How to Make AI Work in Your Planning Organization

AI comes to the supply chain planning organization. What impact is it having today? I'm joined by Lori Harner. She is vice president of product marketing with SAP. Hi, Lori. Hi. How are you today, Bob? — I'm great. Thank you so much for joining me. Let's talk about the state of AI in planning today. Things have happened so fast in such a short amount of time. I'm wondering what impact it's actually having on the planning function right now. Yeah, so if we take a look back in planning, we've always had AI in planning for many, many years, right? So, starting out with embedded AI in our applications, but moving now more towards the Gen AI and then to agents, of course, will be the quick time to market moving forward, right? Okay, well, let's back up and cover each of the embedded AI. Maybe it wasn't even being called AI back in the day, but it had some role in this in analytics, not necessarily decision-making, but certainly at that point. Then you get into generative AI, which I think a lot of us now understand what that is about, but agentic AI, AI agents, what are they doing and how are they changing the picture? So, I really think with agents, it's more about having to augment it goes beyond augmenting tasks, right? Which we're seeing in Gen AI. It goes on to actually functions and the doing of helping through the planning process. So, we're seeing a lot of agents that are cross-organizational agents. So, basically bringing that end-to-end supply chain planning throughout the organization. So, I think with agents, you're going to see a great, you know, revelation in the way planners interact with one another and interact across the organization using both human and agentic AI. When we talk about agents these days in areas like customer service and the like, very often the word that comes up is autonomous. They're doing it themselves. Does that also hold true in supply chain planning? Are they stepping forward in that way? Yes, absolutely. So, we would go beyond again the Gen AI tasks, just task automation, things like that we're using and agents again will be very well used and utilized within planning and the operations across the end-to-end supply chain. So, AI is not just a shiny object. It's something that can be of value, but it can also be something that doesn't deliver value if you don't do it the right way. What needs to happen for AI to be in place and actually have an impact on the planning organization? Yeah, so I boil it down to three main areas, right? AI alone will not be enough. So, you basically need the data, the applications, and then AI on top of that to be able to orchestrate any type of end-to-end process using AI, right? So, AI alone again not alone is not good enough. But when you combine it with the data and the applications, that's really when the power powerful workflows can begin and the powerhouse around it. — Well, in theory, companies have always had data. Maybe they haven't had enough of it or maybe it's been wrong or bad or good, I don't know. But what are we talking about when you say they must have the data? I believe it's getting the data from cross applications, normalizing that data, and then being having that data being able to be used within the applications themselves. So, it's more about not just bringing in plain data or garbage data, if you will, but actually having the data that you can use within the applications through that cleansing within the data warehouse. So, what do the applications do? Is that the analytics part? What what do we talking about when we use the word applications? — applications really are around whether it be your planning application, your transportation application, warehouse application, utilizing that data to make informed decisions. Mhm. Who is making those informed decisions? We just talked about autonomous agentic AI. Is that in fact the AI model making them or is there still a person There's absolutely still going to be a person involved making the helping make those decisions and actually helping to orchestrate those decisions across the supply chain. So, the agents are doing a lot of the heavy lifting with the humans making more of the insight-sightful decisions. It feels like every day more and more gets put over on the AI and less on the human, but not to an end point where there's no human at all. Correct. There there'll always be a human interaction. It just might be that those humans are doing something different than they're doing today. So, the planner of today doesn't really necessarily look like the planner of the future. Right. I imagine the AI in a in one possibility in terms of recommendations type state, it can give you several scenarios that you could then proceed on and then the human could choose from among them. Correct. Correct. — But then you go beyond that and it doesn't even ask the human to choose, it just chooses one. Yeah, and in certain functions and different types of workflows, it'll be up to the organization to make that decision what they want to put completely autonomized and still have some type of human touch around, right? Right. So, AI fails. I mean, we hear a lot of disappointing case studies of AI

never getting out of beta projects. Why is that? What are they not I know they're not doing these three things you say they need to do, but where are companies failing when they improperly employ AI in their organization? I think first you have to set up the stage, right? understanding of what processes you want to use AI around, right? So, it's what are we going to take and put AI around to make more informed decisions, for instance. So, it's I've always said with AI with companies, it has to be a journey that they're on and that they're willing to take. I think it's also around basically deciding, okay, what are these things that we want to do where we're going to put AI around, right? So. And people always going to be in the picture to some degree? I believe for a majority of the tasks, like I said, if you look at planners of the future, they're going to be doing probably something a little bit different than they're doing today, but I believe there's always going to be this human and machine touch. So, it's still going to be a dual type of touch and I and again with the improvements and efficiency and productivity coming from the AI engines. Human plus machine, that's the equation you're proposing to us today. All this is I mean, again, to what degree is AI now, even as we speak, making its mark in these ways we describe versus aspirationally in the future? Yeah, so today, I mean, again, it's around putting processes around AI that are helping to, if you will, expedite productivity tasks, expedite certain types of workflows, expediting things like documentation. So, those thing the pieces of it that businesses have already said this is something we need AI to help with. So, if you think in your own example around even Chat GPT, like Chat GPT came out what, 2 3 years ago? And almost every person that I know within companies are using some form of Chat GPT. So, it's just the adoption is the other piece as well. So, you've got to have the organization willing to actually adopt AI as well to understand to get the strength out of it and the productivity and efficiencies. Yeah. Thanks so much, Lori, for those insights into the role that AI is actually playing in planning organizations today. May I take a moment, however, to ask you specifically about SAP. What role is SAP playing in this world? At SAP, too, we are leading by two different directions. So, AI first is one of the directions and business suite is the other direction that we have from as far as our North Star, our guiding principles, right? So, everything that we're doing, we're looking right now as we're building out the products, we're building out new functionality and capabilities, where we're putting AI within the applications, and we're talking to our customers and our partners and getting that type of feedback, real-time feedback. In some cases, we're even co-developing with some customers. So, it's really around taking that human approach again and understanding from our customers where the need is within their business versus trying to have us do it on our own, right? Cuz we can put AI in just about everything that we do. However, it's got to fit a business need. Yeah, thanks again, Lori. I really appreciate you spending time with me to talk about the role of AI in supply chain and specifically in planning. Thanks very much for your time. Thank you for having me. I've been speaking with Lori Harner of SAP. Thank you very much for watching.