Hey, everybody. It's Patrick Escolas with another Banyan Technology’s Tire Tracks podcast. Today, we're doing the impact of AI and freight procurement part of our mini-series, and with me today is Matt Harding, CTO of Greenscreens. Yes, he is part of the same Bald Barber Club that we all go to. No, but Matt, thanks so much for being here today.

Yes, it's great to come back.

As you can see, you're coming straight from the matrix. So, you are talking to the AI right now. Are we going full on agent Smith? What's happening back there? 

I can actually explain this. This is the visualization that we did, so it's really hard to explain what your systems are doing to lay folk, right? They have the ROI and the value. What this is trying to do is really get across a couple of things. Each one of those dots is a model that's been released into production. Each line would be an individual customer at Greenscreens. So, you can see the darker shade there, that was the early years. This one's about six months old, so there's probably 30, 40 customers that aren't up there, but each dot represents a trained and tested data curation, all of the model validation. Then each dot is also the source of information, we have about 6,000 monthly right now. They're all feeding off that dot, within an organization. 

Right. That makes sense.

And all the APIs that were connected to all of our partners and all of the TMSs. So, this is really sort of just visualizing the production of, the factory floor, if you will, for all that we create.

And you are in base talking to a layperson. So, I did all represent something, what is it telling us within that the pattern or the color matchup thing?

Yes, it's telling us a couple of things. One is that the deployment is sort of asynchronous. It's kind of a constant development, constant cleaning. It's highly automated. We like to say machine learning at scale. So, it's not so much about the model. It's about kind of the perishability of data and staying fresh.

Oh, okay, because there's a half-life to every piece of data you've put in, all right.

Yes, pretty much. So, you constantly have to grind through the data, make sure it's high quality as you're building these models, and then you need all that connective technology on the other end to get it where it needs to be so people can be better at their work.

All right, it will prevent me from asking you how much do you close costs in the matrix, that's a grandma's boy reference for anybody watching, but with that, so we're talking about freight procurement and AI. And that's really the theme of the miniseries we're doing. Greenscreens is spot on. I think if I was to Google those words, I think I would get to Greenscreens as well. So, we'll start with 30,000 feet. What is Greenscreens doing today? Because we have spoke before to support freight procurement or free procurement with AI right now for the market?

Yes, our core solution starts with, we'll connect to brokers primarily. We're doing pilot work with shippers, that's ongoing, but our primary customer is a broker. As you know, you can have a telephone as your primary technology as a broker, one guy, dialing for diesel, or you can have an organization with thousands of people in it.

And everywhere in between.

Yes, so everywhere in between, we're part of that journey. We start with getting the information, we start with cleaning and validating it, and then making sure the models are accurate for our customers. So, at the highest level, we are, think of it as a sort of a clearinghouse for all buying and selling within a broker so they can have access to data. Then we create models that reflect the market. We focus on, gosh, we've got drive and reefer, flatbed. We do some multi-stop, real close on releasing hot shot.

That would be awesome.

So, we do a lot of pricing intelligence to help brokers with their buy rates. Then you can think just logically, what's the next extension to that? It's the understanding your win ratios through your quoting, all the business intelligence that you could put on these data sets that helps some of those sort of mid-market brokers that don't quite have the IT they need, or even smaller ones. And then think about all the tech that you would need for an enterprise, large enterprise where they're looking at APIs and connectivity and they've got some other things that they're doing, maybe even their own modeling or things that they're looking at within their business.

And I was just going to say, that makes sense at the larger scale. It gives you an idea of kind of what's going on at the other end of the building almost and where that's playing into the overall. Now, you talked about hotshot coming along and I'm going to stick there for a second because I know that a big piece of it is kind of the modeling and the predictive analysis. How does that work for something that says, spontaneous or as hotshot to where you don't know that you need it almost even six hours before you're ripping through the phone and trying to find somebody?

Yes. Well, the hotshot that we're looking at are some of these fast, fast loads, maybe not as big loads. Just standard hotshots. The way the modeling works is that like anything, if you work with brokers and you're taking binary for flatbed to start, they're like, “Hey, what about range? What about this? What about that?"

We all want more. It's always just, yes, you got that, or how do I give you everything and not make it my problem anymore?

Right, right. So, we have machine learning talent, if you will, that in addition to building out the core solutions and bringing that scalability, they're also testing by ability, kind of an R&D skunkworks approach to sort of say, “Okay, let's take this idea. We've got really great relationships with lots of brokers.” So, if it's a common problem, we usually work very closely with the handful. We'll go to the whiteboard and start sketching it out and measuring the results. And then that's kind of how the ideas unfold and are developed and then it becomes product. – 

Kind of speaking to this and AI is, I guess technically still relatively new, but within ever since it came out, everybody's been talking about how you use that towards freight. Have we, almost that, what comes first, the chicken or the egg, has AI in the marketplace started changing some of those models in the predictive analysis of what you're actually going to see? Because before, you're basing it on all of these data points and you still are, but now you also have to assume that other players are using AI to kind of model as well. Is there almost an AI versus AI component to it at a certain point? Or am I just thinking too much war games?

No, I think the question is, there's a lot of different factors that I think companies tie into. We've seen a lot of like this ChatGPT/LLM type of modeling.

So, I think that makes me sound professional.

And then there's these sorts of analytical models. On the early stages, they're kind hallucinating, supply chain people. There's usually a lot of accountability in something that, so that hallucination is something that has to evolve. So, if you're trying to look for a framework or like, give me all these different ways to approach a problem, it's very good with the language. I think from the, if you look at those two as separate types of models, I mean, we're really taking this section of AI that's called machine learning to predict a rate. So, things that factor in to that are, you're going to look at seasonality factors, you're going to look at speed of data, you're going to look at how do you distribute this to the industry. At some point, there's just a source of market information that people are going to use.

Where it gets into the AI versus the AI is if you think about parts of the supply chain that have historically had difficulty connecting. I go back to like shippers being able to forecast their volumes accurately or take a sales forecast two months in the future, figure out where you're going to have issues with your lanes. I think that's the area where some of these technologies will come together and really kind of solve problems more systemically just based on difficulties and units of measure and something a database can't handle because the calculations can't, they don't tie together well.

I might be really well suited to bridge areas in the supply chain. I also too, think it begs the question of what types of data can you start requesting as part of this? Because right now we're part of this TMS revolution that's a part of the broker space. So, there's lots of options and there's lots of information that's fairly common across these organizations. But as they start asking more interesting questions, how to get better accuracy or better information through that system, we just look at that as input for the R&D that we would need to go and kind of suss out a solution.

So, I think that using more data is better, having higher quality data with a broader level of visibility within the TMS space or other systems, and then linking that to other processes a little bit better.

I mean, that's always the goal with technology in this space. I think, A, I want to touch on a few things you said, TMS revolution, I'm thinking Viva La TMS. I love that idea. And then kind of going away from the hypothetical back in the machine versus machine, how is what Greenscreen does with their brokers, how is it affecting these brokers today? Are we seeing a and a 10%? Obviously, you've got ROI for success, but you've been doing it long enough now that what is the true trend of a broker that doesn't have, isn't working with Greenscreens, starts working with Greenscreens, and here's where they are today.

Yes, just a more efficient booking. I think, speed decoding. There's cases where companies have information, it's embedded in the ML, and then they can sort of auto quote on that. So, giving the ability to know what a good rate is and just cutting out the time to find a better one. That's one area. I also think two, that if you look at more traditional brokerages, less tech heavy brokerages, there's typically a handful of individuals that really know rates. So, if one day they decide they want to work for a different place, you certainly don't want to – 

Or retire.

Yes, right. Exactly.

Whatever the job, I went fishing. Who has the pricing? John has the pricing, still. Oh, crap.

And then the other thing too, is retaining that information, obviously, and like we said, speed to quote, but also with all of the integrations that we've done. I mean, our integration team is phenomenal. We work with so many different partners. Our APIs go all over the place. So, if you're really trying to minimize the clicks and the time it takes to do your job because if you look at most brokers, they've got 5, 6, 10, dozen tabs up for different types of applications. We're trying to get that information where it needs to be so that the person can get the best experience to get the work done. So yes, higher productivity, more alignment to the market, in terms of what's happening on a load basis, but then even on the market basis, when you go to that next level and just want to know what's happening with rates, what's happening me versus my peers. 

We just launched a new product called Illuminate. It's a BI solution, but what it does is it breaks down the market in different views, whether it's geography, spot or contract. So, brokers have visibility to – what's happening to their margins? How does that compare to the rest of the world? How are they buying? How are they selling?

So, it's kind of an information system relative to everything related to price. We even have individuals that they're logging in and they're doing certain things. And you may have very successful employees that have something working for them and you'd want to be able to kind of map measure that, map it, make sure it's part of a process for the overall health of the business.

Well, I would like to look at the Illuminate. I'm sure that looks pretty cool, especially just as from an analytical nerd point of view too. As you talked about, Margin is within the broker game as everything. I know a bit about how Greenscreens get started. And one of those ways is kind of to go to the historical data of a broker. Now, if they're not using AI in the past, or maybe even not that technological of savvy, is there a lot of dirty data that goes into it? And how does someone like Greenscreens just take the data that's relevant and good to use for that rate generation or at least prediction when who knows, you're looking three years back and like we talked about the guy retiring, they could leave, what about everybody that left there and I don't have a story for it, who am I going to go talk to about why that happened in this way? It's just a number and a date.

Yes. One of the approaches I think that we have that makes us unique is we're very hands-on up front. So, we have a full and customer success team, individual designated to a customer. We get the question sometimes, well, how does that scale? They have a lot of customers assigned to them. 

More success made.

A lot of customers assigned to each individual and they work extremely hard. But I think it's a value prop because when you think about, when you're working with a broker and you've got a data template and you've got a process or ingesting information, we're going to take that information and we're going to put it in the process. One of the beautiful things about ML is that it tells you where the accuracy is. We're very transparent from day one till whatever decades these customers will be working with. So, the whole relationship starts off with we believe we know your data is clean, but it's not, so let's go on to figure out what's going on.

If it's not clean, it's not your fault. We're not mad at you and we're not insulting you.

We'll find the exceptions. There's always exceptions and we'll find ways to codify that. But more importantly, after that pre-work is done and we get things trained and everything looks great, and we go live, we'll start billing. Our threshold is 10% MAPE. We have other metrics we use on the ML team to measure the accuracy, that's just one of them, but MAPE's just the mean absolute percent error.

I was just saying, for all those people that are listening and don't know, I definitely knew that one, yes.

Yes, absolute percent error. So, if you're above $100, below $100, that's $100. It's a thousand-dollar load, that's 10% MAPE. So, we measure to that, but the other thing to know too is it's probably like I would say 50%, 60% of the data is below 5% MAPE, right? So, it's one of those distributions in statistical terms, like that's the worst, that's the average.

So, we're looking at those outliers from a MAPE perspective. We're looking at outliers from a business process and data perspective, and then we're giving customers the ability to report on their accuracy as part of their solution. We're very transparent about that. And more important, this is sort of behind the scenes, but we have push reporting and other types of reporting that will tell us, “Hey, there's some part of this network that doesn't look right.” Let's say something comes in down the road, somebody's keying in data, and it's creating a problem. We have ways of knowing that there's something happening instantly.

So, we're constantly tracking all the data that's coming in, even after the onboarding to make sure that when people are trying to quote, you get the work done, they don't have to worry about the solution being drifting or having issues with inbound data. So, it's a very, very important part of what we do.

I can see we're having a success manager for each of those ones, because as we both know, this industry is so relationship based, even when you're dealing with a broker with thousands of employees, it's still who you work with it's important. To them, the industry as a whole. One of the questions I have is talking about accuracy. How is that accuracy defined? What is it comparing to? Because if I'm taking five or six different brokers and their data are all saying different things, am I just comparing that to an average of those years to see what falls more accurately in line? And you don't have to spill secret sauces, but just generally, how is the accuracy determined in something like that?

Well, for every load that is booked and makes its way through the TMS, that becomes input data at some point for all runs. So, we're able to attach predictions at the time of the shipment to say what our prediction be based on what you actually paid.

It's kind of, it's back-checking your own predictions. Okay. No, that makes sense.

I mean, we will see cases where, let's say something happens really rapidly in a certain geographic area. We'll tune to that, but reality is at the speed of light, right? So, we do have cases where if there is a big shock and that's – I mean, for us, that's like an open area for innovation and how do we account for those. And so we've done a lot of work to take prior years seasonality. Central Florida, every Mother's Day week, places on fire.

What we're doing is not just taking last week's data or last seven days data or whatever and creating that prediction now.

Yes, because you compare it to that Mother’s Day week.

– from other periods to say, do we get some information from the past that we can verifiably use and improve that as it's occurring, right?

So, with that, we've up to this point, a lot of what we've talked about with what how Greenscreens helps is with that rate or price generation, how does AI or Greenscreen specifically help with capacity type questions? Or is that just that data is inherent within the rate generation and it is going to get from there? Because I know that capacity is always something that people are trying to balance. Where does that work?

Well, we have what's called capacity on-tap. And what we've done is based on the APIs of our partners who do that capacity verification, they're integrated into our UX. So, we look at that as a, it's not part of our core solution. We've sort of, I wouldn't say outsourced, it's just there's other companies that do a really, really great job at that work. So, we just make sure that technically we can bring that capacity look up into the solution. We also have a running history. If you're using our solution and we're trying to look at the rates and the pricing, you're going to be able to see everything that you've done on that lane as well. We do not share – one company will never see another company's data, for obvious reasons. But that you will, all of that information is part of that user experience.

One thing I didn't talk about is some of these integrations that we have are actually functionality bundles that are packaged into and share the same screen space with other applications. So, we do have those APIs, obviously machine to machine, but in cases where we need what we call Java bundles, we'll do JavaScript and actually share some screen space so that you get the functionality and the branding of green screens within your workflows.

I got you.

So, we got a lot of that out there as well.

Yes. Within this, that makes a lot of sense with the capacity and with that. One of the things I want to ask is, with all of this happening at the speed of light, or whatever the speed is, which is faster than someone would have done it manually, are we opening ourselves up to more errors? Or are we actually more correct because the data we have to make that decision is more accurate? Is because it's faster, we might have made the wrong choice from a capacity or a carrier choice, or it's fast, and it's accurate because the data we put in is just clean? Or have we seen sometimes that it leads to more errors because it's too fast that you're not thinking as critically anymore. You've just got it go with it.

Yes. It's interesting because there's the market effect, right? Which is what's happening in the market and then you've got brokers are in between a buyer and a seller and so you have some of those effects that are – we just know the shipper is slow. That kind of issue. Or we have maybe brokers and we've trapped like how much reuse is there with your carriers, right? So, those of us tracking like, are you completely turning through your carriers, one and done? And what percent of your volume is that versus the ones where you have relationships?

So, I think there's a lot of, if you really try to assess out accuracy, what I do is going to be a blend between sort of these committed relationships and these other things that you do for shippers where it's predictable, and then just this mattering of unpredictability. So, I think from the standpoint of us learning a broker's network, our machine learning is going to key into those individual areas within their network. And then for what the general market's doing around, that's what we call the network model, that information's going to be there. It's really kind of getting those two things together.

I would say too, after looking, I did a lot of data analytics and data science work in prior lives all during COVID. And as we all know, COVID was probably the most extreme supply chain pressure I think any of us have seen in our life, and there's sort of a rate of inflation that the market had in it based on the route guides not functioning, lots of spot rates. You can track the rate of inflation with – it's going to be single digit percentage per month, but it's pretty hefty over a year. So, the key about reality being the speed of light is that the data systems and the machine learning at scale has to get as close to that as possible to account for that shift.

So, I think the market is less interesting than what brokers might do individually. I think that's part of the reason why our recipes work for them.

Speaking of recipes that work, how much of it is reactive to the data and proactive to where what we believe or what the model might think is going to happen? Or is it just a combination of the two?

Well, I mean, anytime you use data to predict the future, and we have, we're very close in terms of our long-term release, but we're still ironing out some work on there. But the idea being is that you're using historical data to make a prediction. And then the key is to get some of that variability and some of that, your own history and then your own expectation to make a better bid for long-term relationships. So, that exercise is always going to be less accurate into the future, right?

Right.

If I were to ask you in October, or let's say the first day of November of last year, what's your prediction for 2025? There was one big event in November that – and even where we are now.

Could have swung either way, really.

I mean, check your portfolio today, I don't know. So, we've got a lot of – I don't look at that, I like to go through the day with a smile, so I let someone just do that quarterly. Then I just have one day of either happy or sadness. But I have four kids, so I can’t have money anyways.

It’s better to be busy.

So, my point is that there are always going to be external events that are impossible to predict. But the same point, there’s still is a process and we’re trying to harvest these relationships and those rates of change into something that's useful, so that people can have some sense of where the tea leaves are, three, six months out. If you want to get really janky, go 9 or 12 months out. That's up to you.

That's like that's like picking Super Bowl winners before the season starts at that point. Yes.

Yes. So, I prefer predicting Super Bowl winners around the – 

The day of/ I like the day of. That's where – and even then, I'm generally wrong. So, that makes sense to me that it can do kind of both pieces of it. But the closer you are to where the data you're given is going to be the more accurate it is, which again, that's just kind of common sense. But what if within Greenscreens and its integrations and partnering that up with a TMS, where is that kind of – what is the real benefit of having all of these things integrated? And I know that as we kind of talked about it for a second, sure, I can have less tabs open and click less places, but from a comprehensive, not just for the broker, but maybe for the customer of the broker, what does that look like as far as an advantage using all of these AI and integrated and API tools together and other three letter words or acronyms?

Well, I mean, you think about the dynamics that brokers go through, right? Markets are unpredictable. You have to make hay when the sun shines. And so having those systems and that integration is really about all the things that people need to do are centered around that tech. So, having that tech ironed out, having those processes ironed out, and learning from your competitors to some well, how are these things evolving? How are they working? How should I organize my business around it? More important is a broker if you're really successful. I mean, we had, gosh, one broker was started out with a handful of individuals and they're already doing five million a month, that kind of thing.

So, it's more about the business dynamics or the company dynamics as they're growing to ensure that technologies meet them right where they are based on their scale. Day one, the guy doing carrier sales is the CEO, right? And at some point, you have an analyst in that hole of hierarchy there, and at some point you separate. You've got cradle to grave maybe when you start and then you separate your carrier and shipper sales.

There's this evolution that takes place that I think what the technology is allowing these companies to do is to manage the cycles more effectively. Because if you've got to lay people off, if demand dries up and unfortunately, you've got a downscale, that's a temporary situation. Then you're going to have to bring all those people back or add additional people or maybe some out of college. So, I think that's really the game there is just making sure the tech supports that dynamic workforce.

Yes. That makes a lot of sense with, like you're talking about, if you're getting more people on and the scalability of onboarding someone new to the game instead of them to somehow pick up 30 years of tribal anecdotal knowledge, they can play in the space without that kind of gatekeeping. With that, and what we've talked about with AI here, what else will – what do you see happening as AI is introduced more to the freight space, not just in Greenscreens, but in possibly competitors or within something like the claims aspect or the shipping aspect? What happens when AI is, and it probably already is, everywhere? What does that do to our market as a whole?

Well, I mean, all innovations are driving towards better, do more with less, right? There's that aspect. I do think that the tech space, particularly for logistics and supply chain, there was a late 2010's massive inflow of capital. We're starting to see a little bit of this consolidation starting to affect the tech. So, I think there does seem to be a cycle with some of these innovations where they're not quite proven yet. They've got to pass the test. I mean, people will take chances on them and make those investments and make those changes, but maintaining that business, if it becomes difficult, then there may have been less of the do more with less on the sales pitch of that versus reality.

I think there is something too as AI is evolving. There's no question there's going to be – I mean, I saw Billy Bass giving advice through AI on that thing, right? This guy's talking to his Billy Bass and it's telling him how to get through life. It was incredible – 

And you're not just referring to the take me to the water song that you sing.

No, no. This thing is having a conversation. Unless, it was completely staged. I don't know. But I know you can talk to Grok. You can talk to all these other –

That's crazy out there.

Yes. So, you think about a future – 

I just imagine it being fed financial advice. Like, “Oh, I take all of my ticker knowledge from Billy Bass here.” “Yes, that guy knows what he's doing.”

It's crazy. But I think it's an exciting time and I think that you can't dismiss it anymore. I mean, you could dismiss it before 2020, like, “Oh, AI, everyone's got AI on their name.” But you can't dismiss it anymore. It is real. What that means is that you just got to keep your eyes on where the money's going, who's getting the value. You don't want to be left behind, but it requires a lot of prudence in terms of knowing which things truly work. So, it's going to be a better world. It's just going to be very different.

No, I always have dreams of my AI being smart enough to be like, you have all of my programs and know what it is. I want you to know that I'm calling this person, so I want their screen up here. I wanted the last conversation. We're not there yet. I still have to click many things, but I still dream of it. But within that, where's the future of freight and AI go? And maybe that's with Greenscreens or what they’re next cooking up, or just where's the ceiling? Is there going to be a place where it can only go so far because someone always has to tap the button? Or is there a potential for kind of exponential integration, automation, and someone thinks they need a shipment moved and all of a sudden the chip in their brain tells a truck to come autonomously pick up a load?

Yes, I think, the driver, drivers do a lot of jobs and a lot of value. Anytime I go on a long road trip, I always think about the industry I've worked in and think this is really a rough job. I can't imagine doing this all year long. I think the driver is super important, how that gets augmented is, we'll just have to wait and see. But I do think having a more connected, more efficient use of assets, right? Imagine the price comes down 60%, what does that mean? I mean, ideally that's the future that's there, how soon it comes, we don't know, but it's just going to mean that we're going to be able to have a much more efficiency in our supply chains and maybe less congestion on the roads. Who knows? I mean, it's hard to say, man. I'm just a casual observer here. We're really good at this, but the broader implications, you can get really creative on this. 

Well, yes. I think you bring through a good point with the driver. I don't know if you've seen the movie Logan, but ever since then, I haven't really trusted the autonomous truck driving. And would you, if you've got this giant truck on the highway next to you and you look and you don't see a person in it, I'd get a little nervous. But I'm not the one who understands what the dots and the dashes and green behind you mean. So, you might take a little more confidence in what programming can do there.

Yes, reality is hard. I mean, for somebody who's been in data pretty much my whole career with procurement and market trends and everything else and benchmarking, the digitization that's needed to get the full potential of the things that are coming out from an algorithm perspective is just not there. The digitization and the speed and the digitization has to change because people have to consider a lot of things in their day.

You don't even realize you're doing it.

The number one app in supply chain's email. And the second is Excel. So – 

It is scary how – yes, you say it and there's no denying that fact. But then here we are both, talking about our smart and integrated solutions. But yes, at the end of the day, it's email and Excel that are still winning that game by and large. So, with all of this, it's happened with Greenscreens and AI, and this evolution of automation within the space, is this sustainable to keep moving that way? Or again, if there's not a, we talked a little bit about where the algorithm might cap out, but is there a point where, I don't know, morally or environmentally, someone should take a look and slow down and say, “Hey, maybe you always need a person next to at least one agent of AI running?” And have you thought any of that? And it's okay if you haven't?

No, that's kind of getting into that language model space and that, what do they call it when it becomes self-knowing.

I just call it Skynet. I don’t know if it’s self-aware.

There's some $5 word for AI being a human.

It's Terminator 2 to the rest of us, yes.

It's just hard to say. I mean, Patrick, I've got certain problems that I can see and ones that we address and you think about where that next step is there, but how the world unfolds truly from an AI perspective is going to be interesting. I mean, I don't think people will respond well to risk and to harm. And so, there’ll have to be adults in the room, but some will tell you that we won't have any of that control, right?

You listen to Musk and you listen to others out there talking about the future. From a human standpoint, it has to be better, or why would we do it? I just think that's – 

I think that's a good point. Again, if I didn't ask hard questions, there’d be no point in listening. You make me think every time you say something because I got to say something intelligent back and then come up with a question. That was great. So, as we kind of went to an unknown right there, let's lead with a known.

Okay.

We'll end it with this. What makes Greenscreens so good at what they do? And why should someone not using AI with as a broker right now, is kind of what we're talking about specifically, start looking into it and possibly looking into Greenscreen specifically.

We are a great company. I mean, our people are just top notch. This is one of the highest EQ organizations I've ever worked for. Our customers love us. We have to earn that, right? It's not just a hug fest. We've got to deliver on the promise.

Well, it's brokers. If you're not helping their margin, they don't love any of that. 

Yes, there's like a little head fake and you're gone. I get it. Our technology is proven, it helps the smallest broker. It helps the largest broker. And then we meet our customers where their tech is. So, depending on where your tech stack is, we have a solution if it's minimal. If you've got a lot of tech, we can get really creative in terms of where that goes. And then we just make our data accessible, transparency accessible, lots of little tools built around your key capabilities, I should say, not tools, but capabilities built around the whole life cycle of managing individuals, understanding markets, and just keeping track of your business.

So, yes, we'd love to – if anyone's interested, we'd we love to talk to you. Our salespeople are very eager and very helpful. I get asked by them all the time for the – we just got our SOC 2 certification. So, it's like – 

Oh, congratulations. That’s no easy feat.

We need our policy. We need our disaster recovery on that.

You’ll be spending the next 24 to 72 hours writing that up, I'm sure.

Yes, our team did a great job with that, and it just makes those sales information requests so much easier.

That's awesome.

They do great work, yes.

Well, Matt, thank you so much for your time today. I really appreciate you kind of going into what's going on right now, what's going on with Greenscreens and where freight procurement and AI are melding. And thanks for everybody watching. This has been a part of our mini-series on how AI is impacting freight procurement. Thanks for watching another episode of Tire Tracks with Patrick Escolas, that's me. Matt, thanks again, and thanks to Banyan Technology for letting us do this, and I guess, Viva La TMS. Thanks, Matt, I really appreciate you today.

Thanks, Patrick.

All right, have a great one. If you're listening, watching, follow us, engage, say what you like, say what you didn't like, and let us know, and watch the next one. Thanks, everybody, bye.