Supply Ch(AI)n Strategy Session - AI Bootcamp FROM FASTCO WORKS AND SAP
19:43 Share

Welcome and thanks for tuning in to this session about how AI is completely revamping supply chain and procurement strategies. I'm Julianne Pepitone and I'm happy to be your moderator for this event, AI Bootcamp, a Fast Company & Inc. series in partnership with SAP.

As any procurement leader knows, when it comes to the supply chain, the old adage is true. The only constant is change. Climate events, geopolitical, technological changes, global pandemic, as we've seen recently. The thing is that the supply chain can change in an instant and that we're

requires resilience. And that's where AI comes in to embed that resilience into your strategy, to automatically identify and address supply chain issues, and to find sustainability opportunities that reduce carbon footprint and lower costs. Today, we'll talk key tactics and takeaways for integrating AI into your supply chains.

Joining me today are Dr. Walter Sun, SVP and Global Head of AI at SAP, and Dr. Ava Ponce, Director of Omnichannel Supply Chain Lab at MIT. Thanks to you both for being here. You know, as you both know, for many companies, procurement really is the lifeblood.

So I'd love to start here. How can AI modernize procurement and help mitigate supply chain risks while informing procurement strategy for the long haul? Walter, we'll start with you. I think the idea of disruption mitigation is a big key technology that AI can enable. And the idea is that if you're a supply chain planner, you have so many different variables to deal with every day and you can't really keep up with everything. So if there's an algorithm that can track news, events,

different things happening around the world and alerting you with issues that you would make your life a lot easier. On top of that, with generative AI, you can actually marry the idea of alerts with the idea of finding mitigations. And the mitigations could be, you know, if you see something, let's say a disruption in a Suez Canal, you can have the machine algorithm look at your deliveries and say, how many of your deliveries are rotting through a ship, for instance, through the Suez Canal?

and notify you as well. So that's the next level. And then the final level of that is on top of that, you can actually even have generative AI create content, send emails, pair emails to send to the individuals that you, your vendors, people you're supplying with and say, hey, do you want to send these emails to them and ask them if your delays are happening due to this issue? And so that way, all of this becomes like almost like you have a human assistant sitting next to you who can alert you of things that you otherwise wouldn't have known about.

Right. Wonderful. Eva? Yes. So as Walter mentioned, AI and generative AI tools are disrupting the landscape. And I think this is true in education and the industry. Companies are exploring how to use AI to offer more customized products to their customers, but also to use AI to improve their operations. And artificial intelligence has the potential to enable signal

efficiencies across the entire supply chain and the logistic industry. And more specifically, I think has the transformative potential to enhance procurement strategies and supply chain resilience, specifically in the face of dynamic and global changes.

Some examples of how to do that, predictive analytics is a big, big area that is impacting many areas in the supply chain, and more specifically procurement, supplier risk management, I would say another big area, and supply chain visibility. Because this end-to-end visibility is something that the supply chains require currently, AI can help with this supply chain visibility, and specifically if we look upstream the supply chain to their suppliers.

Yes, I mean, you each mentioned several interesting examples there. And I'd love to dig in a little bit more on some of these use cases and just talk to people about what that really looks like on a practical level, you know, either at your own companies or ones that you've encountered in your work, you know, for AI and the supply chain and procurement space.

So, Eva, if you could pick one or two of what you just talked about or maybe a new one, if you could walk us through a use case, an example of what this really looks like in an organization. Indeed, I'm going to pick some of the top number ones. I currently run a survey as part of my Omnichannel lab to better understand the impact of AI specifically because of the growth of e-commerce and Omnichannel. And we asked more than 100 retailers and manufacturers specifically.

Where do they see AI having the greatest impact on supply chains? So I would say that demand forecasting is the top number one. And the reason is that AI analyzes historical data, but also external factors like market trends.

and weather, for example, to predict a future product demand. And this allows, at the end of the day, for a more accurate stock levels, allows to reduce waste, to reduce shortages. So an accurate forecast is something that has been an obsession for supply chain professionals, I think.

So definitely machine learning techniques has been used extensively in order to bring more data to help and improve this forecast. But there is much more than we can do here. And AI, I think it can also start bringing kind of judgment components into this forecasting. And I think this is kind of the more, I would say, revolutionary part of using AI specifically in demand forecasting.

As part of that survey, I also identified customer experience. And specifically, we look at the customer, what happened with AI. They come being also more personalization. So, and this topic recently appears in news headlines. For example, Walmart is launching a new generative AI search service.

feature that will allow customers to search for products by use cases instead of by the specific product. So it's another way to use AI. Also related to customer service are also the chatbots and these AI shopping assistants that are also helping a lot customers when they are shopping and personalize some of the products.

And of course, inventory management. I think this is another alien supply chain that AI is bringing new insights and new addition. And here, same thing, machine learning has been using so far, has been helping to simplify the process, but AI will transform the function. Again, companies are trying to introduce AI just to incorporate in their replenishment solutions.

Just trying to automatize that based on the customer needs. Yeah, I'm struck by so many of those trends that you mentioned. I guess it's a bit of a chicken or egg, but there are trends we hear in the retail space too, right? That people consistently want more of a personalized experience or a personalized product even. So, you know, and that's something that AI is able to give us. So do we want it because AI is able to give it or is AI kind of evolving to give us what we wanted? Yeah.

Exactly. Yes. And it's more kind of based on your previous orders, what you just purchased previously, historical data, what is happening currently. So it's a mix of combination of historical data and current events that are happening that is personalizing this offer just for customers.

And Walter, I'd love to hear some use cases or examples from you either at SAP or some of your clients as it relates to AI in the supply chain space. If you could walk us through an example of what that looks like on a process and practical level. I can think of three examples. One starting with the demand forecasting that Ava was talking about is like

With AI, you can actually, that's not just using time series, which historically you just say, look, what's the pattern historically? Maybe there's seasonality on US Thanksgiving if you're shopping in America, but the history is a good predictor of the future. But now there's exogenous factors, right? Of course, with the pandemic, of course, with shortages in delivery, the Suez Canal blockage, there's different things that can happen. So the idea of having exogenous signals being added to the model. So AI can actually model what ifs, right? You can have a what if model saying, what if these eight things happen?

And you can model each of them happening. And as soon as one of them does happen, you can actually have a mitigation plan ready immediately. So I think that's one area that we're building up. The other thing that Ava was talking about, recommendations. I think you were also talking about it as well, Julianne, that in Ariba, which is our procurement tool, we can actually do recommendations for businesses.

And what the recommendation is, isn't that like, you know, consumers purchasing. It's actually saying, hey, if I buy a certain laptop, what do other people normally buy with that? And maybe there are certain docking stations or products that fit automatically. And it saves, you know, users a lot of time and saying, hey, look, people who typically buy this laptop buy these three other components. So that's a use case that we're actually offering our customers as well. And finally, there's a car company which has a pain point they told us about, manually documenting data. So we have this

transportation management use case where we expedite freight verification and documentation. And so, you know, in the manufacturing industry, a single manufacturer can have hundreds of trucks that arrive every day at production facilities. And there's nothing digital in their manufacturing system. They give you a receipt saying, hey, I dropped this

bag of widgets. Here's a piece of paper. And so now you can digitize that. You can actually process it and create a record keeping of it. And that information to this company who worked at CEDA stays on close to a million a year just for one pilot facility. So you can see there's a lot of value add and savings time and money for these type of technologies and supply chain.

Great examples. Yeah, and in each of those use cases and examples, you both mentioned some technologies that are used. I'm curious, are there additional AI technologies, either specific types of products or perhaps just categories that you think are having the greatest impact on the supply chain and procurement spaces right now? And do you see that evolving? Can you kind of predict a little bit where this technology is going? I think optical character recognition is a big area where you take an analog piece of

paper, OCR, and get the images. But I think the next level is what we call multimodal large models. So it's generated models, which not just take text, it takes images. And this can be useful in supply chain space of detecting issues with the processing. If you have a camera in, let's say, a warehouse, you can actually find issues. If there's a water leak, the camera can quickly detect, oh, the images see something here, and the models can detect aberrations, detect, alert the manager quickly and find issues. So it takes

unstructured information and images and create structure out of it and alerts as well. And I think that's kind of the next thing where you say, look, we've done text to generative models. Now we have this multimodal images and videos as well.

Ava, same question for you. Are there certain types of technologies or categories that you think are having the greatest impact in your field right now? And do you see that evolving in the near term? Yes. The example that Walter wrote about identifying kind of disruptions in warehouses is one of the top ones. And we are using a lot of applications specifically for warehouses.

More specifically for procurement, I would say natural language processing is something that in procurement is also helping to automate and streamline the processing, for example, of uninstructed text data, such as invoices or purchase orders and emails. So are helping to streamline operations.

Another area that is not kind of as impactful as the other, for example, predictive analytics is a top one. But robotic process automation are also being used to automate kind of a basic procurement task

such as order processing and payments, is again helping to streamline operations, is helping to free up human workers and having them more time for strategic works instead of just for more operational tasks.

So all of these things are helping. And of course, Internet of Things, these devices in the supply chain that are providing real-time possibility to tracking their goods, offer transparency and visibility across the end-to-end supply chain are really powerful and bringing this end-to-end visibility that you need. For example, in e-commerce and omni-channel, if you need to provide a seamless customer experience, you need to have inventory visibility.

So there are many tools, AI tools that can help to gain this end-to-end visibility that I would say the growth of e-commerce is bringing underneath of omni-channel and integration of different channels. So as you both enumerated, there are plenty of technologies, plenty of opportunities, a lot to consider for executives in this space who are developing their AI strategy.

I'd love to bring the office of the CIO and CTO into this for those executives thinking about AI, how to incorporate it into their strategy. What are some of the key questions that they should be asking of their CIO and CTO offices? Ava, we'll stick with you. Okay, so for executives developing AI strategies, I'm focusing more on chief information officers or chief technology officers. The first thing will be understand the assist situation. I mean,

What are the current state of their AI capabilities? And how ready is the IT infrastructure in order to support this AI deployment that they want to implement? The second key thing is data, data, data, and data. How will they manage and govern the data required for AI? Because we need to ensure the quality of this data

Security aspects are also key here, compliance with regulations. So data plays a key role, not only the quality of the data, but how we are going to manage this data.

Another thing I will bring here is the strategic alignment, because we need to align any AI implementation, any new strategy like AI that is going to be a disruptive technology with the business goals. Questions kind of how will AI implementation drive business value and competitive advantage in this specific industry for their specific operations?

I think will be key. And one of the most common reasons I have seen to fail when implementing disruptive technologies like AI is when they are rushing in that implementation with a lack of clear vision. So I think this alignment will be clear. And finally,

as an educator, I need to finish with the skills and talent. But at the end of the day, they need to incorporate the necessary talent in order to develop, deploy, and maintain AI solutions. Sometimes they need to maybe think about hiring new talent

or just upskilling their employees in this area. So I would say just to be ready and have the right skills for this journey is something key and can be through upskilling or through hiring new technicians, but definitely something that might need to be in consideration.

Thanks, Ava. Yeah, Walter, same question for you. For those executives thinking about how to develop their AI strategy, how should they think about the tech offices? What are some of the key questions they should ask of the C-suite tech leaders? I think the main question is, are field workers ready, the ones in the supply chain, ready to handle the technology?

And I think the answer is no, as Ava was saying, the upskilling. How do we educate users for the technology? I think also the people who are building from CTO's office, people who are building technology should also think about how we can make it easy to use. Meaning that instead of opening an application on a laptop

find out alerts or mitigations, you can actually hopefully get text alerts, right? Because everyone has a phone, most people have phones, they know how to operate them. So if a supply chain warehouse worker gets a text message saying there's an issue here, then he or she can quickly address that without needing to have to learn too many applications or products to make it work. So I think that it's a combination of

upskilling, you know, educating the field workers, but at the same time, building technology that's easy for field workers to use. So it's kind of both the CIO kind of educating and the CTO finding technology would maybe be easy to use to kind of bring those two together. And finally, for the senior leaders who are joining us today, if you could have them share

takeaway, one truism, or perhaps implement one tactic as they think about integrating AI into their own supply chains or procurement strategy? What would that be? One takeaway or one tactic you'd like to see them implement? Eva? So AI is complex. So I think they need to bring the right talent to help them in the journey. And also they need to put a lot of attention to the data.

the quality collection processing capabilities because these are key for effective AI applications. Wonderful. And Walter, we'll give you the last word. Disruptions are always happening, right? And we can never help. I think you said at the beginning that change is the only constant. I think Julianne, you said something to that extent. And I think that in supply chain, you can predict everything you can, doing what-if analysis or whatever, but

disruptions will happen. So I think the main thing is that there are many points of failure. So if you leverage technology, as we talked about, like disruption mitigation, AI technologies, notifications, these things will be necessary, I think, to make your life easier. I think there's no harm in getting that information, a false positive, you know, over-eager text alert, so long as it catches the real failures, right? And so helping you, leveraging technology to make your life and job easier.

Thank you. And that is the end of our time for this session. Walter, Ava, thank you so much for joining me and for sharing your insights today. And to our audience, thank you for joining this session and for being part of this AI Bootcamp series. On behalf of SAP, Fast Company, and Inc., I'm Julianne Pepitone. Thanks for being here.