AI & Fleets: The Shocking Truth About Your Data!

try. Okay. All >> right. Morning everyone. Uh my name is Yonghu, also commonly known as uh SH. I'm an analyst at Bloomberg New Energy Finance and it is my absolute pleasure to moderate this panel today. Uh so in my day-to-day work, I cover transportation trends and how technology is reshaping mobility and fleet operation. And fleets today, as you know, are dealing with more complexity than ever. rising cost, supply chain volatility, uh shifting customer expectations. So it is not surprising then that AI is increasingly being used uh been integrated into fleet operations to enhance safety, efficiency, resilience, you name it. So to help us unpack this topic, we have today uh Brendon Wiggins from Congress, uh Jamie Boxstrom from Step Energy Services, and Michael Bernice from Motive. Welcome everyone. So, let's start maybe with a

quick round of introductions. Uh, tell us your your title, your role, your company. And for first timers in Nashville like myself, can you recommend one thing to do before I fly out of Nashville tomorrow, please? Brendon, you want to go first? >> This is not in the notes. >> Yeah, I know, right? So, that was that wasn't >> throwing us a curveball. >> Um, yes. I'm I'm Brendan Wiggins. I'm the the vice president of fleet and equipment at Congruix. Um we uh do fiber optic construction. So in in ground and aerial fiber optic construction. Um but just been at Congruix uh maybe two months now to since taking over the fleet there. Um but been an industry guy for a long time. So um in fleet about 15 years. >> Nice. >> Nice. >> I'm Jamie Borstrom. I'm the DOT manager. So, Department of Transportation Manager with Step Energy Services, uh, Canadian division. So, help with the fleet and transportation, uh, helping the field operations. So, I've been with STEP for 15 years, been in the DOT department for like 20 now. So, >> yeah. >> Nice. Nobody's telling you what to do in Nashville. >> I know. >> I know. >> Come on, give me one at least. I skipped it. >> Well, I have one. I'll I'll give him one. Don't worry. Okay. So, hi everyone. I'm Mike. Uh I lead the AI team here at Motive and all those exciting uh product features and things that you heard about in the keynote. My team builds those. So I'm very excited to hear from all of you what is top of mind and to talk to you a little bit about uh what's going on at Motive and SH. There's one thing I know you need to do in Nashville. It's the Rock and Roll History. Oh, no. Country Music History Museum. >> Done on Monday. Next. >> What's that? >> I've done it on Monday. >> You did it already? >> I'm out then. Maybe someone else in the audience has one. >> All right, we can talk. >> We'll take in the Q&A later. >> Yeah. So, so and then Jamie, it sounds like both of your companies are operating in very different environments. Maybe you can walk us

through how exactly you're using AI in your in your day-to-day operations. Just on a high level, >> do you want to first? >> Yeah, you start. >> I think that the the most obvious place we're using it dayto-day is is the mode of cameras, right? is is the safety stack. Um, and how we're able to scale up, you know, a program like that as quickly as we did. Um, really leans on the the the AI portion of the safety cam. And so I would say that's the the biggest place where we're using it right now, but also starting to use a lot of generative AI in all kinds of places across um the fleet organization. So, it's just saving us a lot of time on uh repeat tasks and um and also just when we have big documents that we need to distill or um getting us a head start on creating documentation for other stakeholders inside the organization. Um it's just starting to give my team kind of a leg up uh to to get us started on uh on some of that work that used to take people a long time. and and so I'm very focused on the right people doing the right work and uh and if we can have AI take some of that and then take my folks my subject matter experts and pivot them towards more value added work um anywhere that we're able to do that is is kind of my focus. >> Yeah. So I should say uh step energy services is an oil field company so fracturing now cementing uh coil tubing nitrogen fluid. So um our logistics department is where we use our AI cameras right now and our new uh merger company uh also uses them and uh so yeah before it was just like relying on lagging data right and now we it's helping us with our reports and our um like our fuel consumptions and that and creating reports things we just didn't know before. you know, we have our safety protocols in order, but were they actually being done, right? So, now it just helps us see things like this now and definitely something we want to get more into our fleets to get more data, more reporting and to help our professionals to keep them safe. >> Yeah. And Michael, from M's perspective,

where are you seeing AI create uh the biggest value across today? I mean, when I when I hear from customers and I talk to customers like Jamie and others, um, what we hear is that there's there's a bunch of trends and a lot of that you heard in the keynote this morning. So, companies tell us all the time they have so much data they don't know what to do with it. So, they're data rich but time poor. And so, um, we're trying to help unlock the value of that data. So people who are literally too busy to even analyze it, they can get the benefit of it because it's their data and it it should be something they can uh leverage. Uh we hear about you know the need to reduce manual work across the fleet operations and physical operations economy all all over the place. Um filling out reports and writing documents or um ingesting documents or having workflows that can be automated. So giving people back that time. so that yourmemes can do the work that is most value added for them. Um and then safety right like everybody wants to make their operations as safe as possible and I think with the advances in AI that we see all of us feel it like every day in in our day-to-day life but uh I believe and I know that there are ways we can take that those um huge advances that are happening and bring them to operations that can allow them to be safer to provide like accident prevention or uh safety monitoring other things. And so we're really focused on unlocking that capability for motives customers. >> Right. So Brandon, you mentioned earlier on about work that used to take a lot of time and you Michael just mentioned uh your customers are basically data rich but time pool. So I like to dig deeper into the topic of productivity. Maybe

let's just kick start of discussion in the topic. So Jamie, how how is AI actually reducing manual work for you today? And are you seeing real gains in terms of efficiency or response times? Something that's really measurable across the board? >> Across the board. Um it's I mean for all of us, right? Like it's just that real time data. It's being able to start that conversation sooner with our drivers and stuff. um pulling pulling all this the safety off the AI and uh getting the reports done in that. Um yeah, I mean we went from just ELDs and tracking devices to now having this AI technology and it's just grown and helped us so much. Right. >> Right. And and you Brenda, are you seeing workflows basically move from let's say a reactive problem solving angle to a more proactive approach? Is that happening in your space? >> Yes. And and and I think that there's like you said there's uh things that used to take us a long time manual things. So the everybody wants some, you know, big sexy solution, but just things like when we get an invoice from a third party vendor and we can take a picture of it and it will create a work order in my maintenance system. Those kinds of things save a ton of time where I used to have an admin who would have to go and key all that in. And you know there's a risk that the AI sometimes gets that that transcription wrong, but there was also a chance that the human got it wrong. So we had plenty of data integrity challenges before and I actually think that that's one place that AI is helping us quite a bit is the data integrity. So you said like we're data rich um and that was true but also when you dig into it >> was all of that data useful and and the AI is able to help us kind of distill that as well and go and do some data clean up. We're doing a lot of data integrity work and then also things like scoping documents when we're looking at hey I need to go build a new integration across two different systems in the business. before would require us to get severalmemes in the room and hash that out of what what data is available in that system, what data is available here, what are the API connectors that are available versus we can get a scoping document, the rough cut of a scoping document for a project like that internally done in a couple of hours now and then we're all working on okay, how do we go take the next step? So, it's this force multiplier that's taken some of those complex projects that would have taken us several weeks to scope and collapse that. So now we're having productive conversations almost immediately. immediately. >> Yeah. As it pertains to u productivity,

I'm hearing a lot about the use of edge AI as opposed to cloud AI. Can you clarify the distinction between the two and how they used to reduce manual work? >> Sure. Sure. So I mean as you all know right we have our our devices that are in the vehicles or or on the um tractors wherever and those have compute power in them. they have sensors in them and they have the ability to notify, alert or u speak to the driver or speak to whoever's operating that vehicle. Um and so that we call that the edge AI. That's the the um edge of the system where it's actually on the the kind of the road or the the work site. And you know, if you want to do things that are going to be timely and notify someone before an incident occurs, like instantly or within seconds, there's no time for that to kind of go all the way back to a server that's in a data center and then come back to the device and then warn the driver. That would take too long. So everything that we want to do that's very, you know, instantaneous or reactive has to be on that edge device and running on the actual hardware in the vehicle. And so there's lots of features you heard about, you know, collision prevention and lane swerving, other things that our devices need to process and reconcile on the actual hardware that's in the car. Then there are things that can happen after the fact like AI coaching, right? that doesn't need to be in the moment within seconds. It's typically going to be something that happens after the shift is over. That can happen in the cloud and that's what we call cloud AI. So those are huge big models, very advanced AI systems that can use lots of power and lots of data, but they can't be real- time, reactive, instantaneous. And so the cloud AI complements the edge AI. They work together and they create this unified experience. we use whichever one is most appropriate for the the product that we're trying to to deliver. Um, and you know, I think all of these technologies are offering ways to save time both the edge and cloud-based systems. Uh, you know, things like um processing documents, for example, that doesn't happen have to happen on the edge. It's not a split-second thing. It's like se several minutes fine. Um, and that saves tons of time. And then you saw some of the things where our dash cam in the keynote is able to tell people, hey, you know, you've got a um a good deal on gas coming up in three miles, you should probably stop. And that saves money and time for that driver. So, a combination of all these things together is what makes our platform powerful. The integration of all of it, >> right? Yes. Saving money is just the

perfect segway to our next topic of discussion which is cost optimization. Maybe again for you Michael when evaluating the ROI of AI investments what metrics should senior management look into? >> Yeah, I mean in my mind like money and time they're the the main thing we need to and safety obviously but like when we're talking about you said talk about profit and money. So in that context, I would say time and money are really the things that we need to optimize for and be measuring carefully and accurately. And so um when we launch a new product or when we work with one of you, partner with you, we are obsessive about understanding the workflow, the optimizations that we can provide and how that translates directly into the bottom line for companies like you guys. So um that's something we put a lot of effort into and I would encourage you all to take the same view of hey the hard results are what matter right if this technology is exciting and fun and fancy but it doesn't actually save time or money then it's probably not worth it and so what we want to focus on is those real business metrics we want to move the needle for our customers. >> Yeah. So at the operational level, Brandon, are you seeing any real reduction in expenses for things like maintenance, fuel consumption, asset utilization? >> Uh so I was probably maintenance is the one that's that's we're seeing real results from and and I also don't want to pollute. There's there's machine learning and AI and you know they get conflated often, but we've been doing a lot of that in in the maintenance space for a long time. So getting out ahead of some of these failures and triaging fault codes on which ones are severe and um and starting to prioritize work definitely pays dividends. Um but I think that you know to Michael's point earlier also when we're looking at AI as a tool, I'm going to evaluate it from a business perspective the same way I evaluate any other tool I bring into the garage. Um and it has to meet the same metrics. we have to go, you know, we're going to go look at um technician time and you know, I I have a master diesel tech. I want him turning wrenches as much as possible. Any time he's spending on the laptop or anywhere else in the garage is not really leveraging him as a subject matter expert. So, um really looking at, you know, am I getting time back from from that mechanic? And usually I'm going to go reinvest that time. So, it can be tough to go put a an exact dollar value on what was that worth. So, we're really just looking at am I reducing nonvalue added time from that employee and and increasing his value added time and I view that as as a win and we can kind of pencil what is that worth, >> right? >> right? >> What about uh unexpected disruptions like lost cargo or unplanned downtime? Do you think AI is capable uh in addressing those issues? Anyone? Anyone? >> Lost cargo for sure. Yeah, I mean we have the asset trackers that we provide and like if something is lost, we were talking about it just before $15,000 magic wand. >> Yeah, >> Yeah, >> some you can ask him about it later. >> I lost a couple recently. >> He lost a few. >> Let's not talk. >> So, you know, there's lots of these things that um where you know, a simple solution, but something powerful can unlock a lot of savings and uh recoup lost cost. >> Yeah, that's good to know. Yeah. So we cover productivity and cost optimization. Let's now turn to another key area where AI plays a critical role

and that's safety. Maybe we'll start with you Jamie. Can you perhaps uh give us some insights as to how the AI dash cams have has uh helped you improve safety for your teams and drivers? >> Yeah. So yesterday in one of my interviews I said that the new AI has 15 alerts watching that this morning there's actually 20. Um so that's even better. But yeah just like for our logistics and that that where we use the the dual facing cameras now with the AI technology. It's um it's teaching our drivers that the the cameras are more like a co-driver, right? Like it's just a conscious thing and that's what you have to like get through to them. And so, yeah, just having that um that awareness and I love that it like alerts the drivers first. So, they get a few alerts first and of course you can set everything yourself, right, as a company because we all have our own standards. So, yeah, we all have our safety standards, but we just didn't know before if they were actually being followed, right? And like I said, now we can and we have all these alerts. So the drivers are we we all have a lot of new newer drivers. And so it's teaching them our older drivers that have been driving for 25 to 30 years, getting them out of their bad habits, maybe things that they didn't realize they they did, right? And now with these new alerts in the dash cams, um, yeah, it's just the training that it gives them. And it's not a tattletail thing. It gives them the alerts first before it tells a manager, right? And one thing with the AI is the more the more the supervisors and managers use it, the more it learns our company standards and what we actually want to see, what we make critical, high, medium risk things. So just like any AI, as we're all learning, we have to put in the time and the effort to get the results back to help us with our reporting to help us with the alerts and that. So, >> right, I'm really intrigued though as

the AI system is really getting embedded into the system. Do we run the risk of uh driver distrust of alert fatigue happening out in the field? Is that an issue? You think >> I I can like speak to that a little bit. I think one of the things we are very careful about is that when we launch one of these products or one of these features that we are very um diligent about not having false alerts and you heard about that in the keynote as well that every false alert erodess driver trust >> and can lead to them actually disabling or ignoring some of these things. And so we have a very complex system that processes every event that we detect from in the vehicle on the edge all the way to the cloud like we talked about. And before it reaches a safety manager's dashboard, it's been vetted by all of these different stages. And the idea is that uh by doing that we will prevent that kind of fatigue that you mentioned and people will trust >> the technology more. The last thing we want is for someone to disable a life-saving technology because it's annoying, right? We want to make sure that it it only fires when it should or as much as possible and provides the benefits to them. >> I see in its current form, so it can it can be disabled >> anything. Our platform is fully configurable, right? So, anybody can choose to to do whatever they want. A fleet manager can configure the system how they please. Uh I don't like an individual driver may not be able to do that but at the high level uh fleet manager can configure and disable enable whatever they'd like >> right >> right >> maybe I shut five off that's why I only thought there was 15 turn >> I I I view it as similar to again any kind of new technology we're bringing into the cab for the drivers. So there's the initial push back on the cameras at all. Right? This is and then you know you'll have one big exoneration and the whole driver core is asking like hey when are you installing my camera and uh so you know it's it takes time you have to build that trust and I think AI is going to go the same way right that they'll learn um what once that we're not seeing those false positives as often then I think that driver core will learn to accept it the same way we did with the cameras and other things. I I think honestly the bigger risk is for us maybe at the manager level of trusting it too much too early when it's when it's not ready. Um and and that that's the risk of any technology but um but yeah I think that >> the drivers will will come around once they see the benefits of of us continuing to tune out the noise. >> Right. Yeah. So on the technical side,

how is building AI for physical environments different than say building AI for knowledge work applications since there there's real world consequences? >> Yeah, I mean so you know a lot of the AI we talk about these days is you know kind of I would say low stakes applications you know to like a chatbot or uh trying to decide where to go for dinner. those kinds of things. A mistake there, you know, might lead to a bad meal, but it's not going to be catastrophic for anybody. Uh, in our world, we're dealing with much more high stakes situations. And so, there are several things we need to kind of do differently there. But at the same time, we need to bring to bear the most advanced technology we can because it's such a important problem. But when we develop technology, we have a much higher bar for validation and testing. Uh making sure that we don't just solve for the 80 90% cases, but the 99.99999% cases and we have a very rigorous system to monitor and ensure nothing's going wrong at any time. And the combination of all of that means that we can deliver products fast and safely. And that's really what we aim to do. Um, never sacrificing any element of safety but moving as quickly as we can with our development. >> Yeah. So for our next segment, I'd like

to touch on cross industry impacts. But before I do that, maybe just uh to do a quick survey, show of hands if anyone here is not involved with fleet operations. Hands up. No one. No one. >> Wow. Let the record show nobody raised their hand. >> So, everyone here is involved with fleet operations. Okay, that's cool. So, we can make this really quick because this is just to indulge in my curiosity because I'm in the news and media uh uh sector and I'm really intrigued by everything I've I've uh uncovered thus far through this conference through going through uh the numerous case studies that you have and talking with folks in motive. So in my mind uh the the motive product offering really hinges on integration and automation as mentioned during the keynote this morning and that's something that's applicable across different segment of the industries but I I put up a list that basically look at all the paradigm shifts that are possible with uh AI and automation in general. So here's a list that I've put together. Maybe you guys can complement it. So I only have six on my list. It feels insufficient because typically for a list there should be at least seven to it or 12. >> Seven. >> Yeah, it sounds a lot better. But number one, obviously you got to move uh the system from being fragmented to fully integrated. That's one. Number two, it must be you must move the workflow from one that is manual to fully automated. Number three, the process has to be uh proactive rather than reactive. Number four, you need to be able to measure precise incoming data as opposed to vague subjective uh data. There's the adage that you cannot manage what you don't measure. So >> that falls under this. Number six, as it pertains to administrative work, you need to try to have as much real time processing as possible as opposed to batch or deferred processing. And number six, as opposed to having, you know, a stepbystep uh uh process, you need to have sort of like an end to end agentic process in place. So maybe from your perspective, Michael, is there anything that I should add that's applicable to other, you know, industries out there? >> Well, I think what I've what I've been kind of surprised to find is how similar these problems are across different industries and different applications, even within fleet management. You know, you heard about some of like the waste management applications where we were able to use models which frankly are like very very similar to the models we use in the vehicle, you know, uh, dash cams and we're able to detect overages in waste management. We're able to detect contaminants. We're able to detect missing PPE on work sites. All of this using basically the same six principles that you mentioned shared across all these different applications, use cases and features. And that's been remarkable to me to see. I didn't expect it to be so generalizable, these principles in the way that this works, but really there are so many similarities between these different physical problems in the world. And so what we've been able to do is leverage like basically the same technology to solve all of those problems. and the six items you mentioned, they go into each of those different um applications. So, uh that's been my observation. Very interesting. >> Very good. Jamie, you have anything to add to this? Anything that's transferable to other industries from your experience? it. I mean, working in Canada, we have very harsh weather and that and so I think um Motive has come a long way in their asset trackers and that that it'll work in our minus 40 temperatures up there. Uh yeah, I mean we're we're all going to have something different, right, that we have in our industries or with our fleets or whatever our standards and uh it it's yeah, following those rules, there's just adjustments that can be made and it they're easily like you can easily do it, >> right? I think I've, you know, had the I've been lucky enough to work with a lot of different fleets in my in my career and uh you like everybody's fleet is a little bit different and I always joke that you know you're you're a special snowflake just like everyone else, right? Um and you know trucks are trucks like there's a lot of things that are going to be applicable effectively across the board and then what I do think the AI is going to allow us to do is some of those industry specific niche use cases that we want to go do. um we're going to be able to go dig in on that, right? I want to I want to automate asbuilts for us after we're done with a project and those kinds of things. We can go dig into that, but there's a huge amount of the the groundwork that we're laying right now as an industry that we're going to be able to then go lift and shift and and will give us, you know, a head start on some of these other, you know, kind of more siloed projects. But there's still a ton to ton of work to go do and value to add that's going to be applicable, I think, to basically everybody. >> Mhm. Yeah. But beyond all this hype

surrounding AI and automation, do you think there's a specific area where AI has not yet lived up to its expectation? Well, I mean then again with the the keynote maybe put a pin in some of my original thought of how I was going to answer this question, but um the uh I would say for me at least that one of the pain points that we still deal with a lot is on the logs and and so you know having AI be able to detect when a driver didn't log in or didn't log out or just those kinds of things and and not just having a we're still having to have a person go and and validate logs versus just flagging not just there's something that looks like it's wrong with this log, but here's what I would do to fix it. Um, I think that would be a really great use case that we haven't quite got there yet. But also, I know it was discussed, so um, it's coming. >> All right. Okay. I think we have 15 minutes left. Um, we're saving the, uh, the best for the last. So, the last section here will be on actionable takeaways for everyone in the room. So what what do you guys think is the best

place for folks to start in using AI if they're new to this journey? What would be the first place that they should start? >> Look at me. Okay. Um well, I think what you heard in the keynote, I'm going to go back to that again. I think that Atlas tool that we're la that we've launched is an incredible AI assistant for you all. Uh if you're already on Motive Platform, it's got access to all of your data, your context, everything. uh that'll be a gateway I think into this world and I think from there you can connect it to claude like you saw or or chat GPT or whatever um and then you can start to pull information from everything right the web your data your email your uh your your system so um that is a very compelling place I would start from and I think that will take you on the journey and lead you to new and exciting applications automations things like that which um frankly I think probably I can't even come up with honestly without knowing your your business but you guys should be able to do that and we're giving you now the tools in your hands. >> On the flip side of the coin though, what are some of the mistakes that folks should avoid? >> Well, I think one thing that AI can't do no matter what is take responsibility and I think we can all agree that there is no AI today that can take responsibility if something goes wrong. one of us is going to be on the hook. And so the mistake I see is people not understanding that and maybe trusting too much or not checking or not double-checking things. And I don't think we're in a place where that is we're ready for that today. And so if we're in a environment or an application where you know safety and critical things can occur, uh it's on us to doublech checkck things that come from the AI. And so I think it's a great tool. I'm like obviously 100% AI bought in, but I also want to be honest about the fact that, you know, it's not perfect all of the time. And I think we all have seen that happen. And so we're going to uh be careful when we use it. >> Right. Brendan, Jimmy, anything to add to that? >> No, I agree. We can't just focus on the the false, but also focus on the positives. And Exactly. If if you're going to start integrating more into it, just know it does work with a lot of the programs and systems that we have in place already. Like that's the purpose of the AI is to integrate it all together. And it and it does work and it helps you actually see maybe some of the the flaws you had before too, right? Like it's just that extra tool to help you and but again you have to work with it. So don't Exactly. Don't just rely on it to pick up everything. You still have to do the monitoring. You still have to go in and work with it so that it helps you grow and gives you all the proper data and stuff. So, it's uh it's a growing process, but AI is just it's it's taking over and we just have to all get into it and work with it and just integrate it, >> right? >> right? >> And then the the only thing I'd add is just the back to what I mentioned earlier on the data integrity, right? It's like anything data related, you're garbage in, garbage out in in a way. And um so really owning your data, know what data you're collecting, make sure that it's accurate, have it keyed into all your systems, accurate, the vehicles have all their spec information in there. And you know, the AI can help you clean it up, but but also, you know, just having that good basis of data for it to work with is is going to be critical to your success. >> Yeah. >> Very interesting. Last question for the

day. What do you think the next three to five years is going to look like for fleets? >> Yeah, I mean, I think what what I'm most excited about in the next several years is the the safety side. Frankly, I think the the automations, the workflows, that's that's all great, but really what we're here for is to make sure that the roads get safer and that people get home to their families. And I think what I I mentioned earlier is that there's been incredible advances in AI models even in the last several months, but they haven't translated yet into this application that we are looking at. I think that's going to take a few years, right? And there's a series of challenges. And I think Motive is incredibly well positioned to be a leader in that uh transformation, but it's still going to take some time. And I think over the next several years, what you will see is improved predictive collision avoidance capabilities. So we reduce the number of accidents and we prove that reduce the number of severe accidents. will have better alerting when things are going wrong in the vehicle and automatically notify you or the driver. Um, and things will just become smoother and easier to use. And so that's basically the the trend I'm excited about over the next several years. Um, we'll see what happens. >> Yeah. >> Brandon, any future predictions? >> I I I view it, like we said earlier, as a force multiplier. So, you know, I I tend to I don't want to fight a war on 10 fronts. I'm going to fight a war on two fronts on any given year, right? We're going to go pick our projects and go lean in and get them done. And I think that maybe that'll allow us to do three. I think um so I I I think our goals are going to stay the same, but we're just going to have uh a lot of new tools u to react to the change that's going to be coming. >> Totally. Anything Jamie? Do you want to go live? >> No, I just agree like I I want to get motive more into into our fleet and just um Exactly. the safety side that's that's huge for us and uh yeah, just helping the drivers. So, >> okay, very good. Thank you. So, we're going to open it up now to the audience.

So, if you have any questions, uh we're going to bring the mic to you. So questions anyone? We have 10 minutes uh left for the session. Okay, we have one here. You are talking about predicting the trajectory for example when you identify something coming in the way of a truck. Uh I understand that you are making different scenarios. Would you be able in the future also to react a couple of to predict a couple of scenarios of a driver based on the dynamic of a truck? For example, you see that somebody is coming one way. So you can anticipate that the driver will veer or go left or right, but it depends very much of the speed of the trucks and things like that. So would you be able to have something like a interactive scenario that would allow you to further improve your prediction of what should be done? >> Yes, I think I understand. Let me let me try and repeat uh like the question so I understand. So if we can predict what is going to happen maybe in the other actors in the scene but can we also incorporate >> knowledge about what the driver might do >> as a reaction to what's going to happen >> and characteristic of a truck. I think uh I mean the short answer is yes. We it can be done. I think the way these models work is that it's all about data and it comes back to this point that if you have data and you can teach the AI about these scenarios, if you have enough examples of drivers avoiding accidents or swerving out of the way of something, then the models will learn to anticipate that and incorporate that into their reasoning. And so really what it comes down to is do you have examples of this in your data set? And this is where I think Motive is really well positioned with millions of devices or a million plus devices out there and millions of vehicles. We're able to detect and analyze data from hundreds of thousands of different events that are near collisions, collisions or dangerous scenarios. And so in some of those there will be driver reactions and other ones there won't and the AI will learn that in this kind of environment in this scenario this is what's likely to happen and so I need to alert this or I need to not alert and that's how we would incorporate that kind of knowledge into the system. So definitely possible dependent on having data is what I would say >> and I'm not saying that there will be convergence between what you're doing and autonomous driving but there are simil similar similarities so to speak and I I think there is room there is obviously room for both right because what you're doing is for 99.99% of the trucks and autonomous is.1.1% but there should be more >> uh integration so to speak what you're doing is more like uh preventive because it will teach people and to the M is just it just happens >> definitely. Yeah, good point. >> Okay, we have questions back. >> Excuse me. >> Okay, so my question is when we first set up Motive, we have independent contracts with different cities. So we couldn't put everything into one system to start out. We are working on a merger now. But my question is is would the different atlases be able to speak to each other so that we can get reporting out in that way not necessarily have to merge? >> That's a great question. I unfortunately don't know the answer to that. So I'm going to defer you to like your account account manager or someone who can >> probably find out the the specifics of that. That's it's a bit of an edge case. So >> yeah. Yeah, that's perfect. Like I said, I was just like, "Oh, but wait a minute. If they all have atlases, could they talk to each other?" Right. >> We'll get back to you. >> Okay. Anyone else? >> Yeah. How do you foresee AI impacting headcount? I know we see that a lot in the news. currently with uh especially the tech industry. Um do you foresee that being a big impact to fleet operations? >> Well, I mean the headcount impacts that I'm seeing right now are in my field of engineering. So that's where most of this is happening today uh that I'm observing. I'm curious what you guys think about impacts in your industry. I don't really have visibility into that. So >> yeah, at least right now for me, I'm not looking at it impacting headcount significantly. Um I I like I said, I view it more as I'm getting more value out of the employees that I have. I don't know if right now that's changing how many I need. I I think it's just changing our effectiveness as an organization. Um so when you think about a mechanic or a fleet manager out in the field, um I've more work than time today for those guys. So, uh, is is just helping them get the work done. I I don't think I don't see in the near future, uh, AI materially changing my headcount needs on the fleet side. >> What about from a dispatch perspective or analytics perspective? I'm sure you have teams that do those perform those functions. >> Yeah. So, I do lean on the IT organization for for a lot of the reporting and and analytics side of things. So, um I there probably is somewhere that um you maybe see headcount slow. Uh I again I don't right now it's not impacting us enough that you have people sitting on their hands. So it's it's I view it as a productivity tool still. Um but yeah, dispatch for us is still kind of decentralized. So I haven't really stepped in that pond yet, but um yeah, I'd be interested to hear. Yeah. Not from an oil field perspective either. It's not going to change anything for headcount wise. We still need our our professionals. >> Yeah. >> There's your answer. Right. >> Yeah. >> Not sure if you see anything in the future though. I I understand currently that's how it is. But >> anyone else? >> Anyone else? If not, I'm gonna sneak a quick one here. here. >> Where does what to do in Nashville? >> I'm allowed to do it. Come on. Where does autonomy fit into all of this? Have you guys looked into this >> level four fully driverless trucks? >> Are you guys preparing for this future? >> Obviously, like the dash cam doesn't need to be there if there's no driver, right? But I think the um the fleet operations, the platform itself is still going to be a very valuable asset in that world. And you know, I don't know where that is going to go or how long it's going to take to get there, but as you said, you know, we're focused on serving the needs of the 99.999% of miles driven today, which are done with human drivers in the seat who need this kind of technology to keep them safe. So, uh, who knows where the future will take us, uh, and how that looks, but it's something we're not we're not like focused on that at this point. >> I I would argue a little bit of the that you don't need the camera anymore or the telematics anymore once the vehicle's autonomous. And just that as a fleet operator, right, um when I have an autonomous vehicle driving around and there's an incident, which inevitably there will be, right, >> um the only source of truth I have is the person who built it, >> right? >> Versus me having my own third party device in the in the vehicle. So, I think, you know, as a as a fleet operator in the future of autonomous vehicles, I definitely see me still having my own device in that vehicle that that gives me a a different source of truth. >> Mhm. >> Mhm. >> Yeah. >> Yeah. >> That just scares me. >> Jamie's never ridden in an autonomous vehicle. >> You're going to get her arrival. >> Yeah. Just look. People are also terrifying. >> I know, but I know. Yeah, watch some of those videos, guys. People are bad. Bad in the 80s. >> All right. Do we have any last minute questions? >> Sh, since no one else will give you what to do in Nashville. I'm going to cut to the chase. Broadway is full of touristy stuff. Go enjoy it. But the thing you need to do in Nashville, if you're going to go to the um Honky Tons, we're from Nashville. Um, uh, Robert's Western Western World. Go there, but do it tomorrow night. Kelly's Heroes plays, best band on Broadway. Uh, and get a recession special. It's what you need. It's a bologn, a fried blowing sandwich, a bag of chips, and like a PBR, something like that. A PBR or a highlight one of those, a cheaper beer for like six or seven bucks. >> Yeah, it's crazy. Like, yeah. Yeah. So, you you can skip the other ones or go to them, whatever, but go hit Robert's Western World tomorrow night. >> Okay. >> Thank you so much. That's a pro tip. >> You're welcome. >> Most informative part of this. >> Are you in the band that's playing? >> No. >> Oh, okay. >> Oh, okay. >> Little selfotion. >> Little selfotion. >> I am friends with the guitar. So, but that's >> okay. I felt like there was a self-promotion. >> All right. I think we're out of time here. So, please join me in a round of uh calls for Hamlet. Thank you for coming and have a great rest of the day. See you guys around. on cube.