise, and really appreciate your time and collaboration. So, best of luck to you.

Duncan Angove

>> Brilliant. Great to see you, Dave, and you, George.

Dave Vellante

>> All right.

George Gilbert

>> Thanks, Duncan.

Dave Vellante

>> This is Dave Vellante for George Gilbert, and we will see you next time on theCUBE. Thank you for watching.