We didn't have good visibility of our trailers and where they were, and it was extremely challenging.
We were getting into accidents and we didn't know why. If you don't know why something is happening, you can't fix it.
And so what does world class look like?
If we don't have control and have visibility into and insight, then we can't run our business.
We needed one consistent way of being to identify where our major risks were.
Identify issues that the trucks were having.
Finding issues through CCTV was like finding a needle in a haystack. Having a single view with real-time data is paramount.
Having full visibility into what's going on on those vehicles, what's going on around those vehicles has always been a challenge in our industry for decades.
One of the most amazing things about Samsara is they actually listen to the people performing the day-to-day operations.
One of the things that we do that's unique is we involve customers in that process early on because it helps shape the direction of the product.
The Samsara team has been very receptive to a lot of our feedback and then they've been quick to initiate a lot of that feedback.
Seeing the end users that actually work at our customer sites gives us that real tactile feedback.
We love to engage with Samsara from a partnership perspective and we get a seat at the table.
Please welcome Samsara co-founder and CEO, Sanjit Biswas.
All right. Well, good morning everyone and welcome to Samsara Beyond. It's great to be here in Vegas with you. We've got over 4,000 people who've come in for this event here in person. Got even more people joining online. And I just want to start this morning with a note of appreciation. We know you're all super busy. You've got busy operations to run. So, we truly appreciate you making the trip out. We have dozens of different industries gathered here today to talk about the future of connected operations.
And when I think about what we're doing here, it's really a community that we've built. We're here to share notes. We're here to talk about how to put technology to work and solve real world problems. And this is a community that has been growing year after year. We're actually celebrating an important milestone today. This is our fifth annual customer conference that we've put on. And like I said, this is our biggest one yet.
But I want to start also with a note of thank you and appreciation for the teams that have made it to all five Samsara Beyond — the teams from DHL, Mohawk, Gordon Food Service, and so many more. So let's give them a round of applause.
You guys have been great. You've helped lead breakout sessions, share notes with each other, truly build this community from the ground up. And like I said, this is what it's all about. We're working together to build the future of operations.
And as a co-founder and CEO of the company, I'm fascinated by your operations because it's truly the infrastructure of our planet. It's how this whole place runs every single day. And as we got started with the company, we wanted to learn everything we could about how operations and infrastructure works. But these are industries that you can't read about in a book or you can't just visit a website and figure out how they truly work. You have to put on your boots and go onsite.
And for me, I really think of this as one of the big perks of my job is getting to go behind the scenes. In the last couple months, I've had a chance to go visit sites out in Colorado, see where the natural gas that heats our homes really comes from. I've had a chance to walk below the wing, go see how our airlines run every day, and get millions of people to where they're trying to go on time. I've even had a chance to visit some of the largest food distribution warehouses in the world. One of them fed over 10 million people in a metro. It was absolutely incredible.
And this is where our ideas come from. We love to spend time with you and your teams, learn about some of the real world problems and challenges that you face in your operations, and then we can go back and come up with really exciting new technologies to help solve those problems. And this isn't just a few of us at Samsara. This is cultural for us. This is how our company truly runs.
As we're getting ready for Beyond, we put out a quick message asking people to send in pics from onsites. And before we knew it, hundreds and hundreds of photos started coming in. Let me show you what those look like. It's absolutely incredible. So, these are team pictures of ours of people doing driver ride alongs, spending time in the maintenance shops, really going behind the scenes. And we tried to count how many onsites we did in this last year, and we got to a couple of thousand and we kind of gave up. That's how much we love learning.
So, also a big thank you for not just coming to this event, but for inviting us into your operations, the hospitality, the ride alongs. It really is so great for us to be able to learn up close.
So, while we were out on the road this year, we heard a lot of themes and I mentioned we've got a few dozen different industries here. You might at first think all these industries are so different, but really what we noticed were a lot of the same challenges coming up over and over and over. This year in 2026, the biggest theme that I heard was a rising cost of fuel. The price at the pump has gone up anywhere from 30 to 40% this year because of the war in the Middle East. And that's affecting you guys in a big way. Some of you have fuel budgets that are measured in the tens of millions or hundreds of millions of dollars. So you do the math on that, it's a really big hit to the budget.
And then on that note, I heard a lot about equipment. All of the equipment, the tractors, the trailers, the construction equipment, all the tools, everything has gotten more expensive and harder to come by over the last couple years. So your costs are rising on that front and the maintenance and service costs are rising too. A few of you said your operating budget — about 10% goes to maintenance alone. So that's absolutely enormous.
And if that wasn't enough, you guys are busier than ever. Across industries, you're helping build data centers from the ground up, helping modernize our electrical grid, rebuild our roadways so our towns and communities can operate seamlessly.
We wanted to try to come up with a visual of how busy operations is this year. And we asked the engineers to pull some GPS traces and try to come up with something. And it looks something like this. We did a snapshot from a single construction customer in a single part of Texas. This is up in the Dallas Fort Worth metro. And you see their operations running live here on the screen. This is a single day. We're talking about a thousand points of tractors and trailers, all kinds of equipment moving around. And it looks like a beehive of activity.
Which brings me to what I think is the most important part of the operation, which are the people, the millions of frontline workers who go in every single morning and keep our planet's infrastructure running.
I want to also start this morning with a note of appreciation for them. Let's give it up.
These are the people that we want to work hard to keep safe every day so they come home every night. These are the folks who work so hard that we want to make that job just a little bit easier every day, streamline their processes, make it a little more efficient. And this gives us a real sense of purpose. We're building technology to solve problems for people, to solve problems for you all.
Okay, so let's talk about technology a little bit. Technology is exciting. It's a great way to solve some really interesting challenging problems, but it's also a powerful tool. It's powerful for the front line and it's powerful for the back office, but it can also be overwhelming. We know that with technology, there's also a new problem that's emerged, which is data overload. I've heard that a lot in the field. There's just too many reports to look at, too many different systems. Drivers have told me, 'Wow, there's just too many apps, too many distractions.'
But I do have some good news. I mentioned this last year when we were gathered at Beyond that we're entering an entirely new era of technology. Now, we're entering something we call the age of intelligence. And this is of course all about artificial intelligence or AI. And AI can have a huge impact in operations. And it's getting better really quickly. So, as you all know, we can collect tons and tons of data from out in the field. But now, we can use AI to find insights on that data. We can figure out what's really going on. And in the last year since we met, AI has gotten even more capable, even more powerful. AI can now formulate plans, figure out tools, and really help you run your operation. It can take action on your behalf. And that's exciting. And that's new.
And the most practical part of that is that AI can now start automating away some of the task work that has to happen every single day as part of your operation. And this is a huge unlock. We can now automate away the grind. Think about the reports that many of you have to put together every week with KPIs and summaries for management. AI can now help with that. AI can help you with messages and workflows, help you communicate with the thousands of frontline folks out in the field. And this is all giving you time back so you can spend time on the important stuff, the interactions, the people. And I know you all have more that you want to do in your operations. We're excited to help make that happen. We truly want to help take your operations to the next level.
So, let me give you a couple of examples of practical real world ways that AI and all this technology can help.
Now, let's imagine for a moment that it's snowy out. And I know that's a little hard to do. We're here in Vegas. I think it's already 110 degrees outside. But a couple of months ago, many of you were confronted with some of the most extreme weather we've ever seen. Snowstorms, rainstorms, tornadoes. And just think about it. Wouldn't it be amazing if the system could automatically adjust settings, help drivers be safe on the road by changing the speeding settings, the following distance settings, and do it all automatically, maybe based on the driver's profile and the type of equipment they're operating.
It would be even more amazing if we knew about risky intersections or if there's a low bridge nearby. Doing this manually would take forever. You'd have to do it for thousands of vehicles. You'd have to be watching screens constantly. Most of you don't have teams to be able to do that. But this is the kind of work that AI can now do. Then if we think about the equipment that's behind the scenes in your operations, you can monitor all the usage. What if you knew exactly what had been used, what was dormant, what needed maintenance, how many engine hours did that run? All of this to do manually would be a tremendous amount of work. And I've heard it's a little bit tricky or challenging to get your regional managers to do all that paperwork, let alone every day. But the AI can do this. It can help you figure out which equipment to maintain or even where to position it.
And then on the driver front, we can really reimagine the frontline experience. What if we could give drivers a heads up as they approach a customer site? Many of you are doing thousands and thousands of stops every day across your operations. A lot of you have new drivers, new sites. It'd be amazing if we could greet a driver with a heads up of this is a security gate to go to or the dock door that they're expecting you at. Again, if you had a manual operation behind the scenes, a team doing that, making phone calls, it would take dozens of people. Most of you don't have the extra headcount budget to do that. But this is the kind of work that AI can automate and that's really exciting.
So this is what we're talking about. Reducing risk for the front line, increasing the efficiency of the operation and really helping you digitally transform with this technology. And that's what we're excited about today. So like I said, we're here in Vegas. This is the land of live shows and entertainment. And we've got what is the biggest slate of demos ever at a Beyond. So, we're going to keep things exciting and interesting. So, on that note, let's get into it. I'd love to introduce you all to our chief product officer, Johan Land.
Thank you so much, Sanjit. Friends, good morning. Oh, wow. It's so great to be here with you. I've been really looking forward to this.
And look, over the last 10 years, we've been truly pushing the frontier of what it means to be using tech and AI in physical operations. And every step, we've been doing that together. We embed into your operations to understand your problems. Wherever they are, whether that's in a vehicle, on a job site, or in a warehouse, we're there to understand them.
And I know that lately things have not been so easy. We have drastically increasing nuclear verdicts. So, we got diesel prices drastically up on the next slide. Here diesel prices are up enormously, 40%. We got nuclear verdict sky-high, driver turnover higher than ever. In addition to that, we got insurance costs up, new regulations, cargo theft, environmental requirements, and much, much more.
Look, I feel you. The pain is tangible. It's real. But we're there to help you. And we may not always get everything right, but we are committed. We're committed to you, and we're committed to our partnership.
And let me tell you, commitment, something that's very special to me. Commitment, it's not something that you just say. Commitment, that's something that you do. It's something that you embody. It's a value, something that you live by. Commitment, it takes sacrifice. And speaking of sacrifice, look, this morning, I was having my breakfast right over here. I was having my bacon and eggs. You guys love bacon and eggs? Yeah, we do, right? It's really good. But it had me thinking, see, when you make bacon and eggs, you need a chicken and you need a pig. But the chicken is not really committed. It's not making any sacrifice. It's merely providing an egg. It's just a supplier of an egg, right? But the pig, right? The pig, that's the true commitment. It's making the ultimate sacrifice. And that's how I see our partnership with you.
I'm really grateful that you're here. You know, I know it takes a lot and you put faith in us. We got a great lineup of things for you. Things that I think will make a true difference in your operations.
And to get started with this, in fact, last year we together helped prevent more than 380,000 accidents. In fact, over the last several years, we prevented more than 2 million accidents. That's a real reduction in human suffering. How about we give each other a round of applause for that, right? I mean, we're saving lives here.
Should be really proud of this. But I think we can do even better, you know, and the key to this is to look at the entire situation, everything that the drivers are doing from every angle of the vehicle and then acting on it and acting on it at scale. Therefore, I want to talk to you today about two things. First, how we see everything and then how we act on it in an automatic way using AI.
Let's start off with how we will see everything. Last year here at Beyond, we unveiled the first AI multicam which helped us see fully around the vehicle. And this has truly changed the game in terms of preventing crashes, especially in tight urban spaces where there can be lots of pedestrians.
And since then, you've asked us, what can we do to make visibility even better? Like what would the next level of the AI multicam look like? And I'm excited to show you what we come up with. So, take a look at the truck over there. We've outfitted it with four new types of cameras, one on each side, so that instead of four separate feeds, the AI has actually taken these four cameras and stitched them together. So, the driver gets to see a perspective from above like a bird would see it.
And this is what that looks like. Here on screen, what you're seeing is what the driver sees right inside of that truck. The driver gets this stitched together view right inside of the cab. And this gives the driver the full visibility of everything that is going on around the vehicle. And they get this in this single place. And of course, by the way, these cameras, they have AI inside of them, right? So, I want to showcase this to you. And to help with this, let's bring in Aditi and Evan over there. Thanks.
And now, as they approach the vehicle, they're visible to the driver on the screen, but as you can see around them right now, they're green boxes. That means that they are not in danger. But as they get closer to the vehicle, the AI detects them as being at risk and alerts the driver. In particular, as you now see them starting to move around the vehicle, you can see how the AI keeps track of them. And with that view, the driver can quickly grasp where they are relative to the vehicle because everything is projected into this single picture from above. And that's what true complete visibility looks like. So thanks Evan and Aditi for doing that.
And we call this bird's eye view. And it gives you that extra set of eyes that you need in incredibly tough to maneuver vehicles like school buses, box trucks, and much more. And I'm really excited to let you know that this will be available by the end of this summer.
Now, with the AI Multicam, the most appreciated feature has been the AI alerts. And I'm glad to let you know that we're now expanding to more types of these alerts. And we're talking about crashes. The most frequent thing I hear about is backing up or bumping into things, especially in tough environments. So, we came up with a solution to this. We call it rear collision warning. And we've done it in a quite novel way that I'm quite eager to show you.
So, take a look at this. So, here we're backing up, but as we do that, check out the color-coded areas that start appearing. So, these are the areas that the AI has detected as being in danger of us bumping into. And this can be a pole, a curb, another vehicle, or anything else that might be around us. And that makes those high stress backing up moments a whole lot easier. And you're going to see us continuously adding more detections like these to the AI Multicam.
The reason for that is that this 360 degree visibility changes the fundamentals of safety. And one thing that I've heard you consistently say is that you want visibility on everything. All your vehicles. Visibility for baggage tugs on the tarmac, forklifts in warehouses, excavators on job sites. But look, I got to tell you what you're asking for here is just not reasonable. These are tough environments and we thought that this would be impossible to solve until we did.
Let me show you. This is what we came up with. This here is what we call the 360 camera. It's a sleek single module camera purpose-built for operated equipment now. And you can mount it on any kind of environment. But even more magical, what it has is a full 180 degree view from the front and a 180 degree view from the back. And then we stitch them together, creating a 360 degree visibility from a single device. And of course, like all of our devices, this thing stands up to your toughest jobs. Has an IP67 rating. It's ultra ruggedized to withstand dust, water, and heat. It even has a button, a speaker, and a mic designed for these environments.
And look, I know this truly sounds unbelievable, right? 360 degree view from a single device. To be honest, you shouldn't believe it until you see it. Seeing is believing. Let me show you what this looks like. If you can flip to my screen here. So now this is Alaska Airlines. Let me maximize this. So as you can imagine, having full visibility in this kind of environment is super critical. And right here, we're about to push back this aircraft from the gate. But check out the resolution. It's incredible. But here's the kicker. Because this is a single device, we can zoom around like this. See anything? You can see the driver here, what he's doing. You can see his hand fiddling around. We can zoom in, zoom out, anything we want because this is a single device. How about that? Is that incredible or what?
So, now we got this running already on a bunch of different equipment like forklifts, tail handlers, and much more stuff. In fact, we have this live right now with more than a dozen customers. And I'm psyched to let you know that this will be broadly available later this year.
Last year, all of us came together as a community to do something pretty special. We shared our data with each other to keep each other safe. Millions of vehicles use Samsara every day and we tapped into this data and community to create weather intelligence. And it's been remarkable to see what you've done with this in order to stay safe. And this has been possible only because we created this as a community.
And I'm glad to let you know that we're now having the next wave of features coming out of this. We are detecting potholes and informing cities about it so they can repair them. We're detecting down power lines and then alerting the other drivers so they know before they get there. We're using our collective data to detect dangerous intersections and then informing drivers so we can route around them and avoid it. And all of this is possible just because we came together.
But one consistent feedback that you've been giving is that we have all these cameras. Why can't we use them for other stuff? Not safety stuff, but the other stuff, all the operational stuff. And you've been giving us a lot of ideas, things like really exciting ideas, by the way. Ideas that can truly make a difference in your operation. And I'm really excited to share that we're now starting to develop these non-safety detections like road defects, waste management, passenger transit, and there's much much more to come here. We are going to put your cameras to work for you broadly.
So, we've got new cameras, new alerts, new detections outside of safety. And all of this lets you see everything. And that's absolutely crucial because only when we see everything can we actually act on it.
Now, one thing that you do that has piqued my interest a little bit are ride alongs. That's when a senior driver gets in the cab, rides along, takes notes on the other drivers' driving, catching subtle behaviors that otherwise probably would be lost. Things like, are the mirrors being properly checked? Is there judgment and the ability to anticipate situations before they happen? Signs of aggressiveness or signs of disengagement, subtle signals hard to catch. Ride alongs are very impactful and very complementary to cameras and AI events, but ride alongs are hard to scale. And drivers tend to be on their best behaviors when someone is sitting next to them.
So we put AI to work on analyzing billions of miles of driving to see what is it that truly matters. What are those subtle signals? What are we missing in our current approach to AI and events? And what we found was staggering. We identified 22 behaviors across six principles that are the best predictors of crashes. And then we built it. We built AI ride alongs, infinitely scalable. And the magic is because the AI is doing this, it can be done at any time. I'm really excited to show you this one.
So, this here, this is Tim. So, he's a technician with Turnpoint, nationwide HVAC fleet, and they have truly safety at the heart of their mission. And by the way, just so you get it — I'm not making this up. Like this is real. Like what you're about to witness here is the first AI ride along in the world. That's how excited I am about this.
Now Tim here, the way this works is that the AI, it does just what a human would do. It rides along, takes notes of things, and what we see is that people most of the time they do good things. It's like Tim here. It's green everywhere. He's checking his mirrors. He's paying attention and everything. So, typically people do really well.
If you click on here, it looks like there's some things here that really need attention. Let's take a look. So, here is driving through an intersection. The AI saying, 'Look down while driving to charge a vaping machine.' There. You getting that? The AI saw that he was not paying attention but also identified there was a vaping machine. And in addition to that, believe the AI here will, if we look at this one, it has also assigned a training to Tim to manage these distractions.
And now building this functionality, let me tell you this was not easy. So, I'm quite proud to let you know that these AI ride alongs are available to all of you to use in open beta right now.
But that's just one safety task that AI can automate. A lot of you spend hours sifting through data to understand which of your drivers needs help with coaching and which ones do not. So we've come up with a pretty clever approach to automating also this. The way it works is that the AI looks at drivers' complete history. Not just their event, not just a single ride along, but a holistic view of every data point we have about the driver. More than 45 different risk factors, things like road conditions, weather, coaching history, and much more.
And I want to show you what this looks like in the case of Tim. So all I have to do here is click on Tim and the AI brings up this. The AI here has identified that Tim has a harsh braking problem, mobile usage and then the inattentive driving, the vaping device if you remember. But in addition, the AI has here deemed Tim as very high risk. That means that he is at high risk of causing the next accident.
But here's the kicker. The AI has not just looked at Tim. It has done the same for every driver in Turnpoint's organization. So, let me show you what that looks like. So, here we have it. They have almost 4,000 drivers and the AI has looked at all of them and we chose Turnpoint because they are an absolute leader in running safety programs at scale with automation. And the AI's classified 300 of these drivers as very high risk. That means that the AI has predicted that those 300 are at the highest risk of causing the next accident. And in Turnpoint's case, they prioritize these drivers. They focus their in-person coaching on those 300 and they send their lower risk drivers to automatic self coaching. They coach everyone, but they prioritize the ones that are at the highest risk. And this is quite extraordinary.
We call this coaching priority. I'm really excited to let you know that this is available to all of you right now.
So, we've got phenomenally increased visibility and the ability to act on all of it. And with AI, we're making this truly scalable. In fact, I think that the potential in what I've showed you here is huge. This is a pivotal moment when it comes to safety and running safety programs. We're going to look back at this day as the moment when we shifted safety into the next gear.
But there's one thing, none of this matters. All of this would be irrelevant if we weren't having the drivers with us. That's why we've invested into drivers and driver experience. And to help with this part, I want to introduce you to Arpan. He's our new head of safety. So, please give him a warm welcome into our community.
Wow. Johan, thanks for the warm welcome and for telling a few thousand of our friends that we might be pigs. Let's move from the farm to the front lines to where drivers spend their days. You know, our purpose is to help every frontline worker get home safely at night. For us to do that, we know we need to build products that drivers love. And to us, that means being there anytime they need us, bookending their days, all without creating new distractions.
And so over the last year, our team has been obsessed with answering one question. How do we reimagine the driver experience? And as we look to the future of driving, we really grounded ourselves in the present in all the ways that we inundate drivers today. They have screens. They have more screens. They get beeps and alerts. They have post-its stuck to their dash. And all of these things pull at your driver's attention.
But what's the alternative? When you need to reach your driver, you pick up the phone, call your drivers, right? But we don't want drivers to be tempted by all the distractions that come with phones. So, we challenged ourselves. Could we help you reach drivers without phones, without mobile plans, and without another screen? Well, today I'm excited to share that you can now reach drivers in the moment using Samsara's cameras that are already in the cab. So, I'm going to hop in the truck now and show you what that looks like.
Incoming message to your camera. Please turn off your radio. Hi, Arpan. I'm an AI agent on a recorded line. Are you ready for your start of day briefing?
Your safety score is holding steady at 96 and your safe speed streak is up to 236 miles. Great job. Today's drive is a 90-minute round trip from Vegas to Boulder City. Conditions are clear and traffic is moving well. This week, I noticed you had one following distance event. So, today let's aim for 4 seconds of space, especially when traffic is merging. Do you have any more questions?
That's super helpful. Are there any high-risk areas on my drive?
Yes, there's a construction zone coming up on I-15 in about 10 minutes. The speed limit drops to 45 mph in that stretch, so give yourself extra clearance and watch for sudden stops. That's it. Have a safe day out there.
Glad I could help. Drive safe and goodbye. Call ended.
All right. Wow, that feels like the future. Imagine an AI agent delivering a personalized briefing just like that one to every driver in your fleet. It's a level of service that you could only achieve today by hiring teams of dispatchers. And this is just the beginning. Start of shift will include weather updates, safety tips, driver recognition, road intelligence, and risk alerts. All delivered at the right time in the right place, built on top of custom events designed for your business.
All right, so now that I've got my shift info, I'm going to turn to navigation. So, since we launched Commercial Nav last year, we've made a ton of updates, most recently with CarPlay. CarPlay is an incredible product. The interface, the orchestration of apps and sound. And so we wanted to take a product that we know and love and give it to our commercial drivers. And that's what we did. Here's Samsara powered by CarPlay. So the way this works is as soon as I got into the cab, actually the Samsara driver app paired with CarPlay. And from here, I just accept the route and we'll be on our way. And this integration means that the powerful features of the driver app like turn-by-turn are now accessible by voice and the in-cab display. It means that all that custom routing logic that you've built up over time and all the data that we have across the Samsara network sort of get boiled together and piped back into CarPlay guiding us every step of the way. All right, I've got a trip to take so I'm going to throw it back to Johan.
All right. Thanks a lot, Arpan.
Well, that was great. Now, let's imagine that Arpan is now on his way to a job site, but let's also imagine that he has forgotten something. Let's say he's forgotten a generator or something like that. So, but instead of calling him on his phone, which by the way would be unsafe and also illegal, I'm going to call him on his camera. To be clear, my voice is going to come through on his side of the cab. And to do this, all I have to do is to click on call driver right here.
Incoming call to your camera from Johan Land.
Hey, Arpan. I just realized that you're going to need a generator for this job. There's actually one at our North Vegas yard. Can you swing by and pick it up?
Uh, sure. I can do that, but I have no idea how to get there.
All right. Um, okay. So, let me see here. I should be able of changing the route inside your vehicle on CarPlay. Let me click on dispatch here. Address North Vegas yard and dispatch. Right. You should be seeing it coming through on your end now.
Yep. I've got it here. And I'll just accept that route.
Awesome.
And so that's how we've reimagined the end-to-end driver experience. From start of shift, which gives drivers everything they need to know, to fluid, intuitive navigation powered by CarPlay, to two-way voice calling built on top of the tech that you already have in your cab. You've got AI agents co-piloting your drivers in real time. And it's a safer, smarter way to get around with voice at the center of it all. Now, this is just the beginning for us in reimagining the driver experience, and we can't wait to bring it to life with you. Now, back to Johan.
All right. Thank you so much, Arpan. And I don't think I need to tell this room that drivers are just extremely important. We cannot do anything without them.
And we've been building a lot of what we've just been showing you together with UNFI, one of the largest food distributors in North America. And I want to show you something. Let's roll the video.
In 1976, our founder, Michael Funk, started this business from the back of his truck. He really valued natural, organic, and specialty foods. And with that passion, he started delivering to customers in his neighborhood.
50 years later, we are celebrating that natural organic spirit within UNFI.
We service more than 30,000 customer locations, moving approximately 230,000 products. Every major safety breakthrough in our industry came from leaders willing to try something new in service of protecting our people.
That's the good thing about UNFI is most of our drivers here have over 5 years. 80% of them guys are here 15 or 20 years. People stay here forever because this is a good company I work for. Probably the best job I've ever had.
It's really important to us that when we think about making these large investments in these technologies that we're rooted in the principle of protecting our people so that we can ensure that all of our associates go home the way that they came to work.
We're delivering stuff that people actually are going to use in their everyday life. Everybody needs food. So it's like, hey, I'm doing something for the community, our state. Wouldn't you be proud to be a driver to drop off all the food? It's kind of a good feeling.
All right, so please join me in welcoming to stage UNFI Chief Safety Strategy Officer, Tasin Shadwick.
Hi.
Wow, it's great. Thanks so much for being here — thanks for having me, Johan. I appreciate it.
So that driver just shared with us this is the best job he's ever had. It's like, how do you achieve that?
I wish I could take credit for that, but that's our leadership and culture showing up right there. But at UNFI, as we continue to build our next generation supply chain to operate more effectively and more efficiently, what we're seeing are real results already. So, compared to exactly about a year from now, we have seen a 4% improvement in our on-time deliveries. We've also seen a 5% reduction in the number of miles driven per delivery. But what matters here is what we just heard our driver say, that this is one of the best jobs he's ever had. And that's the signal that we care about. And so it just continues to challenge us to listen and keep getting better every day.
Yeah. So I'm dying to hear, like, what is it that has shaped you as a safety leader to where you are right now.
Thanks for that question, Johan. So for me, this is really personal. I grew up in a developing country in Africa and I dealt with a respiratory illness as a child and access to care was quite limited. So when I did get that help, it just really stayed with me. And I remember thinking originally, I'm going to grow up and become a doctor. That changed early in my career when I actually met a safety leader who had been a nurse before and she said something to me. She said, 'I used to help people recover. Now I help to prevent those injuries from happening altogether.' And that just really stayed with me and it's really shaped how I think about how we go upstream. How do we help prevent injuries before they happen?
Yeah. And I think we all have had some personal moment that has shaped us into what we are right now. Now, you've been helping us with these ride alongs. Can you tell us what that's been like?
Yes, absolutely. So when we started to think about ride alongs, our teams wanted to be involved and so we raised our hands and we said we want to be a part of this beta test and so we tested the tool. We challenged assumptions and what was really interesting is we wanted to provide Samsara with real life insights that you just can't replicate in a lab. And so we were able to provide that information to the Samsara team and just really help build that tool together so that we could have a tool that we believed will provide value to our drivers and most importantly, you know, make them have the information that they need and that they feel confident with the tool.
And this is really key to us to understand your operations. So what were some of the insights that you shared with Samsara?
Yes. So it was really important for us to share practical insights. So when we were going through the tool and the great team at UNFI — that some of them are here — some of the examples that we identified were early in the beta. I remember there was an example where we had identified a situation where an event had been triggered. So that event had been triggered, but the truck hadn't moved. And there was another example where the weather had been overstated by the tool. And those details really matter because if AI are going to be part of the safety conversations, then our drivers need the confidence that the tool is really there to help them improve and not just generate some background noise. And to Samsara's credit, Johan, your team, they took that feedback and they've made some real product improvements. So, really happy to see that.
Well, thank you so much for helping with this. It's truly instrumental in us building these products. I think you've also been testing out the AI multicam.
We have, yes. So at its core, the multicam is about better visibility. And better visibility means better awareness. And so when our drivers have better awareness, they're able to make more informed decisions. And at UNFI, we have a very diverse customer base. So large retail environments to tighter urban spaces. So backing, low-speed maneuvering is just part of our everyday. And what we have found with the AI multicam is that our drivers who are highly trained professionals — that doesn't replace their skills. It's just complementing those skills that they have. So we've really enjoyed it.
Yeah. And what is the feedback they're hearing from drivers?
Yeah. I'm really excited to share that with you. So the feedback has been overwhelmingly positive. One of our drivers actually said that it's like having a second set of eyes and I think that just speaks volumes. So the drivers are feeling confident with this technology. They are enjoying the technology and they're using the technology. And so I believe that, and I think a lot of you would agree, when you have adoption and you have trust, that combination really helps to move the needle on safety.
And Tasin, so you've been in this industry for a while now and in fact I think you're truly a leader among us. So I'm curious to hear, like, what is some advice that you would offer for maybe some people that are newer into this? What advice would you give to them?
Okay. I would say a couple things, Johan. I would say start with purpose, not with hype. Be really clear on the problem that we're trying to solve. And then I would say it's really important for us to remember that every major safety breakthrough in our industry came from leaders like all of you willing to try something new in service of protecting our people. And so it's just really important that we stay focused on that message. And then the last thing I'll say and I'll leave you with here, Johan, is AI isn't the story. It's just the next chapter. And all of us have a role in writing that next chapter.
Well, that's amazing. Yeah.
It's beautiful to see the safety community we're building here, you know, and thank you so much for your partnership, Tasin. How about one more round of applause for her right now?
You just heard it. You know, AI is the next chapter of what we're doing here. So back to you, Sanjit.
Thank you.
All right, thank you Johan and thank you Tasin. It is awesome to see the future of safety unfolding before our eyes and also to see the new reimagined driver experience. But we're here to talk about operations and really the entire operation, the whole operation.
So I want to take a step back and think about a different part of the operation which is related to maintenance and all the equipment that it takes to run your operations every single day.
Most of you on your teams, you've got some folks that come in early. These are the maintenance techs, the shop leaders that get in probably 4-5 a.m. and they're there to make sure that your operations run smoothly, that the drivers are able to get out on time. If you think about this job, it's been around for a long time and it kind of works the way it always has. It's a lot of manual process, walk-around inspections, checklists, talking to drivers about, hey, did you see any check engine lights, things like that going on in your vehicle? And again, the whole idea here is to make sure the operations run smoothly. This is an area where we think we can have a huge impact with connected data and AI.
Let's go over to the laptop and show you what the future maintenance looks like on Samsara.
Okay, so up on the screen we've got the Samsara dashboard and I've clicked into a different area here which is the wrench, which is a maintenance section. You can click on status and in a moment's notice we're able to see the status for the thousand plus vehicles in our fleet. And in this case, the AI has gone through and prioritized that list for me, the maintenance technician, to know which vehicles I need to focus on. So already this has saved me a ton of time. We can see here there's a couple of vehicles that have critical priority attached to them. So I'm going to click on this first one here, which is a Freightliner Cascadia.
So I just went into it and what's going on is the Samsara gateway that's plugged into the truck is on the diagnostic port. That means it's connected to the engine computer and we can see what kinds of check engine lights or fault codes are going on. We can even understand exactly what the code is and figure it out and decode it for us. Now, I don't know about you guys, but I'm not a diesel mechanic, so it's really handy to have a tool that decodes these kind of cryptic numbers. In this case, there's a code that's called 3251. And what that's referring to is there's something going on with the exhaust...
Pressure sensors. And this ends up being specific to the vehicle and even the kind of engine that's in it. We can see we're running a 2024 Freightliner Cascadia and it's got a Detroit Diesel DD13 in it. So in this case, what the system has done is basically cracked the service manual and said this is what 3251 means. There's a lot of text on the screen. This is explaining what exactly the fault code is referring to, what the recommended actions are. Let me scroll down a little bit and show you some more. So on the left hand side, you see some statistics. It's giving us an idea of how important this is. A lot of fault codes are just informational. They're FYIs and that's helpful, but it doesn't mean we need to do anything about it. In this case, the AI is telling us, hey, this is a moderate fault. And you might be wondering, well, how does it know that? So, on the right hand side, you can see some numbers. The Samsara system as a whole, as you just heard, is deployed across millions and millions of vehicles. We see all the different thousands of fault code combinations, and we've seen them over a span of over a decade at this point.
In fact, for the Detroit Diesel DD13, which is very specific type of engine, we've seen over 107,000 of them, and we've seen what happens to them over time. You can see here that this fault only occurs a quarter of a percent of the time. So, that's really handy to know. And that's going to inform the bottom part of the screen, which is the so what, what happens. And this is really helpful because as a maintenance tech, I don't want to call a driver in and take them off the road if they can go about their day. This is telling us that this fault uh is a sensor issue. It would cost somewhere between $100 and $800 to repair. And if we don't do anything about it, about 22% of the time, this turns into the next fault code in 500 miles. That's super handy. That's exactly what I needed to know. I can send that driver out, they can run their route, they can come back and I'll swap their truck out at the end of the day. Really practical, really handy. I also have an idea of how much this is going to cost to address. And again, really practically, what happens if we don't get to it? Let me scroll down a little bit more. And you can see if we don't do anything about it, this is likely to escalate. It's going to turn into this next fault code, and that costs a little more money to repair, maybe up to $2,800. And if we don't do anything about that, it goes even further and becomes uh the most severe fault, which is $3,900 to repair. So, really practical, really real world. This is an application of AI to maintenance.
Okay. Now, let me scroll up here. I know a number of you have been in operations for decades and you have really good instincts. Probably some of you are thinking, 'Hey, that's a pretty new truck. I wonder if that repair is going to be covered under warranty.' So, to figure that out, we can use the system as well. I just click the check warranty coverage button. Now, most of you are buying hundreds or even thousands of trucks every single year. You negotiate specific warranty coverage with your OE partners based on the workloads you have and the routes you drive and so on. You can now upload all of those warranty docs into Samsara and the AI can be aware of them. So when you click a button like check warranty coverage, the AI is able to look at this fault code, compare it to what's covered under the warranty and tell you whether it's it's something that you can get paid back for or not. In this case, if we look at the right hand side, you can see this is actually a covered service test. That's great. This repair should be covered. So we can get some more details here. It tells you what's covered, what's not. Tells you kind of the bottom line. And then you see a question here. Would you like me to draft a work order with the covered service house pre-filled? Sounds great. Yes, please. Um, I always say please and thank you to the AI because, you know, you never know.
So, what this is doing is it's it's firing up the AI and it's going out to the vehicle and pulling all the information we're going to need to get a work order going for the warranty coverage. And it's on the left hand side already opened the work order. You can see here we need a couple of bits of information. We need the odometer from the truck. We need the number of engine hours. It went and pulled that for us. It's got a description there for the shop. I'll scroll down a little bit more. And then you see the service tasks. So, like I said, this specific repair is covered by our Detroit diesel warranty, but usually you have to submit some other docs. So, some photos, a couple of checklist items. It's gone, read through that PDF, figured out everything we need to do, and streamlined it, and put it right here.
Now, let's take a step back and think about how much work just happened and how we would have done this in the old days. You'd had to have a diesel mechanic, an expert who could crack the service manual, figure out what that fault code is. You'd have to make the decision of whether to repair or not. And if you did, you'd have to go look through all that warranty coverage documentation and deal with that. That whole thing might have been two hours of work. That's two hours nobody has in their day, and the system just did it for us in about two minutes. So, that's pretty incredible. So that's for this specific truck. And like I said, I think many of you have great operating instincts. A question that I've gotten a lot from customers is like, well, you know, we typically buy trucks dozens or hundreds or even thousands at a time. If I've got an issue on one truck, I might have it on others. So we can even ask AI to help us with that. So we can go down here and say, uh, do any other trucks have the same fault? And again, this is about being practical. Chances are we bought a batch of vehicles around the same time. We run similar routes in our fleet. This is firing up the assistant. It's actually writing some code to go out and check the engine computers on all the different vehicles that are part of the fleet. And you see it doing it all for us here. Um, and I've triggered this all manually so we can demonstrate it, but you can have this do it every Monday morning or every Friday uh whenever you'd like. So, it's running the code. It's going to query the entire base of the thousand vehicles and said, 'Yes, there is one other truck uh 8639 that's got the same DPF sensor fault and it'll offer to draft a work order for it.' Pretty cool, right? What do you guys think?
So this is our new connected asset maintenance feature and it's got this agent working for you behind the scenes, what we call the maintenance agent. I want to thank the teams that have helped us develop this feature over the last year: the state of Tennessee, Estes, and the team at Herz. So big round of applause for them. Thank you so much.
So, by now I know many of you have wheels in your head that are turning, thinking of lots of ideas, other ways you want to put AI to work in your operation. And that's exactly what we're here to talk about. This is what we're excited about, which is why we're excited to announce a new area of our product that we call Agent Studio. So, Agent Studio is a workshop for your AI ideas. It makes it easy to bring AI like what you just saw into your operations. And it's easy to build. You don't need to write code. It's all in plain English and it's powered by the Samsara data that you've already got in the system. Let's take a look at how that works.
Okay, so I'm back at the laptop. Got a new tab here called agent studio. And like I said, this is like a workshop. The idea is to give you a bunch of templates, a bunch of ideas so you can try different things, customize them for your operation. I'm going to scroll down a little bit on the page and we'll go to this menu of templates. And so you see lots of different categories here. You can explore this um however you'd like. And in the top left uh you'll see one that says weekly KPI report. Click on this. So I've heard this a lot. Many of you have a report that you put together for your suite or other execs because they want to know how things are going in the operation. This might be a bunch of manual work you do first thing on a Monday. You might even do it regionally for all the different op codes or regional operations you have. But the idea is to pull in information not just about safety or not just about maintenance but about the entire operation that might take an hour or two of work and that's repetitive because you're doing it 52 times a year and you might again be doing it for all your different regions. Well, now the AI can automate all that too. Let me show you what that looks like uh with a real world uh KPI report that was generated by AI. So we have a customer reference uh that's been helping us build this feature. They run this report. They sent it up to their leadership and they run a great operation. So they were proud to let us show off their KPI report here on stage. So what you see here is this is a PDF, something that can really easily be attached and emailed out. It's got an exact summary at the top. It's got a scorecard or dashboard here, green, yellow, red, just the way you'd like to glance at it. And it's got all this information about the operation put together in one place. They've customized it for the things they care about. And you can even see some of the big wins and areas that we need to work on in the operation. This is streamlining. This is saving time. And again, it can save a lot of time, hundreds of hours for them over the course of a year. Okay, so that's the KPI report.
Um, and there's a lot that you can do here in agent studio. I won't have a time to go through all of them, but let me show you one more. I'll scroll down here and uh you see here geofence alert. So the agent studio, it's not just for reporting. That's one area that you can uh work with it on. But you can also do things like create voice agents uh similar to what Arpon just uh demoed in the truck. In this case, I'd like to set up a geofence alert so we can give drivers a heads up when they enter a customer site. And I'm going to actually configure it for you here on stage. So there's a quick uh kind informational tab here on what it does. I'm going to configure agent. We'll go into it. And you can see here uh the instructions. Like I said, the code is just plain English. So, we can go in, maximize this, edit it up. Um, and you can customize the voice, you can customize the message that it starts with, and you can even test it out just from the comfort of your laptop. I'm barring Johan's laptop, so it's going to probably greet me as him, but let's just uh let me show you what it sounds like. Give me one sec.
Hi, Johan Land. I'm an AI agent on a recorded line. Is now a safe time to talk?
Yeah, this is good. What's going on?
Quick heads up from your fleet team about Arya Resort venue for Samsara Beyond. When you arrive, follow the signs for receiving dock 5. Expect traffic and follow site specific safety procedures. Let me know if you have any questions.
Okay, I think I'm all set. Thanks.
Thanks. Drive safe. Goodbye.
And that's it.
So this is what we mean by really thinking of creative new ways to engage the drivers, make their jobs a little bit easier every day. This is the kind of call that every one of us would like to have for a driver, but no one can practically run because you'd have to make thousands and thousands of these calls a day. But like I said, the AI doesn't mind and it's infinitely scalable. We can make as many calls as we need to with this system. So these are just two of the ideas for you in Agent Studio. Agent Studio is now available to you all to try out. It's an open beta and if you're here live at the show, we'd love for you to try it out at our expo hall. So, go to the check uh tech showcase and try out a couple agents for yourself. Okay, so now let's take a step back and take a look at how customers are experimenting with both maintenance and agent studio in their operations. Many of you probably know one of our customers, Pitney Bowes. They sort the mail and packages for over 90% of the Fortune 500 companies. If you've received a package or a piece of mail recently, chances are Pitney Bowes had something to do with it. Let's take a look behind the scenes of their operation.
Pitney Bowes is a global shipping and mailing technology company. We are over a 100-year-old company. We help clients of all sizes mail smarter, reduce postage, and increase deliverability. We process over 15 billion pieces a year. There is a lot to how mail moves, how it's created, how it's sorted, how it gets from North Carolina to South Dakota. It truly is supply chain at its finest. Every day is different. Some things always can happen. Um weather conditions, traffic delays. So most important thing is to stay calm and try to focus on solving a problem. I feel that we are provided with good technology and the good equipment and that makes a big difference for us to feel comfortable every day at work.
We have partnered with Samsara since 2020. We are not scared to enter into betas. We love to engage with Samsara from a partnership perspective and we get a seat at the table.
We want our managers spending more time with our drivers and less time behind the desk. We want our managers talking with them, asking them about their families, asking them about how they did this weekend. Retention is the key. Without the drivers behind the seat, we don't work here. This AI, this automation, this is going to bridge that gap. And if we could bridge between tech and practical application, that's where we really win.
All right. Well, please give a warm welcome to the director of network solutions at Pitney Bowes, Michael Keller.
All right, so Michael, we're here in Vegas.
It might be fun to start with a magic trick.
Okay, let's do it. Let's do it.
Do it. Okay. So, Michael's been in mail for a long time. He's seen a lot of packages, a lot of parcels, and I understand that you can take a zip code and tell us where in the US it is.
We do it. Okay. I'm going to give you a number.
You tell us where it is. Okay. 94107.
