AI is Transforming Sustainability

In Part 1 of this two-part conversation, Reminger’s Zachary Pyers sits down with Evan Schwartz, Chief Innovation Officer at AMCS Group, to explore how artificial intelligence is transforming recycling, logistics, and sustainability. 

From “reverse logistics” to predictive truck maintenance, Schwartz shares how AI can connect the dots between processes, uncover efficiencies, and prove that doing the right thing for the planet can also drive profitability.

Stay tuned for Part 2 next month.

Transcript

Copy Transcript Copied

REMINGER REPORT PODCAST ON EMERGING TECHNOLOGIES

Evan Schwartz - Episode 1

ZBP      Zachary B. Pyers, Esq.

ES         Evan Schwartz

 | Narrator | Welcome to the Reminger Report Podcast on Emerging Technologies. Reminger LPA is a full service law firm with over 150 attorneys across 15 offices serving the Midwest.  Join your hosts, Zachary Pyers and Kenton Steele, both partners in Reminger’s Columbus, Ohio office, as they interview entrepreneurs, practitioners and thought leaders. Together they explore how emerging technologies and evolving business models impact our daily lives and how the law is adapting to these changes. This episode is Part 1 of a two-part interview with Chief Innovation Officer Evan Schwartz of AMCS Group.
| ZBP | Welcome to another episode of the Reminger Report Podcast on Emerging Technologies. I'm Zach Pyers and I’m really fortunate today to be joined by our guest, Evan Schwartz. Evan, welcome to the podcast and thanks for joining us today.
| ES | Thank you, Zachary. I’m happy to be here.
| ZBP | So Evan, I know you’ve got a couple of roles. First we’ll talk about you’re the Chief Innovation Officer, is that right, of AMCS Group.
| ES | Correct.
| ZBP | And you also happen to be an adjunct faculty member teaching courses in project management.
| ES | And AI analysis, so I do two fields at Drexel University.
| ZBP | We are excited to have you on the show to be talking about some of the innovations that you’re seeing, both in the fields and I know AI is one of the hot topics of the day and it’s a hot topic in almost every industry and of course in the legal industry. So we’re, I’m glad that you’re going to be able to talk about this. If you would, tell us a little bit about AMCS Group and what you do for them and kind of what they do.
| ES | Yeah, so AMCS has been at the forefront of innovation for decades. The reality of me being appointed to Innovation Officer is just a result of the growth in the size of AMCS. We’re not suddenly innovative because I showed up. We’ve been innovative forever. It’s just we reached to a size that it now needs a declarative focus to maintain the acceleration to get innovation out the door. So we still have all of the same incredible engineers, incredible product organizations, and we have the great vision that is driving the company. Jimmy Martin from beginning to end has said there has to be a better way. In resource intensive industries, we spend thousands of years. You can pick it, whether it’s mining, it’s getting oil out of the ground, it’s forestry, it’s building, you know, metal refineries. We spent thousands of years of being really good at getting it out of the ground, putting it in the supply chain, put it into a product and shipping it. We are not very good at recycling it. We’ve only been doing that for a few decades, and that’s a, that’s a reverse logistics problem. It’s, you know, when I go to the pit to mine or I go to the forest, it’s always there. I know where it is so I can build and optimize it, but the reverse logistics problem is we’ve turned it into products and we’ve sent it to millions of people and now we’ve got to get that material back and bring it back into the supply chain. How do you do that in a way that makes that material still profitable held up against a process that humanity has spent thousands of years making as efficient as humanly possible? That takes a lot of technology. That takes driving deep. If you’re familiar with the process diagrams, here’s the star, here’s the square, here’s the triangle as it goes around the process. You have to dig deep into each one of the squares to understand how to make it efficient and we’ve been doing that probably the first two decades. So now AMCS is committed to a sustainable way to drive our economy, our entire industry across all industries. There’s got to be a better way, and we started to look at the connective tissue between the processes, the handoff from one to the next. And we started to see connections.  Now I, you and I are old enough, you could go back 10 years. Big data. There’s so much data. We have too much data. We don’t know what to do with it. And now with the advent of AI and patternistic, I can now digest that AI and I can see clustering and I can see patterns. That’s all, you get right down to building a model. That’s what it is. It’s feeding it incredible amounts of data and finding patterns. And now we’re connecting the lines between those patterns and finding efficiencies where nobody was looking because it was just too deep of an ocean of data to find it. Example in some of our products is our route optimization. If you’ve got a truck out there running a route, it’s, no matter how optimized it is through computer AI, it’s the ideal rep. That’s, if everything goes right, this is the ideal rep. There’s a lot that goes wrong in reality. We live in a real world. And so one of those is if I’m out servicing a container and I’m expecting to pick up a 90 gallon container with approximately 90 gallons of stuff and that lid’s open and there’s garbage all around it, I’m picking up more than 90 gallons. My ideal route just disappeared. I messed it up. So now I might have to tip sooner. So being able to capture that with Vision AI, I don’t want to burden the driver because then the juice isn’t worth the squeeze if he has to get out and he has to record it and he has to do all this. So again, that’s the same reverse logistics problem we keep bumping up against is I’d love to do the right thing but the cost to do the right thing outweighs the value it brings. So we, we keep pumping and perform that sustainability. Here’s an optimized route to save fuel, to save wear and tear on my equipment. Here’s Vision AI to make sure I’m doing the thing that I sent the truck out to do. And then the health and maintenance of the truck is another one.  If the truck is running bad or the engine’s about to fail, I’m not getting efficient. I’m not getting my fuel savings. It’s not running really well. We now have AI that I can take almost a blood sample from a vehicle or any asset and find out, hey does it need to be serviced. Why would I take a truck out of service? Most predictive analytics for serving a vehicle is kind of like the one for our car - come in every 3,000 miles. You don’t know that I need to be serviced. I just rolled this thing off the lot and I need to take it in 3,000? For crying out loud. So that right there reduces the introduction of oils into the ecosystem. This truck is how my company makes money, so for every day it’s out and has to be serviced and put together is opportunity money I’m not making. So when you stack those three, driven AI analytics based on sensor data, together we found that there’s a huge multiplicative effect. I’m maximizing the value of my ideal route. I’m now catching revenue leakage on overflow containers. Or it can also identify contamination. I gave you a rate based on the fact that I think that’s a value stream, but if you contaminated it, it’s no longer valuable to me. So it’s just all about being able to sense the environment, making sure the hand-down from one step to the next is getting you as close as possible to that ideal outcome. And that’s just, that’s just a logistics thing. When you start to apply that pattern across the entire business, the outcomes are phenomenal. You realize that all the leakage isn’t necessarily in the activity. There was, but we spent a lot of years putting a lot of good scientists on figuring out how to make that activity better. It’s the, it’s the handoff. It’s the connective tissue from one step to the next that we’re now able to see. There’s just so much data, and AI’s really opening that up. So from an AMCS standpoint, the thing that gets my heart pumping is that we’re building software that is performant sustainability. I’m not asking you because there’s a boot on your neck by the government to be compliant and do the right thing. I’m saying if you did the right thing, you’re going to make money. Well, who wouldn’t do the right thing? You know, the ESG reporting as we went across several of our businesses, we were able to identify everywhere you have the highest CO2 output, whatever step that is as you’re reporting it, always is your most efficient, inefficient. That’s where you can build efficiencies. So it’s great that that’s a reporting requirement into whatever country or county you’re trying to get into but also use it as a leading indicator, that’s there opportunity for me to make money there, to be able to either become more efficient, to be able to grow my business. You know, one of our customers had 13 vehicles running all of those routes and by the time we were able to add all this, he was able to part with three of them and still do the same service in the business with only 10. That’s almost $3 million a year savings to him to not put those trucks on the road, and that’s now, he can grow that business without having to invest any more capital funds. Those are the wins. That’s performance sustainability. I could sell that all day long, and now companies are doing the right thing to help with the environment and they’re making money. Why wouldn’t you do it?
| ZBP | You know, one of the terms that I, that you used that I really like is this concept of reverse logistics. I mean I’ve heard the term reverse engineer probably throughout, you know, my life but I haven’t necessarily heard the term reverse logistics. And I think it’s a really interesting concept because in the resource world that we live in where everything comes from somewhere, from the shirts that I’m wearing to the pens that I’m writing with, I mean, we have so many resources, the idea that you would be able to essentially do, reverse the logistics with them to help to obviously benefit sustainability and all of those things, it’s just an interesting concept from a resource perspective that I don’t know that I’ve heard a lot.
| ES | Look it’s, we, I don’t know that we coined the phrase in the industry. I’m not going to claim that, but you will see that there is, there’s awareness of it. Even Elon Musk with his Tesla cars and all the batteries, he’s trying to encourage entrepreneurs to build recycling plants next to his, his lots where he sells them. Because look, there’s another, is scrap doesn’t like to move far. The problem with recycling is any value that’s in that material gets eaten up trying to get it back into the supply chain. So I, there’s no way I’m going to take a car, grind it up, take all the parts and then ship it across the United States and it be more valuable than the thousand year optimization of digging it out of the ground. And it’s, it’s counterintuitive to think about that we’re mining lithium in Africa and other countries, putting it on a truck, putting it to a barge, shipping it to another country like Sweden, making a battery, putting it on a truck, putting it on a barge and shipping it to a manufacturer that puts it in the car, puts it on a truck, puts it on a barge and ships it to America to sell it easier and cheaper than I can collect it locally and then recycle it. And it just shows it’s not because it, that’s, that is such a waste of energy that if we could add the technology to track to get and source the material, it’s an obvious win. It’s clear that it’s obvious. We just need to do it. So most of this we’re going to see starts putting upward pressure all the way back to the producer. We’re seeing movements now in the industry of extended producer responsibility. I could get your product, I could get this thing back into recycling easier if you change the way you build your product, Mr. Wizard. It’s too hard to get your dyes off of it. It’s too hard to break apart these components into simple things, but AI is helping there. You’re seeing out in the MRFs, the material recycling facilities, where they’ve got robot arms and eyes reaching in and putting things and separating. What’s going to get so good, they’ll be able to put, and no offense to any brand I mention here, but you could put a Coca Cola bottle in one bag and a Pepsi bottle in another bag and now I can tell you which one can be recycled easier, which gives me a little bit of a boot on the neck of the other one in going, “Guys get with the program. We would be happy to recycle this if you make it easier to recycle.” But how many product development teams are thinking about how this product gets back into the supply chain during design? Most of them aren’t, right? So we’ve got to change the conversation.
| ZBP | And I know, only I recently went to a landfill/recycling sorting center on a field trip with one of my kids, and it was an eye opening experience to watch them sort because they were still manually sorting the recycling. And I just assumed, I dutifully recycle all that I possibly can, and I just assumed that it was all going somewhere and being turned back in and then they were like, well, actually there’s no market for this, or there's no market for that. And they explained how quickly, for example, like an aluminum can can be recycled and how quickly it goes back into being an aluminum can vs. a plastic bottle. And they said, well, here’s, this is how the plastic bottle, and the process is just wildly different for both of them, which I as a consumer had no idea about until I stepped foot in one of these processing and had some --crosstalk--
| ES | Where this is a big deal, I mean you could across any industry and see the same thing. Aluminum cans, batteries, lead - success stories. Actual virgin fiber, pulp, trees, the tree farm - success stories.  All the log trucks you see on the road today is the growth for that year in the United States. If you think of it that way, so as trees grow, well we’ve cut a tree, yeah, but it’s, we didn’t lose any of the fiber or trees by falling them, okay, which means that we don’t really harvest trees in the United States and we still prep and we still plant. We plant more trees in this country than we cut, and I know that because my background before I came to AMCS was, it was all forestry products and land management, and it wasn’t even 12 years ago in the southeast. And I’m not going to mention the name of the company, but there were 25 virgin pulp mills pulling in tons of trees every day every 30 seconds. There’s now 2, 2. Because we can now source cardboard and white goods and white paper and then we repulp it and turn it back into paper. Now eventually the ligaments break down, but when we first started, we could only recycle a box two or three times. We can now do it up to seven. That’s what science and innovation gets you. At some point they’ll get that up to 10 or maybe I will never beat it down to where the fibers get so broken you can’t make it. So we are just now adding, I don’t know the math here real quick, but 2 out of 25 is the percentage of virgin fiber we have to put it in. You’re seeing virgin polymers now taking that same approach. How do we mix in recycle with virgin polymers to get the same type of recycle? That’s why mostly in the plastics you’ll see 25% recycled plastic because it, we’re getting better and better at the chemistry. But that’s really the point of my first statement is we spent thousands of years figuring out how to make stuff, maybe not that long in plastics, but we’ve gotten really good at taking oil, making polymers and shipping it out. We never thought about taking it back in and what we would be able to do with it. Some of it’s reuse, repurpose, refurbish or reclaim is a way that I can use it as a form of energy. I could just transform it from plastic to fuel through science, turn it back into that. So we’re looking at from a full variety of opportunities. Could you imagine if we even got close to 70% recycling of those materials, how efficient we would become? Some of them can recycle indefinitely - lead, aluminum, lithium can be recycled. So if you really look at what Elon’s doing, love him or hate him, he’s doing a distribution of lithium around the world right now and then at some point, we should just stop mining it out of the ground because it’s all local. I can just sit there and recycle what’s been distributed over and over again. So at some point the hockey stick kicks in so only time will tell whether that’s genius or bad or not. But there’s a lot of those where I don’t think we’re telling the narrative appropriately on the direction of travel.
| ZBP | The lithium example for the, obviously for the electric vehicle batters and all of the batteries that use lithium, is an interesting concept because I know there’s been a big push by a lot of countries to get access to these type of materials and minerals because they’re valuable.
| ES | Right.
| ZBP | And I know that even over the last couple of years, we’ve, we’ve seen countries talking about what lithium reserves they may or may not have, how they’re accessing it, should they open the mining, and the idea that this “precious resource” may not be as valuable in the future isn’t something that even, because to your point, we are so used, I think as a resource-intensive culture, I keep thinking about all the oil that’s been pumping out of the ground for the last 100, 150 years and we haven’t really stopped.
| ES | Oh yeah, you and I are both old enough to remember the ‘80s where we’re running out, it’s gonna run out. We ain’t run out yet.
| ZBP | Nope.
| ES | My actual, my geography teacher when I was in high school, Mr. Higgins, I hope he’s doing well, it’s been a while. But he used to say that we would never run out of anything. It was a, even back then he called it a red herring, and what he was saying is because we become more efficient at using it, so we’re able to get the same out of half and then same, so it’s the, it’s the, I’m cutting the thing in half, you never get to the end type of thing. Now obviously that’s counterintuitive. Humans, you can consume things and it goes away. But as I look back over, when was the ‘80s, oh my God, 50 years ago, some, some ridiculous amount of time, we’re, we’re no, there is no, there is no crisis. There’s no cars lined up trying to get fuel like it was in the ‘70s and early ‘80s, so obviously to some degree he was right. We figured out how to extend the capability to make it more efficient.
| ZBP | Now I know one of the things you talked about when you were talking about the trucks and using the sensors and the AI was, we kind of talk about this role where you’re not replacing the human that’s driving the truck. You’re not getting rid of him or her or they, but you are augmenting what they’re doing and you’re kind of, you’re assisting them in the process. So when people think of AI, I think one of the things that we oftentimes talk about is how it’s going to, we hear this kind of, I don’t want to say doomsday prediction, but
| ES | Sure.
| ZBP | It is. Where AI’s going to replace all humans, and I know that I’ve heard this from people, and I told you before we started recording, I had this question this morning in one of the classes I was teaching, because I’m an adjunct faculty member as well and one of my students, who’s a law student, asked whether this process that we do in these legal proceedings will, will it, will it essentially be totally done by AI in the future. And I’m, I’m curious to kind of get your thoughts on the role that AI plays, not only in your work with AMCS but kind of in a larger, is how you see it impacting the human involvement in so many of these processes.
| ES | Yeah, so let’s, let’s use the truck as an example to explain this. Right now if we were to take all of the AI, the sensors, off of a, of a, let’s just go with the garbage truck. I don’t think people realize it’s a mobile data center. There’s so much tech on a garbage truck today, it’s unbelievable. But if we took all that away, we’re expecting the driver to know where to go, and where I live, they’re always take-all, but in a lot of places every household can shop for their waste service kind of like you would a utility, which would mean that a street could have five or six vendors. How do you know which one to pick up, so there’s a challenge. The second one is any abnormality. There’s a truck parked in front of it, I can’t pick it up today. He’d have to manually know any problem that exists all becomes a burden to the driver. This is a 25 ton vehicle that he’s driving that he’s distracted. He’s trying to do multiple jobs at the same time rather than drive a, trying to be safe, trying to do the service that everyone needs him to do and get it done. You add all of this AI. I didn’t get rid of the driver. The driver now has the ability to focus on driving safe, running his route. He can just quickly get, a glance or audio can tell him to take a right here, you’re picking up at this house. It’s capturing the details of the event. If it’s blocked, he doesn’t have to do anything. I got a camera seeing that it’s blocked and it records it, identifies as blocked. If the lid’s open, it’s overfull, as you’re dumping the hopper I can see it’s contaminated. All of that can happen off stream. The driver just does the job he was originally hired to do before everyone tried to put all of this other compliance on it. That’s a great example. But the reality of it is, there’s nothing where AI today comes with a hundred percent predictable, repeatable perfection. It doesn’t exist anywhere in the models that we have today. There’s always stochastic natures. There’s always errors. There’s always something. So use cases are, it needs to be as good as the human because humans make mistakes, so there’s, no one’s perfect. So as long as it’s as good as the human, I’m okay with that, too. And we’ve now enhanced his experience in his job to do it. If I were to switch to a different role, one that we talked about in the pre-review, is like the cameraman, the guy that is an art photographer. Did he just get replaced because AI can generate beautiful images? Well, how did it do it? It didn’t go off on its own and create a beautiful image. The story I was talking about where he was questioning is whether or not he needs to go back to school and learn something new, is he typed in a prompt, he gave it specific lensing, flaring, contrast, black and white. He described what he wanted and handed it just a selfie of himself and it came back stunning. And I mean it, it had that emotional reaction you would expect an artist with a camera to produce. And he was, he was, at first he was bummed. He thought, alright I’m out, I gotta go do something else, and then after reflecting on it for a while, he came back to the decision is the only thing that changed was the camera. Before he would have to take this camera, go find the setting, get the lighting set up, get his model or art or whatever it is he’s taking a picture of and then snap it with the right camera with the lensing, everything configured just right to get that image that he already had pictured in his mind. Or he was able to describe it to the AI tool he was using and provoke it. So the only thing that changed was the camera. It needed him to describe that to produce that art, so I don’t see that any version of this is replacing the person. It, it’s amplifying the capability so if you’re investing as a business in training your employees how to work with agents, how to do prompt engineering and how to manage the tools. I saw, I saw a post the other day is, you’re not going to be replaced by AI, you’re going to be replaced by the person who knows how to use AI. That’s probably a very realistic approach to it, so you need to get under the hood here. I will be able to scale significantly. Something that a CSR would be able to do in a room of 25, I could probably get one CSR to now do that at the amplification level of x100, being able to orchestrate agents. That doesn’t mean that I fire my 24 other CSRs that have learned it. That just means that I now have incredible growth opportunity and I need to figure out how to leverage these people that I’ve invested in training in other areas of growth. But you, if you’re building AI and you don’t have an on-ramp to a human, you’ve done it wrong. I can tell you that right now. You, if you’re building agents, you need to have a confidence floor that if, because, the AI is a statistics engine. I am, I have this confidence rating on this answer, this one, and this is the highest one that I’ve given to you. So you can decide the level of accuracy that you, you really need in this activity and go, if your confidence level drops below 80%, on-ramp to a human. Give it everything that you’ve learned and let a human take care of it. That is the right marriage of this technology, and now that person may be commanding 200 agents and about 3% of them are escalating to the human to resolve. Humans become the premium service. Humans are the creative ones. That picture that that photographer got out of AI didn’t come from AI, it came from his head. He just told it how to do it. That’s the future to me.
| Narrator | Thank you for listening to the Reminger Report Podcast on Emerging Technologies. If you enjoyed this episode, don’t forget to subscribe and leave a review. For more insights, visit Reminger.com or connect with us on social media. Stay tuned for the next installment in this series where we’ll continue our conversation with Evan Schwartz about managing change in our rapidly evolving relationship with technology.