Everything is Logistics
A podcast for the thinkers in freight. Everything is Logistics is hosted by Blythe Brumleve and we're telling the stories behind how your favorite stuff (and people!) get from point A to B.
Industry topics include freight, logistics, transportation, maritime, warehousing, intermodal, and trucking along with the intersection of technology and marketing within the industry.
126k downloads and rated as a top 5% podcast out of all industries and growing. Follow along to stay curious and become a better thinker in freight.
Everything is Logistics
Shipium’s Jason Murray on Delivery Promises, Multi-Carrier Playbooks, and Agentic AI
Jason Murray—CEO & co-founder of Shipium and 19-year Amazon vet—joins Everything Is Logistics to unpack what really moves the needle for shippers: accurate delivery promises, multi-carrier execution, and where AI agents add real value today.
We get into the “coordination layer” most retailers are missing across OMS/WMS/TMS, building a digital twin of your network, and why faster can actually be cheaper when your method mix and lanes are modeled correctly. Favorite line: Shipium’s role is a “super-sophisticated calculator” for decisions the human brain (and spreadsheets) simply can’t keep up with.
Key takeaways
- The coordination gap is costly. Most enterprises make siloed decisions in OMS/WMS; the win is a horizontal optimizer that weighs transportation, inventory, and cost together.
- Promise precision beats blanket speed. Customers want confidence in “by Thursday,” not generic “2-day.” Model ship dates and probability of arrival—and constantly backtest your predictions.
- Digital twins drive smarter choices. Shipium models lanes, cutoffs, and costs in real time to choose carrier/method, where to ship from, and what to promise.
- AI agents = productivity multipliers. Early wins: a “what happened to this shipment?” chat workflow, and agents watching network signals (the old Amazon “little red button”) to trigger reroutes.
- Money on the table. Method optimization and delivery-promise installs have driven multi-million-dollar annual savings for large retailers.
LINKS:
Feedback? Ideas for a future episode? Shoot us a text here to let us know.
-----------------------------------------
THANK YOU TO OUR SPONSORS!
Are you experienced in freight sales or already an independent freight agent? Listen to our Freight Agent Trenches interviews powered by SPI Logistics to hear from the company's agents on how they took the entrepreneurial leap.
Shipium is the ML-powered shipping platform built by the executives that created Amazon's supply chain technology. They help 3PL & retail leaders like Stord, Ryder, and Saks Off 5th to reduce shipping costs by an average of 12% while improving on-time delivery.
CargoRex is the logistics industry’s go-to search platform—connecting you with the right tools, services, events, and creators to explore, discover, and evolve.
Digital Dispatch manages and maximizes your #1 sales tool with a website that establishes trust and builds rock-solid relationships with your leads and customers.
You add all those paper cuts up, and a lot of what people are doing in these companies are, they're just paper cutting themselves to death, right? It's like, you go look at an analyst screen. They've got 47 Excel spreadsheets open, right? And they're they're moving things around. They're calling people to try to do things. And so if you take that, if they've got 100 steps they're doing per day, and you're able to automate 10 of them, right? I mean that that opens up like this, this massive time, this area for them to start thinking more strategically about where they need to go.
Blythe Brumleve:Welcome into another episode of everything is logistics, a podcast for the thinkers in rate, I am your host, Blythe Milligan, and we are proudly presented by SPI logistics, and we've got another fantastic episode for y'all today, because we are talking to Jason Murray, CEO and co founder of Shipium, and a 19 year Amazon veteran. We're going to be breaking down the delivery promises that convert smarter multi carrier playbooks and where AI actually moves the needle. So Jason,
Jason Murray:welcome to the show. Well, thanks for
Blythe Brumleve:having me here. I'm excited absolutely now, I was listening to a bunch of different podcasts that you were on. Friend of the show, Harshita, with the E comm logistics podcast. She's fantastic episode that you did with her. It was really great at providing some some background, as well as your website, which is website is really well done. We're going to bring that up a little bit later with some of the content that's available on there. But I'm curious as to you know, with your 19 year Amazon history, is that level of, I guess, the today, chaos that we find ourselves in or manage, chaos of E commerce, logistics. Did that exist back then, or is it just kind of on steroids now?
Jason Murray:I think it's, I think it's a good question. I think I think the story with my time at Amazon is probably more the arc and how everything just kind of continues to head in the same DIRECT direction, you know, and it's like every incremental step, and then there's obviously, like these periods of time where it accelerates, right? But I think the note, the notion, or the high level meta narrative, is just that everything kind of has gotten more complicated, right? So there's just, you know, you go back in time to 1990 1990s and you think about, like, what was supply chain, and what was retail? And it was like, Okay, so the pinnacle was probably, like a super Walmart, right? Which is two or 300,000 SKUs. You know, it's all kind of managed through distribution, etc. And now you think about like just as as Amazon followed this arc my career. Follow this arc is just more SKUs, more channels, more carriers, more more kind of ways that people are discovering things, more expectations that are, they're different variable, and it's like, all of those variables push everything, and then you end up on the back end with just like they multiply against each other. So complexity just gets worse and worse and worse. And so, you know, I think, I think the thing that's interesting now is, like, you hear people like, kind of wanting this old time where everything was stable, and they reminisce about, like these, you know, we just have these containers come in and then we sell them in the store and how easy that was. And it's just like, that isn't the reality. And so you've got now all of this complexity that people have to kind of digest, and it's combined with, like exogenous events and and, like accelerating technology. You know, you take the last three or four years and you've got a combination of the pandemic followed by, kind of this onslaught of AI, followed by kind of, like, geopolitical stuff that's causing things to happen. And I just think that's the new norm. I don't see it. I don't see it going away. I can't predict what the next thing's gonna be, but there's gonna be next thing.
Blythe Brumleve:And that is really a catalyst of your, not only career trajectory, but when you started shipping, I believe it was 2019 and then, I mean, everybody knows what happened in 2020 and then the years that followed. And so it really feels like you did. You hit the market at like the the perfect time with, with launching shippium, that, you know, there were all of these issues, these greater issues, that people needed to solve for that level of complexity. And now, you know, you have a solution for that complexity. Did you how did you plan for something like that?
Jason Murray:I can't say that it feels like I hit the market at the perfect time, you know, like, often on the day to day life. But, yeah, I think, I think the need, you know, we're, we're obviously, like any other business where we're skating to, you know, what the customer wants, right? Like, we're moving towards that. And I, I think the whole point of, like, why did we build this, this technology stack in Amazon, this. To solve these problems, right? Why do we do it? And it was really like, okay, you can use data data science, machine learning, coordination, all of this stuff to basically solve these flexibility problems. And I think the need for that just continues to grow, and we're feeding off of that as a company. And so I think in a lot of ways it was, it was good in the sense that the things had kind of reached this point that the demand is a lot clearer, like I you know, if you, if you go back 10 years, it would be much, much harder to sell what we're trying to sell to retailers or or manufacturers, or whoever else, because the the pain was not as acute in that sense, you know. So I think then you kind of go back three or four years, and when we're really started kind of going to market and, and a lot of these factors had kind of driven home the need for technology as a as a way to solve these problems, which I think we discovered at our Amazon a little bit earlier because of the scale that Amazon was running, and kind of how far ahead they were. In in more of these diversified, complex supply chains that Amazon's kind of known for. So so the answer is yes, and I just said it in like 1000 words, which
Blythe Brumleve:is perfect for a podcast interview. You gotta be able to allow, you know, shoot them up and you knock it down. So I am curious around the sort of light bulb moment that you know you work so long at Amazon, what was that moment like to think of? Okay, I want to do this, but I want to do this on my own, and I want to start a company and solve these problems. What did that moment in that journey look like,
Jason Murray:well, some of it is, is kind of personal, right? Like, I'd love to say that the narrative was, was like I had a, you know, I woke up in the middle of the night and had this idea. And it that that maybe plays well in some sense. But, you know, I mean, the thing was about my personal career was just that, when I got out of school in the 90s, I I wanted to be at a startup, and I went to a startup, and then soon after, went to Amazon, because I was, like, really liked what they were doing. Like, I really liked the idea I used Amazon. I was, I was, I was an early adopter of Amazon, and I was buying, like, technical books, and it was the only place I could get them. And I just the concept of it coming to my door was just such a it was such a magical experience for me. And so I had this, like, personal connection to it, and that's why I ultimately went there. But at the time I started there, it was a startup, right? I mean, it was, it was not it obviously, like the the timeframes in that period of time or in that era were much shorter, so they'd already iPod and stuff like that. But, but it like, effectively, it was a total mess there. When I got there in 99 it was like total chaos and and you had a lot of room to grow. There was invention was happening all over the place. The people, like the mix of people, was just amazing, like tons of talent. And so that whole energy system was, was, was just this kind of perfect combination. And then obviously, like over the years, you you, you keep building, and this thing kind of grows, and you end up in this. For me personally, it was kind of like a frog boiling problem, right where you the company is getting bigger and bigger. Your your day to day is becoming more about politics. You're you're less hands on, you're less building, you're less connected to kind of what's happening. And that's, that's a, you know, it's what's required if you're going to, if you're going to move up to the VP or the executive level, you have to kind of move to that state. And so I think if I look back on my Amazon career, I got exposure the kind of 2008 to 2016 period when Amazon really started to scale. It was a fascinating and amazing journey. And I think even more there we were, kind of, you could kind of feel you were part of something like there was a there was this notion of that happening, and so seeing that level of scale and this stuff was really a great experience. So I don't want to take anything away from that. But once, you know, once you kind of get out of that, and then you kind of start looking around like, what I want to do next, and what, what am I, you know, I've got only so many more of these runs before i i effectively am retire, or whatever, you know, like, it's going to be harder and harder to do, right? What do I want to do with this? Right? And I think for me, this was, this was kind of what drove it was I, I wanted to, I wanted to get back to building, you know, and that's, that's kind of how I ended up in this situation. I will say that when I left Amazon, I didn't have a totally baked plan, right? Definitely, this was something that I left Amazon, and then I explored the space a bit and and, you know, I kind of started with, maybe I'll be a CTO or something. But ultimately, kind of got to, like, I, you know, I really want to take, I want to go through this process. I want to, I want to go, I want to be in a three or four person team and see what we can do, and go through this journey and, and there's no guarantee that that is the finance was, financially the right decision. But, like, personally, I. I just love these experiences and building so much that, that that's going to trump everything else that, that I'm kind of talking about. So I it, for me, it was a very personal journey. It was more that got me here. But I'm so glad that I went through this, because I do think, you know, it's like I finally to go from totally bottom nothing to something, right? Is a zero to one, the one to 10, the 10 to 100 process is, is a is a amazing thing to go through, if you have the opportunity.
Blythe Brumleve:Now, did you know right away what you wanted shippium to focus on, or was it kind of a product of iteration? No,
Jason Murray:we definitely went through the product market fit phase. We we had a theory that the original kind of hypothesis was, well, first of all, at the time, like Flexport had a lot of energy, and so the the thought was, is there even, like, you know, should we be even looking, are we looking in the right spot, right? And then we kind of eventually were like, well, let's, let's just follow what we did at Amazon. That's the like, focus on the the end customer experience and, and, but the original theory was we really go hard on delivery promise, and although that is still part of our offering, and I actually think it's starting to pick up some steam Now, the reality was, we went into the pandemic, and people were not really concerned about conversion lift, because they already were selling more than they could handle. So we kind of pivoted to more back end optimization, which is, is not necessarily outside of what the road map looked like or the overall vision. It all fits in that, in that narrative, but the the kind of order you do things, and how you approach it definitely changed, right? And, and, and also, just like, what is the market you're you ultimately have to come to a product that people won't buy, and, and, and you're, you're, you're trying to understand, like, where are the acute pain points that you can focus on, right? And I think we got to this back end optimization piece just was this kind of gaping hole, the deeper we dug and and so getting into, like, transportation Lane optimization, at the at the especially, kind of really focused on, on parcel and diverse networks. There was just a lot of opportunity based on on kind of what we were hearing from our customers, what they were dealing with, they what they were struggling with, and that's, that's kind of how the path look. And so now I think we've, we've the kind of go to market, or how we've gone through this, it ended up being way more focused on, let's, let's, like, solve that, and then kind of build off of that, as opposed to, as opposed to top down that we originally thought so. So to your question, it's changed significantly, like, the way that we the way that we thought we were going to do just was totally different than what actually happened in reality. Yeah, I
Blythe Brumleve:think that's that's a journey that's similar. For a lot of people who are building, you have to, you have this hypothesis, you take it to market, and you hope the market responds positively, but you, more or less, you're going to have to make some iterations to it in order to build something that people want, and they're going to be a repeat buyer of and and so with that, with that all said, for folks who may not be aware, what is sort of the high level overview of what shipping provides.
Jason Murray:So we have a shipping, transportation supply chain optimization platform, right? And what we've been hyper focused on is we have a kind of transportation piece which, which is really focused on, like, if I want to deliver to a customer by a date, how do you do that most possible in the most efficient way possible? So we'll choose carriers, methods, etc. It's all very real time. We model out transportation lanes for all the carriers that we're working with using machine learning, etc. We have a internal cost model to kind of create a digital twin of what, what it's going to cost you, what your you know, what hours your building runs, what are the cutoff times, etc. So we're modeling all of that, and that lets us make a decision. So when you take that and you can roll up to, like, what we call fulfillment engine, which is basically this notion of of where do you ship from. So this is, this is a thing where you're looking at making a decision on how, how you should, where you should ship it from within your network, right? And then you take that same set of data, and you build up to our delivery promise product, which is, what are you telling the customer in terms of expectation?
Blythe Brumleve:And so for a lot of especially with your history at Amazon, you know, they pioneered the two day shipping movement. And so for a lot of folks, they that's kind of the default expectation now is getting that two day shipping. But obviously that's unrealistic for a majority of retailers. Or is, is it realistic? You know, if you optimize properly,
Jason Murray:I think you I think that what happens is it's about it like it's about kind of specificity. And part of the problem is, if you this is where the scenario where your technology is limited in some way, you can't provide the level of detail they need to provide your customer so, so what, to some degree, we're offering is almost. Is almost like, given your network, there are scenarios where, if it's nearby, I can cheap if, if the you know you're in Florida, if the fulfillment center is in Tennessee, I can probably get it to you in two days, very cheaply, right? And so are you communicating to that, that to your customer? Are you taking credit for that? Do you have the tools to kind of measure what is the resin revenue lift from actually making that decision, versus what it's going to actually cost you incrementally, right? And are you offering that to the customer in, you know, in a systematic way, right? And so all of that is like, again, just kind of back to this, this high level, you know, pattern that you it's like we have, you have all this data, you have customer data, you have transportation network data, you have intention, you have you have business goals, and trying to mash that together in some way that that meets the needs of your business is really what we're doing. We're optimizing all of those factors to say, make the right decision for your business. So I like, interestingly, as we've spread out into some of these other industries, like B to B, you know, parks, distribution, or B to B, pharma, or something like that, right? The needs are often different for different reasons, right? They might be more concerned about risk versus cost, but the approach is the same. It's like, I'm going to take huge amounts of data from all these different sources and make a systematic decision on how to actually solve that problem, as opposed to kind of, you know, the, you know, just imagine, like, I'm gonna put some rules in a spreadsheet and mostly hits it, and not all the time, and, and as a customer, you're like, Hey, I'm like, right next to the distribution I can see the distribution center from my window. Why does it take? Why does it take? Why is it saying it's gonna take six days, right? So it's just all of that, and I think that's the technology gap that we're trying to really bridge, right with the with the product, right there just is not you can't get you cannot, you can't possibly, with humans, solve this problem systematically,
Blythe Brumleve:yeah, and I think for the end users, it's not necessarily about, can you get it to me within two days? I think it's more, when can I have an accurate delivery date as soon as
Jason Murray:possible? Yeah, you have the classic. Like, I got the party on Saturday. I need new dress shoes, right? Like it's Monday, so as long as it gets there by Thursday. I'm fine, but I need to know that.
Blythe Brumleve:How complex is that? I mean, I'm sure it's insanely complex in order for a retailer to be able to promise that, is it realistic for, you know, a lot of retail outside of the the Walmarts and the Amazons of the world, is it realistic to be able to provide those estimated delivery dates? Or is that all just dependent on the all of the supply chain?
Jason Murray:No, I mean, this is all just totally doable. I you know, it's not, I wouldn't say it's easy, but, but we, we understand, you know, if you're going to even a smaller retailer, let's imagine you had two buildings, one in one in Jersey and one in Reno or something, right? I mean, we're going to understand this is when you will if, if an order happens. Now, this is when you will ship it, and this is the the probability that it will arrive by a given date, right? And that's all kind of modeled in the system based on understanding of what your networks doing, our understanding of what the transportation looks like, right? It's all a prediction, obviously. So, so you have to kind of constantly be tracking how well did you do against your predictions? Does it? Is it within the tolerance of the process, right? I mean, Amazon itself did not we. There was never an expectation that we were going to hit 100% of the promises, right? It's just, it's a, it's a any physical process is going to have have, like, statistical variation. It's just we, but we can tell you very precisely, like, what is that, right? And, and are you willing to live with it? And can you, are you able to, are you okay with the kind of negative consequences of the spin versus not? So,
Blythe Brumleve:so it's really, I mean, it's, it's not simple, because, as you're talking about, you know, bringing in all of these different data sets, yeah, exactly. I almost, you know, my heart starts having, you know, palpitations a little bit, because it's like, oh, that's a lot of data. That's a lot of, you know, room for error. And so I, you know, as, yeah, thinking about this through, what does that process look like, of just making sure that the customers that are coming in have the data to be able to make those actionable decisions.
Jason Murray:I mean, it ends up being, you know, again, like, oh, so we're going to talk about AI a little bit, but this is kind of why I'm so excited about this. I think you've got validation and and inspection processes that we currently do to kind of solve this, but I think we're on the brink now of these, and they continually speed up as you as you automate these steps, right? Like, how do we read in a contract from someone? Or, how do we, how do we, you know, how do we then verify that the costs are correct after that contract with, you know, is, is ingested? Or, how do we verify our speed? Like, we're kind of back, constantly back, testing our prediction models versus, versus what actually happened? Because we also have the tracking data of. Of the actuals right to model that. So all of that is, all of that has to happen. I think what's so exciting about kind of the AI movement is we should be able to speed this up, right? And if you think about what's AI, really good at, it's, it's, it's really good at taking information and kind of distilling it down to something meaningful. And so it fits this problem just like a glove, and that's that's why we're so excited about this as a future, as a future thing, and whether that's more of an online process that's bringing in new information that we just previously couldn't have, couldn't have possibly fathom getting access to, because it can, it can, you know, dissect and categorize that stuff automatically, or it's just taking existing processes that took too long, and making them, making them 1,000x faster, right? Both of those things are are incredibly powerful in terms of increasing the automation, increasing the flexibility, increasing the kind of ability of your business to run at the way you want it to run. So and
Blythe Brumleve:So for, you know, let's definitely dive into the AI side of things. Because I think when we mentioned AI, you know, there's a kind of, there's a couple different reactions to it. Some people might roll their eyes, of course, but then there's the other perception of, well, I could just, you know, use ask chatgpt, and it should be able to tell me. But there's a an additional layer of complexity, especially at the enterprise level, where you're dealing with all of these different data sets. And so where is, I guess, AI having the most impact in those sectors relative to shipping?
Jason Murray:Well, I think it's, it's super early days, obviously. And I think, I think if you go through like, kind of the diffusion lifecycle, which is common all these, like, new technology things we I think when chatgpt kind of first came out, I think it was like late 22 or early 23 something something in that time period, 2022 I think, yeah, so you started to see so there's obviously, like, Everyone's like, this is this is totally magical. We got to do something. So all the boards call the CEOs. The CEOs call their CTOs. The CTOs tell their team to just put something in place. And, you know, it most of it doesn't work. I think there's been several studies that have come out. It's not surprising. Like, this is what it's just like, let's throw some spaghetti against the wall and see what happens. And I think from the standpoint of someone like me, or someone like a COO, or somebody who's evaluating the stuff like you also have seen the like, Blockchain push in the middle, in the 2010s right where you're like, This is going to ensure that no containers ever get lost and they never went anywhere. Or metaverses are going to run, and are, I mean, the the goggles people are gonna have goggles on fulfillment centers, productivity is gonna triple, right? So, you know, false promises from kind of envision technology versus what actually occurred. And so I think, I think, though, what happens, though, is, you get these personally, at a very you know, as we're going through the cycle, you start to get these proof points of how it's going to work, right? And for me, I think the thing that that starts to bring it home right is if you look at how the software developers are using it, right? And this becomes, as these proof points start building, it becomes easier and easier vision, like, where is this going in terms of our industry, right? And, I mean, I programmed in the 90s, right? And it was this gnarly process with lots of like testing and manual this. And, you know, everything took forever. And it was, it was, you like, you know, to some degree, we were all kind of like these artist wizards who, who were trying to figure all this out, right? And you now, I watch my developers, right? And they're they've got one agent running building tests as as they're writing code. Across on the other side, there's two agents writing code, and they're competing through has better results, right? And the move, the the production of that is, is, is like, the only thing I compare it to, is like when we switched from assembly language to assembly language to C or, you know what, like a really low level computer language to actually having a language at all. And this is that's this kind of jump, right? So the reason I'm going in depth on this topic is because you're starting to actually see how this is going to work, right, and and I so then you kind of like become more as this matures, as the cycle matures, and the use cases start kind of filling in, it becomes very clear of what's what's kind of happening, and where this is going. And I am, like, 100% convinced that this is going to fundamentally change the industry. And I mean, you know, if you kind of like play this forward and say, like 2028 you imagine like a COO or a cseo has now a team of analysts Right? Like they have a bunch of people in the structure, but there's, there's analysts doing things and making sure that operations run right. They're checking buildings. They're talking to GMs. They're they're looking for events, extraneous events. They're looking, they're looking at all these things. And then they're, they're adjusting knobs or adjusting levers or changing the way things are moving. And if you think about 2028, like the way to think back from that is there's going to be, we're going to have, now, a fleet of autonomous agents that are running and checking all of this. Right? This is, this is all of these tasks that were like relegated to humans to deal with all this craziness, because they were the only ones that could process at the end of the day, this weird form of information, like all these different forms of information, it's just such a perfect use case for this is eventually going to be replaced with AI, because it's gonna be able to do it faster, cheaper, etc, etc, etc, right? So you know that that's what I'm working back from. And now we're in the, now we're in the phase of actually kind of building, what does this next, next round of, kind of more AI focused tools that fit this particular narrative,
Blythe Brumleve:right? Yeah, because I keep telling people that it's, it's, you know, the MIT study, quote, unquote, study that just came out recently, which was very flawed study. That's, but it's, obviously, it's, it's getting headlines across social media, you know, I don't, won't go into the it's
Jason Murray:not, it's not surprising, like it, you know, it's like I said, I think you, when you do first, this first, kind of, like, you just, this was kind of, like the first phase of that, the the people's rollouts of the stuff. And now we're getting to, like, this next level. Next level. But anyway, keep going. Well,
Blythe Brumleve:it's a new information era, and that's what I keep trying to tell people, is that it's we are in a new information era, and then we can take action based on where the AI is going to fit the most in certain specific sectors of our life, where it's, it's, it's, it's machine learning, but with an additional intelligence component on on top of it. And so for for a lot of those folks, you know, they probably just didn't realize how bad their data set was, and that's why they can't really make it too actionable. But I do think that there's been some some error in some of the messaging, especially around agents, where they're just going to, you know, come in and solve that problem. You know, as a podcaster, I, you know, I'll be able to get an agent, and they'll be able to take care of all of the post production and editing, and maybe that will happen in by 2028 hopefully, you know, some of those things will be able to happen. But it's more, you know, using it as your calculator, and using it where it makes the most sense. So I'm curious with your your early deployments with AI and, you know, especially around the agent side of things, where have you seen those early winners of where it makes the most sense for retailers and enterprise shippers
Jason Murray:to use? Yeah. I mean, I think it is kind of, it does end up being this, like operator and multiplier, that's, that's at least the current flavor of what, what's working now. And so I think you have, like these mic micro productivity things that that kind of add up to time savings in aggregate. And then you have, then you have, like, larger scale kind of, you know, automation, automation type efforts that help, right? And I, I think, to your machine learning point, the the, the we kind of think of it as, that's something that it's probably not going to actually go it's not, it's, it's more or less building on top of it, because, you know, these, these models that we've spent all these years building and of kind of understanding this data. It's more about now you have this set of tools on top of that that can leverage them. So, you know, so, for example, we have a simulation product that's kind of taking advantage of our models to actually run, run through like, here's a hypothetical scenario. And so to for our and for our kind of folks at our company to actually run the simulations they have to set up, each individually, build out the experiment. And so you can imagine, this becomes, effectively the the the AI is taking in what you're trying to accomplish, and then it's firing off lots of simulations, right? And I think if you even look at like, what if you ever you know, if you're kind of a curious person, and you're using chatgpt, for example, ask it to show its work, right? What you'll see is that it's doing stuff like, if you say, like, Okay, I'm gonna upload an image, can you pull the text out of this image? Right? What you'll see is that it's actually doing stuff like firing off Python scripts. That's of stuff that does OCR. So it's not, you know, it doesn't solve it down to the granular level. What it allows you to do is come orchestrate lots of things. But I think specific examples on the micro level, you know, we have this inevitably, if people are shipping with us, they're asking questions all the time about, like, why did this happen? Right? And so the some of that can be handled by reporting, but, but like in general, there's gnarly ones that come in, and the way that process would work, traditionally is they submit a ticket, an engineer gets to it takes a look like tries to figure out what's going on, reports back, they go back and forth over the ticket, and then some resolution is agreed upon, right? And so we're launching, one of our first things we're launching is just a. Like what happened to the shipment, and that's a that's a case where this is something that someone is potentially spending time on this throughout the day, because inspection is part of any healthy business, but now they're able to just do this in a chat interface to speed up their time, right? And that that naturally works into the workflow. Another example, though, might be we have a product I mentioned at the beginning. We have a product that decides where to ship from. Well, Amazon, you had this, like you have, we had this thing called little red button and big red button where you would basically hit these buttons if it wasn't really a button, but we, we, you know, basically you would push the little red button if a fulfillment center was having issues, right? And it would basically then reroute orders around that. But the presumption, the assumption there, is that someone is sitting there watching for, what are the signals that indicate that this fulfillment center is having a problem, right? It might be the GM is selling you this, or might be that there's weather events starting to happen. There might be like people not logging in. And so this is, this is an example of you have an agent that can basically be watching for these scenarios all the time, and then telling other agent, you know, telling the planning agent, I need to then go and you need to, you need to reroute from this building, because we're having, we're struggling to keep up, right? So these are, these are something that we're kind of in the pilot exploratory phase, these larger things. So it has to be a combo of both. But I will say one more thing, just there is this problem. When I go back to the dev example, you go back to the developer example I talked about a second ago. You go back to like how people are thinking about this. You even think about like the internal DNA of a company, the workflows do change, and that's where some of the resistance comes from, right? People are not wanting to change because it inherently changes the workflow Right? Like we built in software development, the we had these, these stage workflows, because this is realistically what humans can handle. You know, like, this is how you did it, when you have these, this kind of AI notion, and you can write tests live as this thing is going out. All of the plot, the processes you had traditionally don't work anymore. So we've, we've entirely, we're starting to rebuild our entire dev process right to match what this is more going to look like in this AI native agenda. And I think that's going to be the scenario with with supply chains also, and logistics, right? And so we're going to have to kind of fundamentally rethink how you run the teams, what metrics you look at, like, how do you think about who's doing, what, how you organize? And that's all like coming right? It's going to be part of this transition that that I think the main thing is you have to be in this kind of experimental or trial mindset. And so if I'm giving people feedback, it's usually more about culture that needs to shift within these companies, less about, less about like, like a specific thing to do, because it's it's going to keep changing in three month increments, right? Like, what worked, what didn't work now is going to work in three months. So you've got to be in this mindset of, kind of, constantly absorbing and changing and moving, yeah,
Blythe Brumleve:because I think you hit the nail on the head. Because so many people have, are that are, you know, sort of, I'm a tech optimist, but there are so many tech pessimists, and they, you know, they try, AI wants, and it's like, oh, it sucks. It failed. It's worthless. It's overhyped. You know, and that creates this resistance to change, that resistance to being adaptable. And there's, you know, lots of companies out there that, you know, Coinbase, I think their CEO recently just said, like, if you're not going to use AI, you won't have a job here. And I think it's more of a philosophy standpoint that you have to, you have to be able to be able to use these tools. And it sounds like this is, you know, something that you've implemented within your company, and then now can be able to help other companies, you know, adapt to those same, similar philosophies.
Jason Murray:That's exactly right. You have to culturally shift. And we, we have internal shipping them. And I think there's, there's going to be folks who are kind of not on board with that, or they disagree with my assessment of where the future is going. And, you know, I think great, like, you know, I don't have any we're, you know, that isn't it. We'll, we'll ultimately find out whether I'm wrong or overshooting or whatever. But you have to kind of have everyone marching towards the same drum as we go into this and and it does require kind of a wine culture. So I think, I think that totally makes sense. I mean, culture, culture is really important in the sense of, like, who is going to be successful there or not there at your organization, right? It's, it's true of any organization. And so I think you, you, you know we want our we want our folks to be extremely curious. We want them to be trying things. I'm implementing a process internal to my directs, for example, where I want them to install a new tool every month, right? And just to actually feel what it feels like to implement this thing, because there are pains, right? And you don't, you know, it doesn't really do anyone any good. To say, Let's AI stuff, right? You have to, you have to kind of get to that next level. My my general concern, my feeling is, though, when you start trying things, you find things pretty quickly, like, it's not it you, maybe you will have a different experience. But the more I use AI, the more bullish I become on this. And this is kind of where my 220 2028 stance comes from. I don't, I don't use it, and go like, Ah, it's kind of dying off, like going back to say, the blockchain example from from a couple minutes ago, right? It you never, you never got momentum off of that example, because you couldn't find use cases that were fundamentally changed, right? And this one, though, the more you use it, the more you will because it bleeds into it, like we're now starting to use it more and more in our sales process, or SDRs, or our 10x what they could do, 10x maybe 100x what they could do three years ago, right? And so it you, every aspect of the business just continues to as you, as you establish that, that usage pattern and DNA, it's very powerful.
Blythe Brumleve:And so as you're you're building out, you know, your own internal processes and reworking them. And it's probably a process of ripping things out of what used to work and putting things in, and probably doing it several times over, and reiterating, as you go with the with the customers, that you're onboarding. So I would love to get into some of what the onboarding process looks like whenever a new customer decides to do business with you, what do those early days look like? And what does maybe some of the early AI agents that they could take advantage of in their own data sets?
Jason Murray:I think that we are the two couple places we're leaning into. So I would say the onboarding process itself is going to be pretty worked out, reworked over the next six months. So I would say, like attrition. What we've done is like, give us your contracts, give us your details. Let's go to the site and it's it ends up being, you know, let's say more, like weeks, right? But we feel like this is really has the possibility of being days or hours, and if you can bring information in that fast, and maybe it's like, in the virtual sense, maybe it's like 80% correct or 90% correct, but it's still so. So that allows us to effectively bring your data in very quickly and then even prove some points for you early, right? Which is, which is magical. And that's, that's like, you kind of have to part of the same. You know, everybody who's in this space is getting pitched by vendors who are going to save them 10x 5x 3x 12x right? Whatever. The thing is that they're getting hammered by that. I mean, if you are a CSCO, you are getting blasted by vendors all the time making all sorts of promises. And if you added them all up, your business would basically be free. But obviously that that isn't true, and there's and I think therefore people are health police are, in a healthy sense, skeptical, and so our ability to kind of bring your data into our system, that's a huge factor, right? So because then we can prove it quicker, we can move along those axes, I think, I think then you have kind of incremental improvements to platform which I was talking platform, which I was talking about, the question the Q and A thing, right? This is just like this makes their day to day life. We already believe we made their life simpler by being in a console and and kind of having a more modern approach to this. But you throw in the you throw in the question and answer piece that's even that like takes you up to the net stocks, right? And then I think our first kind of, the thing that we feel like is probably the most interesting early is the simulation brought you know, we're effectively creating a digital twin of your network. And where we're getting a lot of the interest is, how are we able to run these simulations using a combination of humans and AI to basically do something that would have taken, you know, weeks to set up and analyze and take that down to days now, right? And so that's that's really where we're leaning in on on some kind of first pass type game changing stuff,
Blythe Brumleve:yeah. And for folks who are just tuning in or maybe even listening, what I have open right now is a screenshot from one of your onboarding videos that's on the website. You know, talked earlier about how I love your website, and I love this video because it just, you know, we could be talking we're talking about, you know, complex things for for a lot of folks, and it's tough to visualize it, but this graphic is perfect at visualizing it, talking about your systems, where it's the website, the OMS, the WMS, the TMS, the ERP, the planning, and then bringing it all together under the shipping platform. And so with a lot of your stuff, with a lot of the things that you've been saying during this conversation, it really centers back to what you call the shipping operating system. So could you kind of explain that where shipping comes into play, bringing in all of this, these different data
Jason Murray:sets. Yeah. I mean, this is really back to kind of even, you know, pre, like AI, just back to kind of fundamental philosophies. There's kind of two things that really separated Amazon's approach to supply chain. After I got out of after I got out of Amazon, I was able. To kind of see how the world ran, right? And it, I think when we were in it, we were just kind of doing it, and we weren't really thinking about what it meant holistically, right? But if you step back from my time at Amazon, there was two big things. The first was kind of what you're alluding to here, which is the, what I call, like this coordination effect. And so you put what typically will happen in enterprises is they will have these vertical slices, right? So the order management system is responsible for everything that's an order, and so therefore it's going to make the decision about where to ship from, right? So the magic at Amazon, we had a team big, there's probably 1000 people on that team that did order management stuff, and we had a big team that did warehouse management stuff, right? Like it was all there was these huge teams that were focused entirely on those problem spaces. And really, what I was doing at Amazon was I lived in this middle area where we were coordinating these decisions. And to use, I heard you say the word calculator a bit ago. And really, to some degree, we're acting as this super sophisticated calculator that could make these decisions systematically. And it really, we really stayed in our lane. I had no interest. I like, there's a bunch of things that OMS do that I have no interest in doing. I'm not, I don't, you know, there's, it's a hard problem. It's complicated. There's rules with all stuff going on. But what we're trying to do is very good at if you make a decision to ship from a given building, there's transportation implications, right? There's inventory implications, there's there's like cost structure implications that unless you are factoring in all these things that we have access to horizontally, you're not going to be able to consider you're you're going to make the wrong decision inherently. And you know, sometimes you'll make the right decision, but if you're making it wrong, let's just say 10% of the time that's that's on the orders of millions of dollars for even a medium sized enterprise, right? So that's the first piece, this kind of coordination thing. And then I think the second piece that was just really magical was, was I took over the supply chain tech stack around 2010, 2009, and that was right about the time that kind of machine learning was starting to take hold, right? And so we had, of course, like traditional operation research scientists on our team and and, you know, but the thing that we did a really good job of is getting them to play together such that we started using these modern techniques for our forecasting, for our predictions, for our our modeling, transportation lanes, all the stuff that's kind of required to make these decisions and and so that is a that is an effort, where these data sets tend to span all of these different stages, right? And so you take those two things together, and you kind of end up with this operating system that the way we're viewing it is, we do not look again. I don't want to be in the WMS space. I don't be in the OMS space. What I want to be is, is this optimization data layer that lives, that lives, kind of below all those that helps you make really good decisions about transportation, inventory, etc. So, and
Blythe Brumleve:for a lot of I just the the industry as a whole, or just retailers as a whole supply chain feels like there's, you know, they have a seat at the the executive table, the boardroom table now, and so in your eyes, with using shipium, they're able to those supply chain managers or executives can now sit down and be able to see if I pull this lever, if I Make this, this switch, x, y, z, is going to happen, and I'll be able to show up to these meetings better informed about some of the future forecastings That that we're going to be making using this dashboard of all of these different levers that we could be pulling. Is, is that kind of the idealistic way that someone is going to sit at a computer,
Jason Murray:and also a great AI use case, right? I mean, like, like, bringing all this together, but it is. It's exactly that. I mean, you have, if you're making a decision about about what you're doing in your network, or how you're doing it, it has implications on customer experience. It has implications on cost, which, which basically just means that you have, you have money you're potentially losing on the conversion and sales side, or gaining, right? And you have cost structure implications on the back end, and there is this kind of ideal trade off, right? Like, it's not a it really is a math problem. And I think we're trying to bring that whole, like holistic thought process to bear. We want people to understand these trade offs and make the right decision for their business.
Blythe Brumleve:So it sounds like it's, it's very reminiscent of maybe some of your, your early days, where it's like, you, I think you mentioned you were kind of these, like, creative wizards, and it maybe sounds like that's exactly what you're doing on this side of the thing.
Jason Murray:Yeah, now it's, now it's coming in the well, I mean, that's the classic thing, right? You figure out the low, you know, you just kind of keep no I think in those days, what I was talking about was just kind of like the mechanics of actually programming required all this and and like the bot, the bottom level programming is these tools have gotten so good they other things help too along the way cloud and. And SAS and APIs and all the other stuff that's kind of happened over the last 30 years, but, um, but that kind of gets more taken away. And so then you, then you end up in this, like, dynamics of the system. What do you do with your data? How do you, how do you kind of use this new set of tools? So obviously, yeah, it's a new puzzle at a different level. But totally, totally like, you know, it's, it's a it get, like, I just, I love these puzzles. This is what I look for, like figuring out how to, how to solve these, how to get value out of this stuff, how to and I love that it has this kind of, like, physical world feel and touch to it. So
Blythe Brumleve:now you talked about some of those early wins that are happening internally at your company. But what about, you know, maybe some of your customers have they had a chance to sort of pilot some of these agents, maybe get some early wins. What does that process look like for from a rollout perspective,
Jason Murray:I think we're kind of in the getting feedback phase, and so the appetite is definitely there. I think, I think there's a, there's a, you know, people are, people are looking for and I think when you demo this stuff, they're there, it's like, very clear to them, there's a there's a win. I think it a lot of it ends up being a little, you know, we've been so as a company focused on kind of the optimization side of it. So just like, what's the cost savings, right? And and so now we're getting into this new territory of like, it's more about automation and productivity and flexibility, which are a little squishier, but the excitement level at these companies, and we show them, like, Okay, what if I was able to, you know, here's this chat bot demo that we're starting to roll out that is going to tell you the state of the shipment. What would that do? And they're like, I mean, this is like, you know, I can, I can now free myself up to be focused on these strategic issues, as opposed to as a piece of grinding away, on, on, on, trying to get through these small tasks, right? And that's so, that's, that's the stuff. That's where it's, like, very experimental as we figure this stuff out, but, but incredible. I just, I think the like, you add all those paper cuts up. And a lot of what people are doing in these companies are, they're just paper cutting themselves to death, right? It's like, you go look at an analyst screen. They've got 47 Excel spreadsheets open, right? And they're, they're moving things around. They're calling people and trying to do things. And so if you take that, if they've got 100 steps they're doing per day, and you're able to automate 10 of them, right? I mean that that opens up like this, this massive time, this area for them to start thinking more strategically about where they need to go, which are, frankly, just things that probably need to happen at the company, whether it's a renegotiation of a contract or following up on things or driving some change within a build, all that stuff needs to happen, and there isn't time for that, because there's just all of these minuscule things around information flow within supply chains. Is
Blythe Brumleve:there anything that's coming down the pipeline? Maybe in the next you know, you mentioned 2028 in reverse engineering, where you think we're going to be about in 2028 but I'm curious as to what is sort of those present day wins that that shipping provides. I read a couple different case studies you know that you had on your your website. I think it was peak readiness with Manhattan WMS, Saks off Fifth Avenue, moved from single carrier to multi carrier fast, and saw that faster is cheaper. Which which of those. I guess levers can be pulled today for some of your customers that that, or maybe even prospects that that are listening to this episode.
Jason Murray:Well, I think I've been trying to kind of keep the conversation mostly focused on kind of the brand new AI stuff, but, yeah, certainly, certainly with, with, within our ecosystem, like we're selling, you know, we open up a whole new carrier network. We are able to unilaterally, almost every scenario save a bunch of money by method optimization. We've got customers who are now using our fulfillment engine to more precisely make decisions where we have delivery promise installations that basically help customers say, you know, tell their customer this is precisely when you will receive it. And so all of this is, all of this is in the mix. And, you know, I apologize for I maybe went too far on the other side, but this is stuff that we that the company has been focused on over the last four years, as we've kind of gone to market with this and built out all this and built out all this infrastructure, and so that, that all that all continues to move along. I think a lot of discussion today, and it's, you know, we pride ourselves in the fact that they almost always have some sort of financial benefit attached to them, like it's, it's always about, it's always about. After we've installed this, we've, you know, we, we saved a big retailer on the order of $17 million right, like annualized, right this year. And so by, by optimizing how they're running their methods and and kind of eliminating having a very, very precise model of what their cost structure is compared to their carriers, we're able to wring out all of this money out of their current spend. So that that's, that's and, you know, from their standpoint, it's all in, almost entirely automated. They don't have to do anything this money. It's just, it's just free on the books, works, etc, right? So all of that is going extremely well, and is, is part of the, you know, continues to be what we build upon. And I, like I kind of said, I think very early in this episode, really, we think of this new AI stuff as accelerants for that. So how do you get to that stuff faster, right? Like, how do you, how do you, if you want to run scenarios against our simulations? Like, instead of having to set it up, work with our customer success, etc, can you just basically interact with a chat bot that gets this thing going? And that's the, that's kind of the, it's back to your point of, like, if I'm sitting at the table, I'm the cseo, and I want to, I want to talk to people about about what are some of these scenarios. I want to be able to turn those around fast, because you can't be at the table and say, we're thinking about implementing this customer experience piece, right? You can't be, well, okay, in four months, I'll be able to give you the impact of that, right? It has to be very fast. It has to be accurate, or you're going to get grilled on the back end. And so that is all of this, of this notion of, like, moving faster, flexibility, autonomous, right? And that's, that's what we're kind of building off of. This is new AI stuff,
Blythe Brumleve:yeah. I mean, it sounds like you're, freeing up from an overall philosophy. So you're getting all the data in to be able to make smarter decisions and make those smarter decisions faster, but then also adding the agentic AI model on top of that, to be able to get rid of the mundane, so those same folks that are sitting at the table then be able to focus on the strategy, the money side of things, anytime you can get more profit back, that is, you're doing your part as a, you know, as an executive. So
Jason Murray:yeah, and I think culturally, the even the process of bringing the data in that's going to be dramatically improved as we get better and better at this and that's that's part of this kind of, like cultural transformation of how we think about all these problems, like we, we solve problems a certain way three or four years ago, that now you just fundamentally solve them differently. It's just, it's part of the game.
Blythe Brumleve:Yeah, I used to tell people that I would and I would do this, that I would hand transcribe a podcast episode, and now I don't have to think about that
Jason Murray:and that that's
Blythe Brumleve:incredible. Tell you how many hours in the day that that saves,
Jason Murray:yeah, yeah. And it is really amazing how much better these things are at that. I just, it's, you know, there's obviously been, like, dictation software for years, right? But it's just, it's like, the the the the level and the quality, like, you know, language translation and all this, you know, super minor thing. But we, I remember early at Amazon, we were launched when we launched a new country, like the late the localization process. It took probably six months right to get through the the kind of us to Germany translation, right? And it's like, you know, we, when we launched our kind of the the our Spanish version of our console for a Mexico installation we did, you know, it was like, just, it just happened, right? I mean, it's just really incredible how these things just speed up certain parts of the information flow that were just so difficult,
Blythe Brumleve:yeah, well said, because that's just, you know, there's all these different aspects of whenever you're implementing new technology and making these iterations that you you can't see until you try and so not a try once, and then, you know, throw it in the trash can and never read it again. It's, you know, let's think about how creatively we can solve these big problems, because big problems probably also have big costs associated with them, and every business is looking to save money at all levels. So Jason, I know we've, we've talked about, you know, a lot of different aspects within this episode, but I'm curious where you kind of think the, you know, obviously the holiday season is approaching, you know, where e commerce isn't going anywhere. It obviously it's a mainstay here. Where do you kind of see, you know, the next, I guess, six to 12 months shaping up, and how can retailers and and shippers be best prepared for that?
Jason Murray:Yeah, I mean, I think, I think you're, I love your kind of quick summary there. I mean, I think obviously we have all of our peak readiness stuff going on right now. Everybody's been, been, been, you know, the tariff thing that's kind of dominated the the that part of the narrative all year has been, has been floating around in the background, and I think is mostly settled at this point. I think we, we saw that, you know, our shippers, our customers, were mostly upset about the unpredictability of it, less, though, than the actual kind of what it specifically was. But I think you take all that, and I think, you know, I just hate to be a broken record, but I just think 2026, is going to be, kind of, you're going to start to see the acceleration of of these AI installations and and. And a lot of companies, probably a lot of people, are kind of in similar situation that we are, is that we are just frantically building for the next six months to try to kind of get ready for this onslaught, right? Because it the the the kind of hunger for it, the the attitude towards it, the whole it's very, very apparent in the market. It shows up in even the stock market, right? Obviously, the best top stocks are, like, I think, I think that's, I think 2026, is going to be largely about that story. And if I, if I'm going to make a I think, you know, the the interesting, like, kind of curveball stuff, is, how does it translate to physical and when so like, self driving trucks, self driving cars, when does that stuff all start to hook up also? But I think that's a mess. That's that's harder to predict.
Blythe Brumleve:Yeah, it's technology. It's if you're, if you're anti technology is going to be a bad time for you. Yeah,
Jason Murray:yeah. I think, you know, use also ask specifically what I would do. I'm advocating cultural changes to the way we think about this stuff, right? And I think you have to push in your org for this stuff to be, to be front and center, and kind of start leaning in towards this, because I think otherwise you're kind of getting, you're going to get caught flat
Blythe Brumleve:footed, yeah. Well, well said. Now, is there anything that you feel is important to mention that we haven't already talked about? No,
Jason Murray:I feel like, I feel like we've covered everything we've been pretty sure, honestly.
Blythe Brumleve:I mean, I had a whole list of questions, but it just that the conversation went in a great direction. And I love diving deep into this, the AI agent, you know, sort of world. So this is this lined up perfectly. So
Jason Murray:my marketing guy will yell at me. He's always, he always gives me these things I'm supposed to do, and then I end up talking about, you know, whatever I'm doing, all this stuff. But anyway, well,
Blythe Brumleve:I think we hit all the important stuff. You know, it's AI, it's technology, it's data, and then it's, how can folks actually use this to make real impact within their company? So, alright, Jason, where can folks, you know, follow you, follow your more of your work, maybe get signed up for shipping demo.
Jason Murray:Yeah, so shipping.com, obviously you can request a demo there, LinkedIn. Jason Murray, shipping them. CEO, should go find me. You know, always, I think LinkedIn is probably our most accurate social media platform. So, so anything on that front, we try to keep pretty up to date on what we're posting and doing. But we'd love to talk to anyone who wants to talk. So, great,
Blythe Brumleve:awesome. Well, this was a fascinating conversation. So, so thank you so much for for joining us, and we'll have to have you back on in the in the future, whenever you know the AI agents are, you know what taking hold, and you know how we can kind of take advantage of
Jason Murray:them. Thank you so much.
Blythe Brumleve:Thanks for tuning in to another episode of everything is logistics, where we talk all things supply chain for the thinkers in freight, if you liked this episode, there's plenty more where that came from. Be sure to follow or subscribe on your favorite podcast app so you never miss a conversation. The show is also available in video format over on YouTube, just by searching everything is logistics, and if you're working in freight logistics or supply chain marketing, check out my company, digital dispatch, we help you build smarter websites and marketing systems that actually drive results, not just vanity metrics. Additionally, if you're trying to find the right freight tech tools or partners without getting buried in buzzwords, head on over to cargorex.io where we're building the largest database of logistics services and solutions. All the links you need are in the show notes. I'll catch you in the Next episode in Go jags.
Unknown:You you.
