Rethinking Supply Chain Planning: AI, Risk, and Sustainability with JF Gagné

UPDATE: Since recording this episode JF has informed me he's no longer working with Pendulum.

AI is everywhere right now, but is it actually helping supply chain leaders make smarter, more sustainable decisions?

In this episode of the Sustainable Supply Chain podcast, I’m joined by JF Gagné, Chief Product and Strategy Officer at Pendulum, who’s spent over two decades building AI systems that do more than just spit out forecasts. JF shares how generative AI can integrate data across the enterprise to drive contextual, collaborative decisions - helping businesses manage risk, cut waste, and improve sustainability outcomes.

We dig into:

• Why traditional supply chain planning systems are too rigid for today’s volatile environment
• How AI can move from giving “perfect outputs” to helping teams reach consensus and make informed trade-offs
• Practical examples of using AI to reduce carbon emissions, track forced labour risk, and optimise inventory
• What most companies get wrong when implementing AI in supply chains
• Why continuous risk assessment, not just better forecasting, is key to agility and resilience

JF also offers a blunt reality check: the world we built our supply chains in no longer exists. If we keep treating today’s disruptions as isolated events, we’re planning for a past that’s not coming back.

Whether you’re piloting AI projects or just trying to get a handle on growing ESG requirements, this episode will help you think more critically about what real innovation in supply chains should look like.

🎧 Listen now and learn how to build a smarter, more sustainable supply chain.

 If disruption hit tomorrow, would you know where your supply chain was most exposed? In 15 minutes my free scorecard helps you assess 27 resilience statements, calculate your score, and turn the result into three priorities and a 30 day action plan. You can download the scorecard free at tomraftery.com/scorecard.

Transcript

Copy

0:03

We have to let go of where we come from. The world we've operated over the last 60 years no longer exists. I think the biggest blind spot is to think that these are isolated events and things are gonna go back to how it was before, and it just won't. Good morning, good afternoon, or good evening, wherever you are in the world. Welcome to episode 66 of the Sustainable Supply Chain Podcast, the number one show focusing exclusively on the intersection of sustainability and supply chains. I'm your host, Tom Raftery, and I'm thrilled to have you here today. A huge thank you to this podcast's, amazing supporters. You make this podcast possible. If you're not a supporter and you'd like to join this community support starts at just three euros or dollars a month, which is less than the price of a cup of coffee. And you can find the link in the show notes of this episode or at tinyurl.com slash ssc pod. Now you know how supply chains are constantly getting wrecked by everything from port strikes to wild weather, and we're all trying to make them smarter, faster, and greener at the same time. Well meet the guy who's actually doing something about it. Today's guest, Jean Francois Gne jf, to his friends, has spent over 20 years building AI that helps supply chains think, not just react from slashing carbon by rethinking shipping routes to cutting waste with smarter planning. He's now at pendulum, a company using AI to connect every decision across a business so supply chains can finally work like they should. Today we're talking about how AI can help. Today we're talking about how AI can help you stop firefighting and start planning for resilience, for sustainability, and for whatever chaos comes next. But before we get into that, in the next few weeks, I will be chatting with Don Weatherby, who's the CEO of RegenX, Danny He, CEO of soapbox, Jim McCullen, CTO of Century, and Gary Loh, CEO of DiMuto. But. Back to today's show, and as I said, my special guest today is JF. JF, welcome to the podcast. Would you like to introduce yourself? Sure. Pleasure to be here. So I'm JF Gagne. I lead product and strategy at Pendulum. And yeah, really looking forward for the conversation. Okay. So for people who are unaware, what is Pendulum? So Pendulum builds a platform to help people plan better. We're building the first generative AI environment to actually bring a lot of insights and a lot of information from across the enterprise into each decision. So the spirit here is that every aspect of how you run your business should impact every single decision you make. Either it's an inventory decision, it's a pricing decision, it's a transportation decision. Ideally, these decisions should be aware of everything else going on in the business. And so far it's been impossible, too complex, too costly to make this happen. And with a lot of the progress we've seen in the last couple of years in artificial intelligence we can now make it happen. Okay. And what's your own professional arc? You didn't just start where you are in Pendulum. I know you've joined Pendulum recently. What were you doing up until now? It's been a long time in this field for me. I, I started in 2001, with a few friends here out of Montreal Canada my first company with a desire to help people make better decisions. This has been kind of. the thread all, all across my career. Whatever project I've been part of has always been centered around this, idea and lucky enough for me about 500 meters from from where I live now is one of the best lab in the world in artificial intelligence led by my friend Yoshua Bengio that I met very early on in my career, and it's been tremendously influential. For people don't know the work of of Yoshua with Geoff Hinton, and Yann LeCun and they've been spearheading a lot of core idea around machine learning deep neural networks, and they've been essentially highlighting the foundation of, what is artificial intelligence today. So lucky enough for me him and a lot of other people around him at his lab and, in the community here enabled me to do a lot of innovation over the years and transfer a lot of these concept, these ideas, these scientific discovery into products. I've done that, over three startups. My first one I sold in 2007, the second one in 2012 to an organisation that is known by people listening. Back then called Red Prairie. And then Red Prairie merged with JDA Software soon after and after that with Blue Yonder. So they're now operating under the name of Blue Yonder. With that acquisition I took the role of Chief Product Officer over there. I helped operate that merger between Red Prairie and and JDA Software and learned a lot about the supply chain. Fast forward three and a half year later I decided to leave and took a little bit of a pause to think about what was going to come next. And with Yoshua and, a few other co-founders, we launched a company named Element AI, which aimed at accelerating the transition between advances in AI to the enterprise. We operated that business for, for a few years and ended up selling this business to ServiceNow in 2021. ServiceNow did not have at that point much AI capabilities. Didn't really have a team at that point. So, Element AI became the core of ServiceNow AI platform. It's been tremendously impactful for, for the company and the customers. So very happy with how these things turns out. So, here I am now going at it again would say, but you know, I really think this is the right moment. This is quite of an opportunity to change things again. And what attracted you to Pendulum? I ended up running into Benjamin a year ago, and they were working on some very novel way of combining optimisation and mesh learning together. A lot of that work was coming out of the research of a very amazing professor Shobhit Shra and I started the discussion with them. I was formulating this thesis around how AI could change the way we plan. And there was a meaning of mind there, like really a good alignment looking into the technology that they've been working on and building, felt like the perfect starting point to then, evolve and build upon their technology and some of my ideas. We've talked a lot, collaborated, and finally made sense for me to just go and, and over and start building with them. Okay. Interesting. And you said that Pendulum is a platform that allows companies to make decisions based on all of the company data. How does it know what's relevant? Because not all data's gonna be relevant. True, and this is, a lot of the benefit we get from, the new AI models is we can now really have a good grasp of the understanding of the world. And I mean it like literally all things. This enables to properly characterise information as it relates to other part of information. A bit abstract, I understand, but the, opportunity here is that of these big model, we can distill these effects, these relationships, the affinity the information has and can be contextualized. And using simulation environment, you can properly balance the importance of these things to one another. And, then just test based on historical data your hypothesis. But you no longer need to formulate these hypothesis with humans. We can now formulate these hypothesis with AI and then you can loop and doing this at scale enables you to really create a body of knowledge that understand all things supply chain and can translate all of these data stream and compare and contrast them very very accurately and very fast. Okay. And this is the Sustainable Supply Chain podcast, JF. So how is one thing related to the other? At the heart of this sustainable means that you're fully aware of the impact of your decisions on your different goals. If you're to build an organisation that respect its environment, and makes sense of it. And also is aware of the impact that it has on its environment. You can make not only better decisions, but also you can track these things. And you can be conscious and take measures whenever certain decisions you're making that inevitably will cause environmental impact or other things are compensated or taking care of later. Talking with customers, what we hear a lot is, yeah, we do get new insights. We get, we do that. like carbon, I'll just use carbon as an example, 'cause that's something that a lot of customers have to now report on. Okay, we get carbon information but, beside maybe a study we'll do of like the impact of, let's say, I'll use transportation on carbon footprint, should we ship through boats or should we use planes? And when do we use planes? What's the impact? And how do we calculate this and they're gonna do a study once come up with a set of policy once and they're gonna implement them. I'm not saying that this doesn't have impact, but it's static. And the ability for them to use this in an ongoing way, for most customers, it's impossible. So, the opportunity here is to truly integrate end to end this understanding of every aspect of the business and as you're making decisions, even though sometimes it's small, but still you can consider that Hey, like this order we could use a boat and then we're gonna, we're gonna play it that way, even though our playbook, our standard playbook would call for us to use the plane, because we want the lead time to be shortened because of risk and this and that, but like, in this situation, we're pretty much convinced that this is gonna happen. So let's use the boat and then we're gonna ship it there and, we're gonna be able to reduce our carbon footprint impact by doing so. So, I think that's the opportunity here from a sustainable standpoint. And you really enable organisation to be less overwhelmed by everything they now need to consider at every single point. Okay, and I mean, I'm sure it's not just carbon, but something like this would also be possible for, for example, checking your supplier base for the likes of forced labor or other things like that, that companies are required to report on. Yes, yes, correct. and so is, you know, for waste and for all of these other consideration that you may want to include. Correct. And there are, I'm gonna say thousands now of AI solutions for supply chains. What do you think is the biggest mistake companies make when trying to implement AI? I think, a lot of customers are approaching this, trying to get the best answer. It's contrarian a bit, but I don't think there is such thing, and a lot of people are fighting to say, hey, I'm more accurate, I'm more this, I'm more that. That's all good, but in reality, that's not how the world works. I like to think about the real goal to it being to reach a consensus between stakeholders and a consensus with now the systems your AI, right? So how you get to a decision, how you get to an output, the policies that guides you there, and then the rational behind why is this policy in place? Why is this rule in place? Why are we working with the supplier instead of another one, is as important as the final decision. Most solutions you're gonna find are gonna talk to you about the output and the way I see things, is that the solution should be there to help you through the journey to the output. And yeah, the output will be hopefully good. But the rational and how you will operate your business, who you will do business with, why are you gonna make this decision or another one is more important. And reaching consensus between stakeholders on this is hard. And I'd say that actually a lot of systems are in the way of making this happen 'cause people come up with the final output and say, look at my output. This is good. We should do this. They forget about how they get there. And then the debate starts with someone else with like a different set of output. Let's say it's finance, and you got the two plans, basically banging head against each other. And a real conversation should be about how you get there, why you get there. And, what, what are your data points and references that justify that, we should operate this way or that way? This is really how we're, we're designing our system, and this is really the distinction. At the end, we calculate amazing outputs but again, like it doesn't matter if you don't get the buy-in into why this is the output. And I think this is a change that needs to happen. By focusing more of the experience around that process, you, you become more agile. There's that conversation about, oh, we do this. Why? Well, because that's how we've done things and, and it works. Well if you have dynamic conversation about your different policies or different practices, you're less anchored, you are more confident about changing them. And then you, can evolve more agility as an organisation. And I mean, the world is quite volatile these days. Now people, even if they don't want to, they have to adapt. I think that's the sport that we want our our users and our customers to to excel at, not the one of, outputting, what's like decisions at the end. That's not what's running a business is about, in my opinion. Okay. And supply chains have, I would say never been under more pressure from pandemics to port strikes to politics to God knows what else is coming next. What do you think are the biggest blind spots that companies still are not addressing? They have it all on their plate. You're correct. People shooting at their boats boats getting stuck in canals. Yeah. it's just it is, a, it is, a very, very intense. I'd say it's hard to, predict all, all of this will, unfold, but certainly people need to understand that we're in a different world. We talk about these events and I don't wanna minimise them, but we have to let go of where we come from. The world we've operated over the last 60 years is, is, is no longer exists. I think the biggest blind spot is to think that these are isolated events and things are gonna go back to how it was before. and it just won't, So, it's in, in front of all of us, but we're, we're still anchored in that old perception, that oh yeah, well, it still makes sense to build stuff in six different countries and assemble it here and ship it there and, we're gonna bring it over I think a lot of that has to, has to change and evolve. Okay. So AI is now being hyped as the new answer to everything. What do you think it can actually do well in supply chain management and where does it still fall short? Yeah. It is part of the answer for everything. I'm biased, but uh, I I actually think, we're entering in, an an era where intelligence is becoming a commodity. And the ability to reason on all aspect of what we do as human is now being augmented. And the cost of doing so is no longer that high. So, intelligence will become ubiquitous wherever compute can be deployed, you are gonna see that intelligence will follow. So that's true for products we make. It's true for for things around us. It's true for infrastructure. It's true for our assets. They're all in the process of becoming somewhat aware and intelligent. We, We've been hearing a lot about self-driving cars, self-driving trucks. But this is happening and this ability of AI models to learn to program, and it's now been made public that the latest version of OpenAI model, as an example, is performing better than 99.8% of every software engineer on the planet. Think of what this means, right? It means that anything that is software based or digital, if AI can't do it right away, it can certainly help you create the program that will do it, that will automate a task, a decision or like do some data analysis or transform information or move things around. We see all the progress in robotics, that is happening extremely fast. And, in the next coming years, we'll have some very flexible. Flexible in a sense of generalized robots. That will be able to produce at scale are gonna, that are gonna be able to, to, change the way we operate. Now, it doesn't mean that everybody will take advantage of it, but some will and, and yeah, I mean, I don't see AI as a, as a technology anymore. It's it's, we've, we've, learned how to. create reasoning entity, like reasoning digital entities and they, they just self-improve at this point and this is gonna continue to drive massive change across. And given all the instability that we are seeing, what should businesses be doing right now to future proof their supply chains? I think first it's not to treat innovation as a separate process. Start to think of and evolve their organisation as a ever innovating organisation. Supply chain needs to start to engage more than ever with all the stakeholders across the business and then set the foundation for that part of the business to to become strategic assets to the organisation. Thinking of that part of the business as a cost of doing business is, is is straight wrong. organisation needs to think of this as a differentiating advantage for them to their customers, and, innovation area. The cost and, and that usually it represents for, for organisation is, is, is huge. But it also is the opportunity to do better, be agile and, dynamic, and, innovate. And I think that the organisations will be approaching their supply chain with that perspective are the ones who are gonna not only survive, but thrive. And given how unstable things are at the moment, do you think we're relying too much on just in time supply chains? Do we need to rethink resilience and build more slack into the system? I mean, part of that evolution is now how important it is to assess for risk, right? How confident are we to see things going the way they are, they've been going. Therefore, once you get into the habit of qualifying, quantifying that risk, the immediate reaction will be to have a robust plan, and robust organisation to mitigate it. And I wanna say like, Oh yeah, you need more robust plans. You need to have contingency and all these things. So these are the consequences of you connecting, understanding and quantifying risk. And it starts there. If you are not in a position to evaluate it, to talk about it, to quantify it, I think it's a, it's a dangerous, slippery slope to just plan contingency, or you know, attempt to put measure in place without anything to evaluate them based on. You're just gonna increase your cost and, it's gonna be hard for you to make an argument without this. So I think what I'd say is the call to action to somehow start to formalise your assessment of how risky, how certain are the different outcome you're planning for. And at every aspect of what you're doing. And then the reaction should, I would say, will be obvious, but should then be based on this. But the right conversation to have is, is around the how volatile, how risky, how certain are things and likely to, to happen or not happen, like, and it's difficult to go from not necessarily having these things in place, to having them, but I really believe that's the right path. Let's talk a little bit about waste and the ability of AI to help with that because companies are often throwing away billions in excess inventory. How can or can AI help reduce overproduction and cut down on markdowns? Yeah, back to this risk assessment. This is really an area where, this technology can help you have a more educated guess on not just demand, but also the probability of things being where they're supposed to be. in the quantity that they're supposed to be. I mean, there's different root causes when you think about, about waste and again, depending on the industry part of it is, the way sometimes suppliers will offer deals deals. They'll, tell you if you order 10,000 more, I'll I'll give it to you to the same price or like just some crazy things like this. Sometimes that you're gonna see. And that pushes buyers to go totally over and stock a lot of inventory around certain products because their objectives, their incentives are, are centered around just a cost. It's like, Oh, I reduced the unit cost. This is great. Part of the answer is, yeah, we want to be a bit more accurate. Yes, we want to understand risk better, but also coming back to, to. the idea of being conscious of, of the whole, is that, incentive mechanisms and the metrics you look at ideally are also aware of the rest of the organisation. And a sustainable supply chain means that your buyers needs to also factor in waste. So if their only metric is to, buy at the smallest unit cost for a given budget, you're gonna get with some, you know, questionable behaviors sometimes from a waste perspective. Connecting the dots here and making sure that you do have full visibility on an ongoing basis as like each people are making these decision across the business, I guess is the is the real answer here. We can get better output and more precision. It's gonna help the waste, but it's also that, approach, that mentality of, of making sure we got 360 perspective as we're making decision across the business. The world is now obsessed with AI and there's so many dystopian visions of AI out there, you know, hello Skynet kind of thing. To be a bit contrarian, what's a utopian version of AI in supply chains that we should actually be aiming for? Uh, Oh. I mean, there's a, there's just that opportunity to localise a lot of the making of things. There's that, amazing opportunity also to understand what to make. I think we got unprecedented access to people desires and ability to understand trends and, and have that shape how everything we make can evolve the ability now to not only design things that people want and, and be agile about it, but to also change the way we make things, taking advantage of new material, taking advantage of, new processes, enabling current processes to, to adapt and evolve at a cheaper price and faster than ever I think can open the door for unprecedented abundance. I think that's the process that we're all trying to work towards. Not that there isn't a lot of danger in getting there, but my utopian view is can create a new, a new era of, of abundance. I think for for everybody if we truly take advantage of what's in front of us as a, as a technology. For companies struggling with supply chain inefficiencies, what do you think would be the one change they could make today that would have the biggest impact? I think it goes back to your question around waste for me. It's starting to ask yourself the right question and to, measure what's in front of you. And even if your tools are limited, try to quantify these things. You'll get better at quantifying them. You feel pain sometimes let's say your, your back hurts, but ends up that, you know, you have, a a foot problem, I don't know. And, then the way you walk essentially is what creates the problem in your back. I dunno if it's the right analogy, but the first discipline that they need to get better at is in assessing their environment, quantifying it and, and at least like starting to do it manually, like, just like, state how you think things are, and then take action based on this and then look at if like your appreciation of this changes or not, and then try to measure it better. But the starting point for all of this is to be conscious and mindful of how things evolves around, around your supply chain. You need to start there. There's a million people proposing tons of solutions. Again, sometimes what you need is, is not a heat on your back. Sometimes what you need is, is someone to take care of your, your foot problem. The only way you get there is by starting to, to, to quantify and make sense of the surroundings. You operate in the surroundings around around your supply chain. And once you start to structure this understanding, it can be very rudimentary. Like You'll get better at assessing how things are, and then you'll figure out the breadcrumbs that leads to the root cause. If you had unlimited resources and zero constraints, what's the dream supply chain innovation you'd build? Oh I, again, I look biased, but what we're working on right now is I think the single most important innovation that can be brought to supply chains, I fundamentally believe that the time is is now right to truly have this true business digital twin and, that partner to help organisation figure out the solution at the back and implement them for the right reason. I think it's at the heart of everything else, 'cause this shapes your investments, it shapes how you operate, it shapes who you work with, it shapes what market you go after, it shapes how you serve your customers. It's fundamental. And I think this is where, when I say intelligence needs to go all across and it will end up everywhere, I think this is the first place where it needs to go. Current infrastructure, current system, current software or not, designed take advantage of what's possible. So, yeah, I know, I fundamentally believe in what I'm doing right now. Great. That's, that's important. Left field question. If you could have any person or character alive or dead, real or fictional as a spokesperson for AI in supply chain, who would it be and why? Tony Stark. Why not? Cause when you, when you look and think about how he's interacting with his AI and the impact he, he is not for the necessarily the military side of things, but for everything else that his, his industry does. I think, I think that's a version of the future. Okay, superb, superb. We're coming towards the end of the podcast now, JF. Is there any question I haven't asked that you wish I did or any aspect of this we haven't touched on that you think it's important for people to be aware of? No, I think it's pretty good. I liked the conversation. Fantastic, great. Okay, JF, if people would like to know more about yourself or any of the things we discussed in the podcast today, where would you have me direct them? They can find me on LinkedIn. JF Gagne or, and obviously look at our website pendulum dot global, and you'll find contact form and, and then different pieces of information, hopefully that, you know, we'll tell a little bit more about myself and the company. Fantastic. Ok JF, that's been fascinating. Thanks a million for coming on the podcast today Great pleasure. Thank you, Tom. Okay. Thank you all for tuning into this episode of the Sustainable Supply Chain Podcast with me, Tom Raftery. Each week, thousands of supply chain professionals listen to this show. If you or your organization want to connect with this dedicated audience, consider becoming a sponsor. You can opt for exclusive episode branding where you choose the guests or a personalized 30 second ad roll. It's a unique opportunity to reach industry experts and influencers. For more details, hit me up on Twitter or LinkedIn, or drop me an email to tomraftery at outlook. com. Together, let's shape the future of sustainable supply chains. Thanks. Catch you all next time.