Scott Jennings 0:00:11.9: 

Welcome. Is everyone here to talk a little bit of AI in the supply chain? Good, because that is going to be today's presentation. We're going to be talking about how AI is impacting supply chain planning from disruption to decision, and that's going to be the framework for what we talk about today. Was everyone in the keynote earlier today? Alright. So that was pretty awesome. All of that AI technology, all of that AI innovation is being activated in our product today, and today we're going to talk about some of the bigger trends that we're seeing in the space. Now from an agenda perspective, we're going to talk a little around AI and what we're seeing out in the marketplace generically, and then we're going to host an expert panel with some of our friends from AWS and Boston Consulting Group. I think we really have a lively show here. I think we're going to have some really interesting topics. We've actually done this before, so hold us to a high standard because we've had some practice and we think we have a pretty good show. So we're excited you're here. Thank you for coming and let's begin. 

 

Scott Jennings 0:01:13.4:  

So first, who has either played sports or had children that play sports? For years - I'm going to use baseball terminology, so if you're from the UK, I'm sorry, but I'm going to use baseball terminology, and all of us, if we played baseball, or our kids did, we went to a team picture. It was really awkward. You got together with your team. Usually the shortest kid had to hold a sign, and you'd take this awkward picture, and actually, my kids who are eight and six, that's exactly what they did last year. This year, when they showed up to picture day, they actually didn't see their teammates. They showed up and they stood on this board that you see here. On the left is Hunter. He's the biggest kid on the team. My oldest son. On the right is Logan. He's the smallest kid, so the book ends, and they didn't see their teammates. They stood on that panel, and they had a picture taken. Now, what was created was an AI picture. So they never saw their teammates we got it, we got a picture, and then their team picture was a stitched together point of view of everything you're seeing here.  

 

Scott Jennings 0:02:24.8 

Life has changed around AI and life has changed in a big way. Now, last year, I went to a conference called GS One Connect. It's a supply chain conference, and it was in Orlando, and I actually had my headshot taken. I think AI did a really good job, because my headshot really gave me a little bit of a youthful appearance, and I'm really pleased that AI was able to get that done for me. AI, sometimes it needs the human in the loop, but it is taking over just about every piece of our life, not just in business, but also in our personal life as well, and it is everywhere. That's one of the reasons we want to talk about how can you use AI to impact your lives and supply chain from a supply chain planning perspective. Now, the disruptions we always hear about in the supply chain are the Black Swan events. They're not going away. So of course, we talk about COVID, but every single year there's going to be some type of disruption, whether that's COVID or whether the Suez Canal has a blockage, whether there are storms that impact your local areas. I think [unclear name 0:03:31.2] and I earlier were talking about Hurricane Helene. I live in Tampa. We had people living with us for six months as a result of that storm. No doubt that impacted supply chains along the way as well.  

 

Scott Jennings 0:03:42.4 

It impacts shipments around tariffs. We hear all over the place, there's tariff roulette going on absolutely everywhere. Depending upon what your business does and who your suppliers are, tariffs are going to impact you in different ways, and of course, those impacts are going to impact different industries and strains of industries as well. These disruptions are what are in place today, but as we know, in the supply chain, disruptions are never going away. You're never going to get completely past the Black Swan. Nothing's ever going to be perfect, and you need to be able to respond to all that disruption in the marketplace. Now, when you think about supply chain and supply chain trends, the Association for Supply Chain Management has a very large membership across the world of supply chain professionals, and they come up with trends that they crowdsource every single year from all of their members. Some of the global trends that they typically report every year are their top ten trends, and of course, at the top of the list for the 2025 trends is artificial intelligence, followed by trade dynamics, big data, cyber security, and on and on, but from a supply chain professional perspective, AI is at the top of that list.  

 

Scott Jennings 0:04:50.4 

So I would imagine part of the reason that you're in the room today is that your organization, AI is bubbling up more and more and more. What I can tell you, working on the vendor side of the house, is probably the last ten meetings I've had with different large companies, we were there for a specific reason to talk about retail planning or supply chain planning, and at the end of the day, the meeting started off with, 'I know we're here to talk about this, but really what I want to start the meeting with is, what are you guys doing for AI, because our board has asked us specifically, what are you going to do to activate AI inside of our organization?' This is a really big deal inside of every company we're talking about. If you take that ASCM, or Association for Supply Chain Management top trends, they have their score model. They've changed it a couple of times, so I'll probably get the processes wrong. Classically, it's plan, source, make, deliver and return, but when you look at AI and where the penetration is and you look at that big green box, there really isn't a better place for AI than supply chain planning, and that to me, is the opportunity that is in front of us. So everything Adam talked about earlier today, that's the opportunity and supply chain is the place to go make that happen.  

 

Scott Jennings 0:06:05.9 

Now, it's not just me saying that. We have our friends from AWS who are looking at a wide range of use cases across their entire business, and you can see the world they're looking at, not just from the score metrics perspective, but also from their point of view on product development, sourcing, planning, manufacturing, logistics, merchandising, sales engagement and customer service, and that's just really in their consumer industries. You can see that AI is playing a big role from their point of view in demand forecasting, risk management, supply chain bottlenecks, emissions, CO2, and all things ESG. This is a really hot topic and when you drill into that, and building off Adam's presentation earlier, there's a lot of words on this slide, but really the three areas you typically hear about are around traditional predictive models for things like forecasting that have been in the market for many, many years. Generative AI, having a conversation with your data, the recommendations coming from large language models, and agentic AI, the ability to actually orchestrate decisions based upon those recommendations that either come from your predictions or come from those generative models.  

 

Scott Jennings 0:07:09.2 

That orchestration and that agentic piece is a massive opportunity, and that's a lot of what Adam talked about earlier today, and a lot of what our discussion and panel discussion is going to entail for the balance of this particular session, which, oh, by the way, I forgot to tell you, this is going to be the best session you're going to attend today. That much, we're committed to. So when we think about AI and scenario modelling, because a lot of planning platforms, you can run what-if scenario plans with, the repeatable pieces, the routine building blocks, we believe AI powered playbooks can handle those routine situations and make you significantly more efficient. We also believe in the complex world where you do want the human in the loop. That ability to have dynamic scenario planning is very valuable to large businesses to really understand, so you can see, okay, if we make these decisions, how does that ripple through the rest of the business? That's not just in your supply chain, that connects your entire business from your supply chain to your product planning office to finance, workforce, the whole gamut across the entire business, how do tariffs impact your business? Can you answer that question? What pieces and parts of your business are impacted by tariffs?  

 

Scott Jennings 0:08:18.9 

That what-if scenario modelling becomes very, very important, and it's not necessarily an agentic AI type of discussion. It really is, the human really needs to understand what are the impacts of decisions we're making. Now you saw it earlier, but I'm going to play it again, where we are headed at Anaplan from an agentic AI perspective is our supply chain analyst. So I'm going to replay this video because I think it's important for everyone to walk away, this supply chain analyst is going to be part of Anaplan's agentic AI strategy and we have analysts that are coming in, in all of our major COEs, from our workforce to finance to the sales piece of the business as well. 

 

Video plays 0:09:01.4 

Your supply chain faces constant change. You need AI that helps you stay in control in real time. Meet Anaplan Supply Chain Analyst, your AI agent for real time supply chain planning, trained on over a decade of planning best practices, on supply chain specific knowledge and experience with millions of planning models, because the best operational decisions require planning for what's next right now. Your AI agent is always on and ready to alert you to potential disruptions, flagging things like delays, cost fluctuations, and demand spikes that can impact your plans. In one simple chat, ask your Supply Chain Analyst to address high-risk scenarios, factoring in things like supplier availability, lead times, and cost impact. Your analyst will provide you with scenarios to see the impact across your entire network so you can leverage insights quickly and minimize disruption. Your supply chain analyst is a true partner no matter where you're working. It's always there to keep your plans on track. Decision-making has never been faster. What used to be slow and manual now happens instantly, so you can keep your supply chain moving at the speed of AI. Anaplan's Supply Chain Analyst built, trained, tested, delivering the next era of productivity today. 

 

Scott Jennings 0:10:22.5 

So that's pretty fantastic. These analysts that we're building are going to change the game from a planning point of view and Anaplan is making investments across the board relative to supply chain retail planning and a handful of other specific industries. As you may have heard yesterday, Anaplan announced an agreement to acquire a company called Syrup Tech. Syrup Tech is an advanced forecasting engine that is vertical specific, retail and CPG, and where it really gets into the nitty gritty is when you think about getting down to very granular levels of detail, that's what retail is all about. You're always going to hear about the SKU level. You're always going to hear about the granular level of information that needs to be part of your forecasting process, that needs to be part of your allocations process, that needs to be part of your replenishment and assortment process. Syrup Tech focuses specifically on that piece of the business, and it plugs into Anaplan's overall retail application suite strategy. This was a purpose-built acquisition for the retail market, and what you're going to see coming from Anaplan in the future is these are the types of investments we're making to make sure that we have purpose-built solutions that really move the needle based upon an industry use case. So we're really excited about this.  

 

Scott Jennings 0:11:39.3 

The acquisition is one day old, so of course I know very little, but nonetheless we do have customers that are using both of our technologies together, and there's a lot more announcements to come. So we're really excited about that, and that is going to lead us into our panel. So what I'd like to do now is I'd like to welcome to the stage, Abhijeet Shetty from Boston Consulting Group and Ivan Fernandes from AWS. 

 

Scott Jennings 0:12:10.6 

Please take a seat. 

 

Abhijeet Shetty 0:12:11.8:  

Thanks, Scott, and thank you for the invitation. We've done this a few times before, so I'm super excited because this is one topic you can guarantee, it's state of the art, changes so much. What I'm going to say today differs from what I've said three months before on the same stage. 

 

Ivan Fernandez 0:12:28.8: 

Thanks everyone. I agree this is going to be the best of all the conversations. 

 

Scott Jennings 0:12:34.5: 

If we could just get about 30, 40 more, I think we'll be all set. I do take a lot of positive feedback, and I like positive feedback, so just remember to leave some. So let's just do - I know you briefly introduced yourselves, so let's just do a brief introduction of who you are, what you do at your organizations and where your organization is focused. We'll start with you, Shetty. 

 

Abhijeet Shetty 0:12:52.2: 

Sure. I can start. Managing director and partner with BCG. I've been running operations for the last 20 years. I started work in AI before AI was a buzzword. About ten years ago, we used to call it big data and analytics. That then became a data science unit, and now it's a fully built out, product build unit called BCGX within BCG. I sit on the leadership team for both BCGX as well as operations. So I know a little bit about AI. 

 

Scott Jennings 0:13:22.2: 

Ivan. 

 

Ivan Fernandez 0:13:24.4: 

Ivan Fernandez, I've been in supply chain for 25-ish years. So again, we've seen it since the early days of what we had before any of these capabilities were available to us. My current role is worldwide leader for AWS, Amazon, for supply chain, and we see it in a specialized team around AI solutions. So our team is driving, how can we enable partners like Anaplan, partners like BCG to further innovate with us on behalf of the customer's needs, on everything gen AI and agentics, as we're going to talk about today. 

 

Scott Jennings 0:14:11.1: 

That's great, and I did not introduce myself at the beginning for the very purpose of being able to introduce myself now. This is my real headshot. My name is Scott Jennings. I am on the domain expertise team at Anaplan, specifically working with retail companies but I float across the board across a number of different supply chain pieces and parts. Anaplan has made a lot of really big investments, not just in technology, but also in people to get closer to the businesses that we're working with out in the field. Today, we're going to have a little moderator panel here. So it's going to be a little bit rapid fire and Shetty and Ivan are going to respond based upon their respective fields and their respective businesses. If you have any questions along the way, we're happy to field them. We're here for you. We want to make sure you walk away and feel like this was a good investment of your time. Ready to roll? Alright.  

 

Abhijeet Shetty 0:14:57.6: 

Let's do it. 

 

Scott Jennings 0:14:58.4: 

Let's do it. Question number one, and we'll start with you, Ivan. What are the biggest AI trends in supply chain management that you see across the market in AWS? 

 

Ivan Fernandez 0:15:10.0: 

Yes. So currently, and for some of you that may be AWS customers or just following the trends, this year, we've doubled down on everything agent, releasing what we call Agent Core, but basically what we see is the start - and I will steal some of Adam's words earlier. We're at the beginning of the beginning of this new wave of leveraging gen AI into autonomy, into agents, and for that, what we're seeing is that shift from I would say more the need to be an interaction from the user, to start the conversation with a gen AI into a more proactive way to building these autonomous recommendation alerting frameworks. So that shift is one thing, but also the other, and I would say highlighting Anaplan's capabilities for analyzing large data sets of data in near real time is to layer these gen AI agentic frameworks on top of it to surface proactively much more external variables. So what we've been talking in the industry around demand sensing capabilities, and it's been hard, now we have the foundation to really enable real time proactive alerting of anomalies, exceptions around the supply chain for any vertical retail CPG and others. So those are the trends that we are actively investing, looking to innovate along with the customers to enable these frameworks for the customers. 

 

Scott Jennings 0:17:00.1: 

That's great. Shetty, what are you seeing at BCG? 

 

Abhijeet Shetty 0:17:02.5: 

Yes. So I would say three big things. The first is obviously the incredible excitement around agentic and agentic AI. This is more of curiosity of what agentic AI can do for an enterprise and supply chain leaders are consistently asking about this. That's not surprising. What is interesting is the second, which is a maturation of AI sophistication with our clients and BCG in the supply chain space, has obviously a five-level sophistication maturity level framework and we see clients all the way at level five, who are really putting agents in pilots and testing them and extracting value in pilots. Then we are seeing clients at level one who are somewhat pre-APS but are trying to leapfrog using AI up to level four or even level three. So that sophistication in the conversation and then the conversation maturing into how do I get value from AI has become a very big theme that we are trying to answer as a strategic consulting firm. The third is a fairly interesting trend that is starting, and I think people are again trying to leapfrog to level four, level five, which is physical AI and embodied AI. Obviously, gen AI works on tokenization of language. That concept has now been brought forward to the tokenization of physical movement, and so there are a number of humanoid robotic companies out there who are trying to get into supply chain, and their thesis is very simple. Our world is a world built by humans for humans. So you can never fully automate supply chain using infrastructure that is physical robots. So you have to introduce humanoid robots, and they are starting to use AI as a massive multiplier to start to train humanoids to go into warehouses, to go into retail stores as labor and to even go into factories. So interesting trend. I think that's four years away, but that's coming. 

 

Scott Jennings 0:19:08.7: 

One quick follow up for you, Shetty. Who's calling you? Where are you getting calls from? 

 

Abhijeet Shetty 0:19:15.8: 

Right from the CEO to the CFO and of course, the COO. So this is no longer just a COO topic and just a supply chain officer topic. It is a core and center CEO topic, and CIOs are way in over their heads in terms of understanding AI, crafting a roadmap, thinking through their architecture and tech stack, and then servicing all these tasks that are coming in from the functions. 

 

Scott Jennings 0:19:45.5: 

Let's move on to question two. Question two: Most supply chain organizations have experimented with AI, agentic AI. What allows an organization to scale and prove value? Go with you, Ivan. 

 

Ivan Fernandez 0:19:59.6: 

Yes. So back to where we're heavily investing as AWS is to create the framework. We were talking prior to this session, is that the solution is not going to come out of a single agent, and we should not build an agent that is overly complex. Rather, we should task those agents with simpler tasks but allow them to collaborate, to create armies of agents that can audit, supervise, correct other agents. So again, those agents are not going to come from a single place. So again, what customers need is to be able to experiment in a framework where you can have the application layer being Anaplan, where you can have a consultant building agents, where you can build your own agents. So that's where we see the focus for being able to first ensure security, ensure scalability, and then I would say a consulting advice approach to allow for simple use cases to prove the thesis in order to then crawl, walk, run, and scale, which is where we also see a lot of failures in companies that have just tried to embrace AI for the purpose of saying that they are doing some AI, and then they shut them down without being able to scale into production. So that's what we're trying to partner with you guys in order to allow the customers to have those sandboxes where they can experiment and as we call it, fail fast or then go into broader use cases. 

 

Scott Jennings 0:21:44.5: 

That's interesting. So before Shetty answers, out of curiosity, who in the room at your organization today is building out agents? So we do have some folks that are building out agents so that's good. From previous sessions, the room was relatively - was very tentative about that. I saw a few hands stick up. So that really tracks a lot of what we're hearing in the marketplace. Shetty, I don't know if you want to double click on that. 

 

Abhijeet Shetty 0:22:12.4: 

Is there a slide behind this that talks about how to get value? 

 

Scott Jennings 0:22:15.7: 

It's like we've done this before! 

 

Abhijeet Shetty 0:22:18.8: 

So you no longer have to guess what drives value for a use case. The time for experimentation has passed and we are on the other end where there are enough pioneers who have experimented with use cases. So there is no reason to guess which use cases deliver value and also the pathway to value. So the first point that we would state is start with the business. Start with use cases that are proven, proven, proven to deliver value. Planning delivers value. Procurement delivers value. There are a bunch of other use cases that we can talk about. Pricing delivers value, but in the supply chain space, planning is a very clear value driver. Focus on planning. Go deep. Don't experiment. Don't try and create agents. If you are at level one. There's no need for sophistication. You're just trying to get to level three, level four. The second is construct your execution such that there is fast impact. Choose a tight small business unit that is representative of the larger enterprise. Bound the scope of the pilot such that you know you're going to create value. Find a business leader who's actually the GM of the business. Do not do this in supply chain. Get a business leader to own the pilot, drive it, and think about the dollars that that pilot will unlock, and then start investing and doubling down on the pilot with clear milestones that drive impact for you.  

 

Abhijeet Shetty 0:23:45.1: 

The last is be aware of both change as a barrier as well as data as a barrier. Every single client organization that I've walked into has got data available, not data usable, and the fact that you've got a data lake or a data fabric does not mean it is at the right level of granularity that can translate to value from a use case, and then humans will not trust a black box model. They will not trust an agent. So the question of autonomy comes much later. First, build trust with the model and build the ability for humans to say, that's a good recommendation. I'm going to decide basis that. 

 

Scott Jennings 0:24:22.7: 

That's a great answer. I feel like it's a jeopardy. My last name is Jennings, so I feel like a Ken Jennings here. We'll go with question number three. You don't have to phrase it in a form of question, the answer. How do different forms of AI work together for supply chain planning excellence? We'll start with you, Ivan. 

 

Ivan Fernandez 0:24:40.1: 

Yes. So, Amazon, we've been testing or playing, experimenting for over two decades. So we started with the traditional AI and the ML, as Adam was saying earlier, the linear programming as a foundation, and then we went into a deep learning which was a precursor to gen AI, but with gen AI, what we saw, at least for supply chain use cases, was that it was marginal. It could analyze vast amounts of data, provide recommendations, summarize, but the real shift that we're seeing right now is with the agentic capabilities, where you can start, as we said, letting those agents shadow and provide recommendations. They need to learn, another thing that we say a lot and that I like Adam saying, that it's a very smart toddler. He's like, even the best gen AI or large language models, you should not give them the keys to the kingdom. It's like, they need to learn from the users and over time, those recommendations and that reinforcement is what's going to allow for autonomy, but we are still not there yet. So that's where we're seeing most of the innovations that we are trying to bring down to use cases, to experimentations with our customers. 

 

Scott Jennings 0:26:15.3: 

How about you, Shetty? What do you see? 

 

Abhijeet Shetty 0:26:17.5: 

We are all familiar with a tech stack, right? So you've got an infrastructure layer, you've got a data layer, you've got an app layer and then you've got a decision layer. You have to think about AI also as an AI stack. There are simple AI forecasting models, machine learning, that have to be part of your workflow, which then you layer on advanced stuff like simple gen AI call systems that then allow you to summarize information, input, etc., and then there are agents that are running around fully democratized AI, etc., etc. I think this is as much now a CEO and a COO decision as it is a CIO decision what your AI stack is going to look like, and not many of our clients are really thinking about this. So I think to the pointed question of how do different forms of AI work together, you have to construct the AI architecture right from the start, depending on the use case that you're driving. Planning is again phenomenal. You can use all kinds of AI forms for planning as a use case. There is regression and time series. That's not even AI, but then you've got machine learning, and then you've got gen AI, and then you've got full blown agents, and all of that can be part of your planning stack. 

 

Scott Jennings 0:27:33.8: 

That's good. Moving on to question four. This is one of my favorite questions around data because I love data. What role do you see data management playing in creating positive business outcomes for AI and supply chain? I was at a supply chain conference, I can't take credit for this, but I was at a supply chain conference and the speaker said, sometimes even with data, you run into a Minnesota situation and people are, of course, like, what's a Minnesota situation? And he said, well, it's the land of a thousand data lakes, and you run into that quite frequently where there is data everywhere and then all of a sudden you have this AI strategy that's looking for very specific data. So with that is the backdrop, Ivan, what do you see from a data management point of view coming into positive business outcomes for AI? 

 

Ivan Fernandez 0:28:18.9: 

Yes. I couldn't agree more and that's the reality of more customers. Companies have made huge investments in ERPs and supply chain solutions and HR solutions, and then they have done some home-grown, and as I always say, and in some part of the organization, they are still running with Excellence or Excel spreadsheets. So the data is all siloed. So there are two areas where we try to innovate, provide a foundation, and then we partner heavily with companies like Anaplan to try to bridge all these siloed data. That's the one thing. In some cases, Anaplan has a great foundation. In some other cases where a customer doesn't benefit from this, we need to provide the data lake or the data warehouse where you can have, I would say a holistic - we call it canonical model. So one of the early usages of AI was not to try to do the mapping, and I would say with BCG, in some of our deployments, it will take weeks or months to manually map all of the data that is dispersed around all of these applications. But now with AI, at least at AWS, we dump all of the data and we let AI run and come back with the recommendations of, we believe this makes sense, and you can drag and drop and say, well, this is custom for us, or yes, all of the heavy lifting that otherwise would have been manual, we do it in minutes.  

 

Ivan Fernandez 0:30:08.6: 

The other thing is to analyze the data before you start running decisions or planning or anything else. Again, Adam was very, very clear in the morning, the first rule of garbage in, garbage out. So don't just dump the data and try to solve, okay, the data is in a single data lake, and let's just throw the agent and let the agent run wild. No, you need to validate for accuracy. You need to audit on the way in into whatever you're going to run at an genetic or AI layer, but then also you need to have audits on the output of those recommendations. 

 

Scott Jennings 0:30:49.0: 

Yes. So if I was to summarize this, you really can't have AI without proper data management. I think, Shetty, it'd be interesting to hear, if you're talking to the CEO and the COO, do they feel the same way in your response? 

 

Abhijeet Shetty 0:31:01.5: 

Yes. It's interesting how CEOs and COOs are now talking about data. The second or third question that we get is, do you think you can make our data work for this use case, because I buy your value, I get that you've done this before, but we've stumbled so many times with our data that I don't believe that we will be able to do this. This is not the CTO or the CIO asking that question. This is the COO and the CEO. The reason is that data management frequently is the number one barrier to executing a well-crafted AI program. You cannot run a Ferrari without fuel. So you might have the best AI engine, without data feeding it you, you will fail. Our programs end up becoming as much about data governance being set up as they are about creating the AI forecasting model and without good data governance and data management. So you need a small data team that manages the data, but then you need domain owners in each of the functions, marketing, sales, guaranteeing that the data that is being produced by those functions and being fed into the data lake is accurate. You get Nielsen data that is garbage. You get IRI data that has got three months, four months missing, and then people are scratching their heads and some data engineer in the back is thinking, why is the forecast wrong? Well, the forecast is wrong because you fed it garbage data.  

 

Abhijeet Shetty 0:32:31.0: 

So somebody in the business should then own whether the data quality is good or not. Sign off on the data every three months, six months, and that is when you've got a very clean pipe running into your exceptional forecasting engine that then generates a forecast, does the plan for you, generates answers to your queries. I'm sure when you guys are rolling out Supply Chain Analyst, data is going to be of prime importance, because where will the query go to if there's no data underlying to answer the query? So this is, I would say, the number one problem and challenge and it is now becoming a C-suite discussion. Absolutely. 

 

Scott Jennings 0:33:09.9: 

Which is fascinating, and I think… Go ahead, Ivan. 

 

Ivan Fernandez 0:33:11.9: 

No, just to complement on that is the power of gen AI and agents to not only tap into the company's data, which is the first thing we need to tackle, but just to go into the wild and start tapping into external data sets. That's the first reason why those other companies start hallucinating, and so you need to provide guardrails and you need to provide audit agents in order to alert if there is a drift from where the data, the external feeds are accurate or reliable. So the problem becomes much bigger now that we have access to those other vast data sets that are like Nielsen, or that may not be accurate. 

 

Scott Jennings 0:34:07.9: 

When you think about planning, planning inevitably, especially for big companies, the data is growing because everyone wants to get more granular and they want to be more real time and as you add those two elements, you're going to explode the amount of data that you're working with. One of the things you'll hear from Anaplan, from a data management perspective, is that at all times, we have 7.9 petabytes of information in memory at all times. That's 15 zeros, and that number is only growing. That's an old stat. I'm sure it's even exploded from there. That's what we're seeing as we're working with big companies, that type of planning scale. You heard some of those data points that Adam dropped around the quintillions. These are massive numbers, but it's what you need to scale and succeed relative to AI and supply chain. So our question five is what do you see as the greatest opportunity for AI enabled supply chain planning? We'll start with you, Shetty. 

 

Abhijeet Shetty 0:34:59.7: 

I would say the greatest opportunity, and I would say it's more of the opportunity cost, not investing in planning is a lost opportunity. Like. we said from the start, we are all looking at AI. We are all looking at value from AI. We are all looking at use cases that deliver value. Planning is an absolute no brainer. So if I think about planning itself and the greatest opportunity, the greatest opportunity is to press the trigger right now, start investing. I'm on your stage. I'm not necessarily biased one way or the other but pick a platform and just go with a trusted provider, partner, platform that you trust and start investing. There is no greater time than now. Supply chain planning actually is proving to be a competitive advantage because you started your conversation by saying disruptions are a way of life. We are in a supply chain supercycle for the next five to ten years, where disruptions are going to be a part of our life journey in supply chain. Tariffs are being ladled on right now. They're being switched off. Three years down the line, administration changes. The tariffs framework might get dismantled. Supply chains are getting rerouted globally. Geopolitical shifts are causing manufacturing to move from far shore to near shore to maybe inshore. This is going to continue.  

 

Abhijeet Shetty 0:36:25.1: 

Planning is actually going to prove a competitive advantage because the faster you plan and erstwhile if you had a eight week planning cycle or a 12 week planning cycle, but now you've got a three week planning cycle or a two week planning cycle, and you do scenario planning on top, that gives you a significant advantage versus competition, where other players might be taking longer time. So your time to decision-making can get accelerated if you have really, really good level four, level five planning maturity and abilities, but again, the time to start is now. In fact, it was yesterday, but the second-best time is today. 

 

Scott Jennings 0:37:03.5: 

We actually were talking to a company not too long ago, and they mentioned that they had manufacturing in China, and they wanted to move it to Cambodia, much cheaper in Cambodia, but when they went to go do it, they realized the port itself is not as deep, so they had to bring in more boats. Those are the types of things that we're hearing companies talk about regularly, these costs they didn't even know were costs popping up and then impacting their strategy. But on that end, Ivan, what do you see as the greatest opportunity for AI enabled supply chain planning? 

 

Ivan Fernandez 0:37:33.2: 

I would say on top of all the capabilities that you can empower with these newer innovations, I would agree that we needed to start yesterday. Now, the good news is that this is a starting point for your companies and all of your competitors. The point is, you need to start experimenting right now. There are proven use cases, like planning, where gen AI can have a dramatic improvement, but it's only through experimentation that you are going to embrace it, and that's going to come along with the change management of how it's not only agents or gen AI is going to take your jobs, but humans need to leverage. It needs to learn to have armies of agents working for them, and then through that gradual transition from the agents shadowing, learning, having positive or negative reinforcement through their recommendations to the point where you can start shifting into autonomy. That is also some something that you need to take into account, that from the pre-trained large language models and whatever the task or the prompt or the goal that you empower a specific agent, they need to learn from your business and that data needs to stay within your instance with all the confidentiality, with all the guardrails.  

 

Ivan Fernandez 0:39:09.5: 

So it's only through - and I would say not trying to go big bang but also trying to run multiple experiments because these can apply to planning. These can apply to HR. These can apply to customer service. So the sooner you start learning how to embrace these capabilities, the better that you will be against your peers in the next, I would say two to three years as also we software vendors, hyperscalers continue innovating. So definitely we need to start. 

 

Scott Jennings 0:39:48.5: 

Both of our panelists for all of their insights from AWS as well as BCG, and I'd like to thank the audience for coming. I hope you enjoyed today's show, and please come by and see us, and remember in the supply chain relative to AI, Anaplan is an outstanding solution, so please keep us in mind.