Supply Chain
Technology Podcast

SUMMARY KEYWORDS

Active risk management, passive risk management, supply chain visibility, machine learning, data contextualization, real-time monitoring, prediction models, intervention capabilities, root cause analysis, device performance, carrier performance, automation, risk quantification, control tower, predictive time of arrival.

SPEAKERS

Scott Mears, Ilya Preston

Scott Mears  00:07

Welcome to the Supply Chain Tech Podcast with Roambee. Scott Mears here, Global Field Marketing Manager at Roambee and one of the hosts of the Supply Chain Tech Podcast. We thank you for joining us today, in this episode, we speak with Ilya Preston. Ilya is the CEO of Paxafe. Today, we discuss how active risk management is evolving the perishable goods industry’s approach to risk in supply chains globally, we learn about innovative technology that allows us to predict risk, and we also have an interesting conversation about the challenges solution providers face when trying to enable this technology with their customers. Welcome. Ilya, thank you for coming onto the podcast today.

Ilya Preston  00:57

Thanks for having me, Scott.

Scott Mears  00:58

Awesome. I love to always start off with a bit of an icebreaker. And I’m actually quite excited for this question, because I’ve not asked it before, so I’m quite keen to know what you say is, I would love to know if you had a different role in supply chain, what would it be and why?

Ilya Preston  01:19

Yeah, that’s an intriguing question. I think for me, at the end of the day, it’s I find myself living vicariously through our data and machine learning teams almost daily. And you know, they’re doing some of the most interesting and unbelievable work around supply chain prediction, around risk quantification, around data contextualization, I think at the end of the day, it would probably be a machine learning engineer. I think there’s just so much new data, so many new data sources, and unbelievable insights that are still lack or that are still trapped and, you know, and are not coming to the forefront for companies that invest in supply chain visibility. And I think being a part of that effort to unleash the value of visibility is something that’s very intriguing to me. So I think it would have to be, you know, and we have a number of great data scientists and machine learning engineers on the team that come from all walks of life, that, you know, some of them were a part of the Last Mile Data Science Amazon team. So they’ve done some really, really interesting things in the past, and I think what they’re working on is contagious, and I find myself struggling to not be a part of every and any, any meeting that they have, because it’s just so interesting.

Scott Mears  02:51

That’s a wonderful answer. And yeah, more and more we’re using machine learning, and more and more it’s going where I’m standing, the power it has. So that’s that’s really a great insight to you and where you would be placed differently if you weren’t CEO of Paxafe. That’s really interesting to know. And what I want to know first is I want to first, at the start of the podcast, I want to really define between passive and active risk management. So what I want us to do together is, I want us to first to define and highlight the key differences between the two different approaches of risk management today. So the traditional approach of the passive approach, and with the more adopted approach today, which is or the becoming the more adopted approach today, the active risk management. So I want to define what these what these are, and the key differences between these two approaches.

Ilya Preston  03:51

Yeah, so, I mean, I think the key difference is passive is obviously not real time, right? Active is real time when it pertains to location and condition monitoring. You know, for me personally, when I think about weighing the pros and cons of each I have no dog in the fight. And, you know, Paxafe has no dog in the fight because we support both active and passive loggers within our platform quite agnostically, right? That being said, I think historically, the calculus of a company trying to determine whether passive or active is the best fit for the lanes has always been very unilateral. It’s always been essentially, you know, is this a high criticality product? Is is this is the lead time on this super urgent? Is it a highly sensitive product? And it was almost a binary yes, no amongst those items, and if the answer to the majority of those questions was yes, active and real time visibility is necessary. If the answer is, this is. That critical product, it doesn’t have an urgent lead time. Maybe it doesn’t require, you know, 24/7, visibility to the customer. Then then passive may be a better fit. I think that line of thinking, that calculus has has evolved over recent years, where companies are now starting to understand some of the some of the value that can be unlocked that’s more difficult to quantify and not as a parent. So for example, the ability to automate workflows by using real time active devices, the ability to build prediction models and actually identify risks before they occur, right? Those are things that are that are not possible with with passive data loggers that are sharing similar data with you. But after the shipments already been complete, after the shipment has has arrived at the destination, obviously, you know, one of the key differences is the ability to intervene right into and salvage any product that may be at risk with with passive data loggers, you can’t do that. You don’t have a real time feed on location and condition data of the product. So if there are any sort of critical delays, if there are any sort of temperature excursions. It’s, it’s, you’re going to know about them after the fact, but you’re not going to be able to salvage the product, the product, it may already be lost. Obviously, with with active devices, you have the ability to intervene. You know, you typically have some sort of a control tower that’s monitoring these dots on the screen. And once you see that an excursion has occurred, and once you see that you know temperature is out of spec, or a product is been sitting at a particular location for six hours without movement, you have the option to then jump in and whatever the intervention process may look like. It may be direct, it may be through the carrier, it may be through the broker, the three PL, you have the ability to kick off that process and have a chance at salvaging the product. So I think, you know, I think there are a number of things that are impossible to do with passive data loggers as compared to active. But I also think that the calculus towards the decision making between active and passive is very fast evolving. And I think what used to be a very binary decision between Hey, should I should I use passive, or should I use active is now a much more nuanced and intentional calculation and consideration. So I think that is very much evolving. And I think you know, Scott, you and I were both at LogiPharma just a few weeks ago, before this conference, I don’t know that I heard universally from pharmaceutical companies, healthcare companies, that they were 100% on board by with with a vision to go active right for most, if not all, of their lanes before this, at least the conferences I had gone to, the discussions that I had had, I think there was always this sentiment of, you know, we’re we’re going to invest in active but we’re still going to keep x percent of our lanes passive. We just don’t have the need to go 100% real time across everything. That wasn’t the case this last conference, right? I heard universally that we intend, in the next three to five years to be 100% active, from a majority of you know, the healthcare organizations that I spoke with. And I think you know there’s, there’s a number of contributing factors to that, obviously, improving device performance, reduced component device component cost. I think there are a number of things that are making active more and more attractive, healthcare organizations.

Scott Mears  09:13

That’s great point. Very much to sum up the difference between those two is, do you want to know problem has happened before, after it’s happened. Or do you want to know the problem either it could happen or it’s happening in that moment, and you can do something about that. It really is that simple, between passive and active and and I agree seeing that change, because I’ve been to other events in farm supply chain, where there’s been a real resistance, as you had mentioned there, to that move to even take 100% of real time, sometimes not to take it on board at all. But I completely agree it was a real move in this particular in lodge farm, USA, to people really pushing forward and even. Going down to the level of robotic process automation, and automating it to even a higher level. So this was really good, seeing a great sign for things to come. And what I want to do, because there are still a portion of the industry that in, and I’m talking now within cola chain, within food and beverage, within pharma, that still have resistance to move from that, that passive tactic. What I want to do, before we really jump into how active is changing the game, I want to dive into a few things that comes up quite regularly that would be that create a resistance to for that move to happen. So I’m just gonna, sort of, we’re gonna sort of go through them one by one, and I feel these are just some of the things we hear a lot. I think one which technology companies feel a lot is the impact of buzzwords. I mean, I mean, for Roambee real time. I mean, I even refrain from saying those words in conversations, because it’s just so loosely used, it’s kind of misunderstood a lot of the time. AI machine learning these sort of buzz words. I would love to learn more from your side of how we can overcome this as technology companies, to overcome that problem of all these buzz words creating confusion, miscommunication?

Ilya Preston  11:25

Yeah, that’s a great question. I you know, so delineating between real time or near real time, for example, yeah, is something. When you speak to another professional that lives in visibility, they’re going to know that nothing is real time every second, right? At least nothing that is economically feasible, that can last from an energy budget perspective. So I think when you when you say real time, and you’re dealing with a professional from the industry, they’re going to understand that real time could mean anywhere from five minute updates, maybe more standard 15 minute updates, but you know at most an hour at once, an hour, right? But when you’re talking to you know, potentially their IT teams, or you know their digital transformation teams, that that may not have lived in visibility before this. They take that real time to be they take it quite literally, so you know, to your point on generally using these buzzwords and the maybe the effect that it has on both the industry and the startup, I think upfront, there is a an immediate resistance and almost a loss of credibility. Because these, a lot of times, we do find that these words are massively overused, right? I could, you know, I’ve been monitoring the competitive landscape. Even bigger companies do this. They’ll, they’ll say, you know, we’re, we’re AI enabled, and we have all this machine learning capability. And then you go look at their employees, and they have no data people, zero, maybe, maybe one data analyst, right? That’s, that’s typically a telltale sign that something isn’t adding up. And, you know, it’s, it’s, I find, even in my own journey, you know, when, when we start to speak to somebody fresh, and they’ve already tried somebody who’s AI enabled or has machine learning, and it flopped, right? It didn’t work. That’s, that’s a loss of credibility for for startups, right? So I think, I think it’s important to, number one, be intentional with the use of those words. Number two, I think it’s important to have a standard definition of what they mean in our particular world. You may have heard PTA predictive time of arrival as another one, right? And I don’t know how many times I’ve heard you know when, when we disclose that and share that we have a PTA that’s 95% accurate? Well, my carrier can get me a PTA. They’ve, they’ve offered me a PTA, right? And then we dig a little bit deeper, and we find that it’s, it’s a carrier promise date, right? Or, or the the intended lead time of the shipment. It’s not a PTA at all. And that’s come up many, many times. So I think having a a standard, almost universal definition of all of this terminology, what it actually means, what would be a net positive for the industry? You know, at the end of the day, in our case, how we kind of overcome this is we’re confident in our taxability, so we open. Up our platform, and we say, go try it. Go try it for a month for free. And you can typically very quickly see if it’s working or if it’s not. And once they realize that once, once our customers realize, Oh yeah, this, this is much better than anything I’ve tried before from a capability standpoint, specifically when I’m talking about like aI enabled features, I think that tends to build that credibility back. But to your point, it is a huge barrier at the beginning of the relationship, and I think the best ways to overcome it are to be very intentional with the use of these words in your marketing language. And number two, to align on a standard, universal definition of what each of these mean in the context of the service that you provide, and make sure that the industry has access to that universal dictionary, if you will.

Scott Mears  16:00

I like that. It’s be intentional, align the standard and and what you also do is you prove a value. Go, look, have this for 30 days. Check this out. Look, go in and play with it, and that’s where the real value is. And you suddenly realize, okay, this is true. This is true. AI that that’s really good insight. And I think learning that, that for all technology companies and companies trying to push this out, and just the industry as a whole, because it’s, it’s a struggle, and it’s, it’s a shame when these buzzwords could prevent something positive happening and bring in, bringing bringing more value to to the industry. I’d love to tap into another one that I hear a lot as well, is, and this is an interesting one, because I sort of took me back when I heard this one. But when I thought on it, I went, Okay, I understand this is what if the root cause that you but that we find is too costly or cannot be fixed, or cannot be fixed immediately. What? What? How do you overcome this with that sort of reply to to bring in more visibility?

Ilya Preston  17:14

Yeah, so I think there are two parts to the question. I think you know, the the root cause being expensive pieces one, but I think even before that, you have to make sure you get an accurate root cause. And right now that’s that’s very difficult to do without additional AI enabled contextualization around the data, right? So, for example, when you say root cause, you’re, I’m assuming you’re, you’re probably referring to something like, why was the it’s not enough to share that hey, this shipment was delayed, or, Hey, this shipment has a temperature excursion. You need to understand what caused that delay? Yeah, what caused that temperature excursion? Was it insufficient thermal life on the packaging. Was it a malfunctioning reefer? Was it an, you know, an open truck door? Auto diagnosing that and getting to an accurate root cause is a really, really, really long journey, unless you’re going to do it manually for each and every single event, right? So that’s the, I think that’s the first half of the equation that you have to get over because or that you have to help your customers get through they’re going to write. Rightfully so be skeptical if you can auto diagnose Root Cause without with limited manual intervention, right? So classification models and the ability to get those right, you know, are pretty vital here. And then the second part of your question. So what can you do about it if the root cause is costly? I mean, I I tend to think that it’s better to know than not know, regardless of what the cost is. I think that at the end of the day, you have to do, you know, each company has to do their own ROI model on what it’s going to what it’s going to cost to to address it versus to let it ride. And you know, if the costs are are too costly, then, then maybe you leave it, but if it’s something that’s recurring, you know, twice a month, and it’s costing your organization millions of dollars, I you know, I’d have to assume that at that point it’s probably more expensive to to not fix it, right? I think understanding root cause is going to help you uncover the severity, right, financially speaking, the frequency, the liability, to some extent, accountability, who, where is it happening? Who is responsible, or who should be responsible? And I think you know in those instances, if you. Have the answers to those questions. The solution isn’t always, doesn’t necessarily always have to be. I am the sole person responsible for fixing this issue. It may be a shared burden, right? It may be something that’s that’s not caused by your organization at all. It may be something that’s caused entirely by the three PL or the carrier. It may be something that’s that the the customer is doing wrong on their receiving doc or through their process. So I think the ability to pinpoint where it’s happening, when it’s happening, who has custody, and then and then drive a shared resolution through the value chain with your partners is something that you can’t do without having root cause to begin with. So I think there’s a again, here, a net benefit to having root cause and having the optionality to to fix it or to not fix it.

Scott Mears  20:58

I like how you put that. It’s giving making sure that we clarify exactly what that records is what you said. It might not even be that that company that’s using that technology that’s impacting that cause and and I’m with you on that it’s surely it’s better, better to note and than not to know. I mean being just blind to it, just because we feel we may not be able to manage the cost, or may not be able to fix it immediately, due to whatever reasons that that’s not really going to grow grow things forward and move things forward.

Ilya Preston  21:35

Well, and cost is subjective to right, what may not be valuable to You, may may also be extremely valuable to your customer. What may cost you? You know, 50 grand might cost your customer 2 million bucks. So I think the ability to quantify that the true cost all the way through the value chain, as well as any sort of not just the wastage of the product, but cost of replacement, cost of inventory sitting there, and not cost of delay from the sale, cost of lead time of replacement product, like all of those things need to be a part of the equation. And then, you know the I think the decision becomes much more clear.

Scott Mears  22:21

Agreed and moving on, I know we did touch on it, having this level of visibility on the level of risk, and you’re going to be able to rate your carriers and understand who is less risk. How do you feel this could impact the relationships and that worry about the impact this could have with the relationships with the caries, if they start rating them on sort of out of 10 scale of where they’re going to go?

Ilya Preston  22:53

Yeah, that’s a that’s a great question, Scott, and not to scare you off, but we’re also rating the device companies as well. So we, you know, we look at, you know, ping, ping rate consistency, you know, connectivity around the world. And so I think there are a number, number of stakeholders that that go increasingly under the microscope. And obviously, again, this is a, this is intended to be an optimization and decision making support tool for shippers. I think, at the end of the day, whether you’re talking about device companies or carriers and these ratings, I think you’re going to have a sub, you’re going to have two groups within the population. You’re gonna have groups that embrace it and that actually use this information and data to improve themselves and drive change and meaningful improvement. And you’re gonna have groups that reject it. I think in the short term, obviously, you know, it has the potential to to create some strain, you know. And you can think you can equate this to the, I don’t know if you want to call it like the storming phase. And you know, nobody wants their operations to be under a microscope. Nobody wants to shine a spotlight, you know, on what’s happening internally. I think that, you know, sometimes we hear who should own visibility, right? Who should it be the should it be the enterprise? Should it be the shippers that own the product? A lot of some of them come back to us and they say, Well, I want the carrier to own this. I want the carrier to own the devices. I want the carrier to own the intelligence part, and I want them to provide it all for us. And I think, you know, in some cases it may make sense, but in a lot of cases, there needs to be a separation of power right where there, there is, there can be a conflict of interest, if the carriers are owning, you know, 100 Of the visibility and the intelligence that they’re not going to want to shine a spotlight on problem areas, right? So they can pick and choose what they show. And I think in most cases, it makes sense for the shippers to own this, to have the ability to have a, you know, an unbiased view of what’s happening, both from from a device performance standpoint as well as a carrier performance standpoint. I think in the long term, this is going to be great for the industry. We see more and more carriers, shippers, device companies, buying into this universe like this, this ecosystem where everybody pitches in, and it drives benefit for all right, but I think on the to your point, on the flip side, you are going to have companies that push back on this in the short term, I think there’s going to be a fairly apparent contrast between those that embrace it and those that push back on it. The ones that embrace it are going to improve. They’re going to improve their own internal operations. They’re going to offer better service. They’re going to delight their customers, and more business is going to go to them. The ones that that are not as eager to to adopt this, I think, are probably going to see negative business impact in the longer term, and, you know, and that’s going to drive then, I think, a secondary wave of adoption where many of those are then going to say, Okay, fine, we’ll, we’ll do this, right? So I think, as with anything, you’ll have the early adopters, you know, you’ll have the ones that resist it. You’ll have the ones that kind of live in the middle and maybe test it, try it. But I think at the end of the day, this drives such a positive through the value chain for each stakeholder, whether it be the shipper, the customer, the three PL  the carrier, that you can ignore it for long, and you’re going to have to embrace it too and use the data to improve your own operations.

Scott Mears  27:06

Absolutely, it’s coming, and it very much is here, and it’s it’s going to drive, drive the growth of the industry forward, and drivers to adopt and adapt to to be the best carers, be the best shippers. The final line I want on this, on this segment, I want to discuss is, is we, and there was a great panel actually on this. I think it was with Lucy from AstraZeneca. It might have been yourself where we got into a bit of a recruitment conversation and a mindset conversation, especially with the amount of AI learning and machine learning being used, little understanding of the vast company, little understanding of what this is and how it operates, there’s going to be a real redesigning of roles, mindset shift to to be able to now manage this, Mount mountains of data and then take action on it. How? How are you combating that challenge? Because I know that’s a long term thing that’s going to take over time. How do you feel? You’re going to we’re going to see that change over time, over time with this.

Ilya Preston  28:22

So for us, you know, for our company, 30% of our company is data scientists and machine learning engineers. So it’s we were fortunate in that we decided almost from day one, that that is our core competency. Um, where companies I think are, are going to have a tougher time with it. And obviously, you know, there are a lot of benefits with making this your core competency, right? This, this is essentially what we live and breathe every single day. You know, a lot of our collective work and vision centers around the ability to build accurate prediction models, risk quantification models, contextualization models, it’s exciting work. So we’re able to recruit great, great, great people to do this when we look at, you know, our peers in the industry that may have started out, you know, building hardware as their core competency, or may have any other sort of core competency. It’s, it’s, it’s more difficult to, then later on, incorporate data and have that be a part of your core unless you’re already in a sizable enough scale and you have, you know, enough capital behind you where you can make those kinds of investments. So I think, I think it is, it is difficult, both from a recruiting perspective and just a core competency perspective, but it’s something again, that that you have to embrace, right? Because if you don’t, you’re you’re going to be left behind. I think there’s been a major evolution in visibility. Even just over the last two or three years, where we started out just focusing with the.on the map right and putting the.on the map and getting real time location and condition streaming data. Fantastic. Great. Like wow, mind boggling. Everybody is super impressed. And now two, three years later, I think, you know, shippers kind of, you know, they started asking themselves, or they had assumed that once I deploy all of these dots on the map, my problems are virtually going to disappear overnight, no more delayed shipments, no more temperature excursions. And unfortunately, that’s not always the case. I think there’s this realization that, okay, now we have to surround this visibility initiative and program with labor. We have to bring on data, people to interpret what this data is telling us. We have to bring on operations, people to intervene, right and a control tower to intervene with with shipments that go rogue or that go bad. We have to bring on quality people to make sure that all this data is makes sense, right, and everything is in order. So at the end of the day, you have to bring on a full team around this, and data is an integral component of that. So I think you have to embrace it at this point.

Scott Mears  31:22

I agree. It’s, it’s a shift that’s happening, and everyone in supply chain feels that pain of recruitment. It’s, it’s a real pain in it, and it’s one that’s going to continue for some time, and we need to move, and we need to adapt and and I love that that Paxafe has 20% data scientists. I mean, that’s such a great line that you have. I mean, that is such just that line in itself, just shows the value of Paxafe. And if people can work to have to really adopt, bring in the right resource, into into their companies to deal with this shift like Paxafe has that’s that’s really, gonna really move this, this industry and company forward. So I want to now dive into now we’ve gone through the resistance to to technology, this sort of risk management technology. I want to now dive into a bit more about active risk management. I really want to know what are some key challenges imperishable goods industries that passive risk management struggles with today versus active is answering for today. Just so, if you can use some key examples of some big pain points of pressure for good industries that activists answering for today.

Ilya Preston  32:49

Yeah, so I think, for one, you know the the immediate one that jumps to mind is intervention, right? So the ability to have your control tower, or what, again, whatever process you have in place to intervene with shipments that go bad during transit, to have a chance at salvaging the product. Now, I think one of the things that active can can enable is a lot more of the advanced capability that isn’t even mainstream yet, or isn’t even universally available for all, for their shippers, right? So I think first and foremost, not all risk is created equal. When you look at the sea of you know, if you’re a shipper and you have 1000 shipments out today, you might have excursions on 15% of them, right? That’s 150 red dots that you’re going to be looking at on your map. How do you auto prioritize those? Well, so I guess part of the answer today is you have a team that’s sorting through and trying to determine, is this a real risk or not? Is this an important risk? Do I need to have somebody on this right this minute or not, and you have a team that’s that’s analyzing all this data in real time, and that’s great, but you know, one of the things that we are trying to help push towards is this auto prioritization and understanding that not all of these red dots that have Risk are created equal. For example, you know a product that may have 20% stability life remaining, right on shelf life right the equivalent of shelf life remaining, but it’s still 60% of the way from the destination, has a very, very, very different risk profile than a product that has 90% shelf life or stability budget remaining, but and it’s 10% away from the destination that latter one may not require any intervention at all. It’s a low risk. It’s it’s virtually, you know, there’s no risk to the product. The first one on the flip side is a high risk. There is a strong risk to the product. How can we sort through all of that data amongst all of these red dots and bring only those that are truly risky based on your organization’s risk profile to the top? So then you don’t have to have, you know, five people sorting through all this information and doing this manually. I think that’s one of the things that real time enables that passive obviously cannot do. I think another one is prediction, and prediction is obviously still very new to the industry, but we you know, for our customers, we offer two different prediction models. We offer a predictive time of arrival, and we can also predict temperature excursions before they occur. We can’t do that without real time data. Right? Both of these models, as a part of their inputs, take in real time sensor data, and it’s not possible to do that solely based on historical data or passive data. Those are features that are highly valuable that you know if you can, if you can tell your customer that this shipment is going to be late, 24 hours before, so that they can orchestrate and synchronize their internal receiving labor in accordance when with when that product is actually going to arrive. That’s that’s a huge win. Same thing with temperature excursions. If you can identify a temperature excursion hours before it actually happens, it’s much cheaper to them to react to it or proactively address it than to react to it once it’s already hours, you know, pinging and out of spec. So I think these are all things that are enabled with real time data that you we can’t do with passive.

Scott Mears  36:54

I think that’s a key one that you first mentioned, that it’s because you’re right with when you start lighting up your supply chain with that real time sensing and seeing that risk that’s happening, you can get lost within it. It can be a lot of dots, and having a system that can define what are the dots I need to focus on is so important, because otherwise we’re just running around like a bull in a china shop, trying to figure out what which, oh my god, like all this is going wrong. But actually, now there’s reasons to why that’s that’s pinging, and it’s actually these three you need to focus on right now. I think that’s key to what you just said there.

Ilya Preston  37:35

That’s absolutely right, yeah, it’s, it’s otherwise, you know the what I was saying earlier, you know, visibility teams, they have this unrealistic expert expectation from their visibility programs, that once they implement the program, all of their problems are going to disappear. A big portion of their problems are going to disappear, right? And that’s not the case. It’s not going to happen unless you bring on a team around this to, you know, around the visibility program to address these issues. And you know, in our case, we try to automate away as much of the quantification work, right, the risk quantification, the prediction capability, as we can so that you can support, so that these teams can support it. Maybe not with 15 people, but maybe with, you know, five people, yeah, and it becomes, it’s, it’s a much, and those people can utilize the tool to make their lives easier, right? Looking at 150 dots on the map is not the most fun thing to do in the first place they can, you know, they can utilize the tool to make their work more meaningful and more productive.

Scott Mears  38:48

And of course, would you agree that we’re, we’re moving forward further and further to try automatically, like you’re saying that automate more and more to need less and less human involvement?

Ilya Preston  39:01

Yeah, yeah. I think that’s a fair statement. I think, you know, human involvement is always going to be necessary. I think the human involvement is is going to evolve to, like I just said, more meaningful human involvement, so actually making decisions based off of the data, right? I think, I think taking the boring part of diagnosing and quantifying and stacking the risk up against each other from shipment to shipment, those are all things that you typically have to follow stringent SOPs. And it’s not the most exciting stuff. If you already have that information at your fingertips. The human now can make a decision on what action to take. And I think it the work becomes much more meaningful and it becomes different. I think it goes back to your or, you know, original point on these, these jobs around visibility, evolving humans become much more of the decision. Decision maker, and they can make a lot more decisions. The throughput increases, productivity increases. So I won’t say that. I think, you know, this 100% replaces human intervention, but I think that just the nature of the work changes.

Scott Mears  40:15

Yeah, it’s, it’s very much exciting to me, the evolvement this happens, and the the the new things this brings out, and the the moving forward of the industry and company. It’s going to be very interesting to see, and it’s fantastic to be the forefront of that. I want to ask one final question, what? What does risk look like in the future, in supply chain? Is it, are we. Do we eradicate it? Is it just a lot less, or are we just better informed? What do you feel risk looks like in the in the far future?

Ilya Preston  40:53

That’s a great question. I I think if you wake up and you imagine your most perfect day where nothing goes wrong and no issues come up, and you’re just, you just are able to be as productive as possible. That’s what I envision, is all of these. It’s not that these issues will disappear and won’t be there. It’s that they will be, to a great extent, automated, and will enable to the human to focus on the important work, right, and that is to to make decisions. I think risk won’t necessarily be eradicated. I think it will be quantified to where companies understand what risk is impacting their lanes, you know. And not just you know, that’s a very broad statement, but actually being able to break out lanes by, by routes, right by, by legs, of routes, by individual shipments within legs, and understand which features which inputs are impacting different parts of their supply chain. So building hotspots around, you know, on this leg, this is, this is the greatest input that’s contributing to my risk, whereas on this leg, it’s a completely different input, and on this leg it’s the same input, right? So, what are the, what are the commonalities between these two legs? I think being able to answer those questions is the first step understanding deeply where your risk lies and how it’s classified and what’s contributing to it is the first step in being able to, you know, automate it away. So I think that’s, that’s, that’s part one, and then being able to generate recommendations from on how to minimize that risk is the is the second component of that so you’ve you’ve identified where the risk lives, what the hot spots are, what’s contributing To that risk, what the So, what factor, right? So now that I know what it is and how to quantify it, how do I eradicate it? That’s not always obvious. I think there’s, there’s a heavy amount of of, I mean, I hate to use the the buzz word, machine learning, but there is a pretty substantial amount of machine learning that has to go into looking at all of the different patterns, the correlations to then help you auto generate what recommendations should be on the table to minimize that risk. And then the human can take those recommendations, you know, weigh each of them and make a decision on which ones makes the most sense for their particular business. So I think in a perfect world, risk will be quantified, it will be known, and to a large extent, you’ll have automated, auto generated recommendations of how to address it, so that way people can can do what they do best and be as productive as possible without having to continuously combat fires and chase things that you know don’t necessarily require the most, that aren’t the most exciting things to do.

Scott Mears  44:20

I mean, for me, that’s, that’s a future I want. It’s very much, in basic terms, very much taking out all that admin and just allowing us to make those top level decisions. There’s a, b, c and d, which decision do we go with? That’s, that’s much more we can focus our energy on being much more productive, and what we’re doing so much better informed. And to be fair, if you’ve ever seen the good place the TV show on Netflix, there’s a bad place in a good place when you die, and in the good place, it’s a perfect day every day. There’s no risk. It’s just a perfect day every day. And they actually become like very numb to happiness. So maybe you always need a bit of risk. A bit of pain in our eyes.

Ilya Preston  45:02

I was trying to think of a movie or a show that I could equate it to that was very utopian. I don’t remember, maybe The Truman Show. Yeah, something, something of the sort. But no, no, nothing popped into my head. I think if you’ve read the book. I don’t know if you had to read the book The Giver when you were younger. I think, I think it was the giver that they grew up in a utopia, right, where everything was, I mean, I everything was perfect, right? Yeah, I, I’m not necessarily saying that that’s what the ideal state of risk is. But I think just this, this world where you don’t have to continuously combat fires and chase down things, and you can really focus on the meaningful aspects of your work, I think, is the desirable end state.

Scott Mears  45:58

Absolutely. Well. Thank you very much. Ilya, it’s great to hear your knowledge on this and see what Paxafe are doing. You’re doing some really fantastic things. And it’s really interesting to hear your two cents on this. And I feel it’s very interesting to the industry to understand this, and technology companies out there to really get more clarification on how they can fight these battles out there.

Ilya Preston  46:24

Thanks for having me, Scott, this was a great experience.

Scott Mears  46:28

If we just do a little wave to the camera, we’ll do a little wave to the camera.

Ilya Preston  46:33

Probably do it with my other hand.

Scott Mears  46:35

Yeah, yeah. You don’t want.

Ilya Preston  46:37

Nursing this shoulder.

Scott Mears  46:42

Thank you very much Ilya, and thank you very much for watching.

Ilya Preston  46:47

Thank you, Scott.

Scott Mears  46:49

Hi, my name is Scott Mears, and I’m one of the hosts of the Supply Chain Tech Podcast with Roambee. On this podcast we talk to supply chain heroes from around the world about everything, ranging from the disruptions related to supply chains, their personal experiences with tracking technologies, strategies to build resilience, and much, much more. We already have some recommended videos for you to the side of me, and if any of this sounds interesting to you, do subscribe to our Youtube channel and hit the bell icon so you don’t miss another Roambee video. I’ll see you next time.