The Keys to Success for Robotic Deployments

[MUSIC PLAYING] Hello. Thank you all for joining. I'm Vatche Arabian and I'm here with my colleagues, Caleb and Adam. We're here to talk to you about Spot, Spot deployment, the keys to success, and how to really get ready for a future of walking robots at your facility. Before we begin, I thought we could do some intros. Caleb, you want to start us off? Yeah. Hey, everyone. I'm Caleb Sylvester. I'm the senior director of growth at Boston Dynamics, and I've been here for about six years. And I'm Adam Camilletti. I lead our sales engineering team here at Boston Dynamics. And I've been here for about four years now. Awesome. I'm Vatche. I've been here for almost five, and I feel like I've had a front row seat to a lot of the work you guys do with our customers. And we've seen a lot of these customers be really successful. We've seen people expand. And I really want to talk to you, guys, about the secret behind that and what these customers have been doing as they deploy these robots at their facilities. Before we dive too deep into all that, there's folks that are probably joining us today that are unfamiliar with the value of Spot or why people use Spot. So maybe let's start there. What's your elevator pitch? Yeah, I'll give it a shot. And it should be good because it's my job. But let's see where it goes. This robot is the most cost-effective way to get mountains of data in an area that might be really difficult to get data and difficult because there's stairs, it's large, spread out area, maybe it's dangerous, maybe it's hot. These are areas that are perfect for autonomous data capture. This is an autonomous roaming IoT platform. It's going to go out into the world and collect all different types of data that you wouldn't otherwise be able to get. And to add on to what Caleb said, most customers are accomplishing this now with either fixed sensors, which, are great on your very critical assets, but that doesn't scale well, or they're just sending someone around or a contractor around to collect this data, and that contractor might have a thermal camera with them, and maybe they're visually looking at things, but that's maybe 1 and 1/2 modalities. With Spot, you're getting a platform that has a camera, which we can run visual AI on, a thermal camera, so you can look for thermal anomalies across your facility, an acoustic camera that can do multiple things. The first application there is gas leaks. So going around your facility and looking for compressed gas leaks, whether it's air, hydrogen or any other compressed gas, it can pinpoint those leaks for you so you can fix them. My favorite thing that acoustic sensor can do, though, is it can listen to changes in noise over time. So we set the robot up to collect a baseline sound signature of say, a rotating asset. And over time, if that deviates from that baseline we set, we're going to let you know. And that's going to tell you, hey, your bearing needs to be lubricated. It's misaligned. Those are things that are going to prevent your production from going down. And the thing I've been really excited about recently, about Spot, is a multi-modal inspection feature that we recently introduced. So you're getting readouts from these assets, asset-based readouts. It's almost like when you go to the doctor, I get the pulse, I listen, I hear, I smell. We're doing a little bit of everything now with that one sensor package, and you're getting that readout within our software. And then also to pull on the AI visual inspections too, we've expanded into even more EHS applications compliance issues. And my really short elevator pitch recently has been see everything, miss nothing. Spot's moving around and you've got this historical record through site view, the street view that we have in the software of what's going on in your facility. That reminds me-- I was thinking about this the other day --the advancements that we're making, and the speed with which we're making them because of what we're seeing with VLMs and these large foundational models, Gen AI. These robots out there getting all this data and the ability for us to extract really interesting insights. It's just getting better and better. Every day I hear about something new that we're able to detect and then connect that back to some value proposition that's really insightful for our customers. So I keep talking to customers about data and insights out. And you talked about the doctor's office. The reason you're there is to get insights. In all of the data, all the questions are answered, all the data that's being gathered is leading to those insights. It's the same thing with Spot. We're really looking at the health of a facility. And the health of a facility could mean the health of the equipment, but it could also mean literally safety, fire watch, EH&S type of issues, even security. What's the full picture of the environment? Spot is a unique way to close the gap to being able to get that full picture. Traditionally, you can't get a full picture of the health of an environment because you don't have all the data to do that. You can't put cameras everywhere, you can't put thermal cameras everywhere, you can't put listening devices everywhere. But you can have an autonomous system carry those devices and gather that data. And I hate to be the guy that's like, and there's AI. But there's so much with AI. Yeah. If you think about it, Spot can integrate with other systems. And if those systems, say, have agentic workflows, well, now Spot, on your factory floor, can become the embodiment of that AI. That system can prompt Spot to do tasks. And now you have put that agentic workflow into the physical world. Yeah, it's a simple concept. But AI engines need data for training. So it's like cooking garbage in, garbage out. Most of the data that exists in these factories doesn't live in your typical off-the-shelf AI tool. So there's this whole effort to use roaming systems like this, autonomous systems to gather this data up and start to build insights about places that traditionally we don't have these insights on. You need lots of data, series of data over time. Adam alluded to this ability to listen to things and benchmark what good is. You need data over time to know that's what that's supposed to sound like. That's the temperature that should be when we're operating at our most optimal clip. When we have the best output in our factory is when everything looks like this. You don't know that if you don't have that data. And so gathering up all that data gives our customers the key that unlocks those insights. So shifting gears a little bit. In the years you've been here, have you seen any unifying reason that customers look to us for Spot on their facility? Something that's similar, maybe not exactly the same. Every company is different, but why are people looking to us? It's because their leaders are saying to them, we need more insights. We need to get much better and tighter at the way we're operating so that our yield, our throughput is at maximum capacity. So how do you do that? You eliminate downtime events. You eliminate those costly scenarios whereby you're interrupted. And what continues to drive it is leadership in these organizations saying, give me these insights. And then the people on the ground who are charged with executing that going, oh, I can't get the insights because I don't have the data. I have some of the data, but I don't have the full picture. That is ultimately what's driving the motivation for all of the customers who've been successful is a leadership vision for better operations through insights and the people who have to go execute that realizing the only way they get that is with more and better data. I mean, exactly that. I mean, the key to Spot is it's a mobile platform that can collect really any data you want if a sensor exists that can collect that data. And I think that's the huge value prop for Spot. Because it comes back to the data, all that data that Spot's gathering at your fingertips in our software, you've got a historical 360 view of all your equipment of your facility. How is Spot helping-- you hear a lot about digital transformation, industry 5.0, 6.0. There's all these buzzwords, but how is Spot helping folks with that digital transformation? I mean, I'd say a lot of the customers we have Spot's one of the main drivers of their digitalization efforts. And that's usually how the conversations start with them, their executives, their stakeholders, everyone. For me, digital transformation has become one of those phrases that means everything and nothing all at once. And I have to remind myself not to glaze over when I hear it is a walking robot in a factory, a form of transformation. In a way, yes, but the real transformation that's realized that can be associated back to a real business plan is what we're able to do with the data. So imagine for a second that you've got a factory where sound is a key indicator of what might be going wrong in that factory. Usually the only person who would know that are the people who've been there long enough to know what good sounds like. We've found a way using Spot to autonomously roam around, listen to equipment baseline what good sounds like, and then come back and tell our customers when something should be paid attention to. That's transformative. That's the type of thing that you can really say we've put a win in the digital transformation column, similar with something like an EH&S patrol, where most of our customers have someone walking around looking for things that are unsafe. Do we block a fire door? Did the fire extinguisher move? Did someone leave a ladder or a tool cart out where they shouldn't have? These are all potentially violations. They're safety hazards. Typically, you'd have to send a person to go do that. So how do you digitally transform that? Well, you send a roaming robot that walks around every day and alerts you as to whether or not those risks exist in your environment. So yeah, I think there are ways to get to digital transformation. You can't just introduce walking robots. You have to do something with the data. So what stakeholders need to be in that room when they're starting to talk about actual deployment? That's a good question. I think we found there is a mix that works really well, especially in their first deployment. You want operations, people who are thinking about efficiency and downtime in the room. You want maintenance people, people who are maintaining the equipment that lets the factory be successful, the environment be successful. You also want health and safety. But we've also found that because it's such a new species of technology, there's a lot to be gained from bringing the skeptics in. The people who are maybe scared of robots or concerned about, is this thing a toy? Is it actually useful? Bring those people in the room and let them be a part of that planning process and that information gathering process. And to add to that, it's very important to have someone on the customer side that is the robot owner, the robot wrangler. We like to call them sometimes-- especially at that first site, we've seen a lot where that person might become the program manager for Spot across the customer's network. And when we see that signal, that is a very big signal to us that the customer is going to be successful scaling Spot across all their facilities. What assumptions do organizations get wrong at this phase? Is there an underestimation of something being easier than it is? What do people usually get wrong? Spot is-- we've been selling it for what? About four or five years now. But it's still relatively new and it's an advanced technology. And we're working with people that are very used to automation projects, capital equipment projects that they're not a switch, but they're very standard process, install test, good to go. With Spot, there's a lot more nuance to that. So I think we need to do a good job prepping our customers to understand that, hey, this is going to be a project that we're going to work with you on to be successful. But it is a project. It's not just install this piece of equipment, turn the switch on. It's going to be a project, and we need to work through some kinks before the robot's going to be stable and successful. I think there are-- we've seen two types of customers where you've got the customers who are really used to doing like serial pilots. I've got a new tech. Let me just try it out, try it out. And they just have hundreds of pilot projects. And then there's the customers-- and sometimes you get customers that have both of these people in this org. You have the customer who goes, I want to operationalize this. I'm spending this money, I need to get a return out of it. I think if you can find a way to make sure that the customers aren't treating this tech as a science project, but they're actually looking at, what does scale look like when-- what does operationalization look like when this is successful? I think that's key. I think oftentimes customers don't engage their entire site in what's going on like, hey, we're introducing this thing. We don't have any other walking machines. This is new. This is different. Here's how to think about this. I think that's one thing that sometimes where we see some slip ups is the customers who have faster time to success and value and faster time to adoption are really open with their customers, or excuse me, with their team about what they're doing, what they're up to. And then the biggest one for me that I see is this idea that, this is an autonomous system, set it and forget it. It'll magically make things happen. It's going to generate data and that data needs to be acted on. All of the value assumes that the customer is going to do something with the data. So I think that's huge. You have to have a plan to address the data. And I think a plan to deal with the elephants in the room like, how do we talk to our safety teams about this? How do we talk to our unions about this? How do we show both our workers and our customer base that this type of automation is necessary and it's good, and it's driving out a better product and a more reliable work environment? So, I guess, color in the deployment process. What goes into that? I know you do a lot of pre-work before our team goes out to help deploy. Can you tell me about that? Of course, I can. So my wonderful sales engineering team, before we ever saw the customer anything, goes to the customer's facility and we're going to put together a full evaluation and a solution proposal for that customer. So that technical evaluation is going to say, hey, here's what Spot can do. Here's what it can't do. Just everything across that facility. That solution proposal is going to include some parts of that, but it's more of an executive summary. So it's going to include high-level what Spot's going to do. But then we're going to dive into the business case that we worked with the customer to build. And that's the proposal. Once the deal closes and we move into a project with the customer, they get a dedicated project manager and two application engineers that are dedicated to their project, and they're doing, say, four weeks of pre-work before they even come on site, making sure your network is set up, making sure you're doing your infrastructure changes, just holding your hand throughout the process before they come on site. And then once they're on site, I mean, it's go time. They're spending that whole first week integrating Spot, creating maps and missions, testing it, getting it stable. After that, maybe after about two weeks, we enter our acceptance testing phase. From there, we define metrics before the project, and we're going to measure to those metrics at that acceptance testing period. Once we meet those metrics, we'll hand the project off to the on-site team. And from there, we figure out, do you want to scale this? Do you need help? Do you need more training? We figure out the process from there. It's all about understanding the goals, achieving those goals, understanding the pain, alleviating that pain. So there's a lot of work that we do upfront to understand, what are the corporate goals and initiatives? What are the site goals and initiatives? How do we translate that into a deliverable that we can measure and say, hey, we were successful delivering this? And Adam's team does a wonderful job building that out. Everything from the corporate goals down to, how do we make the actual day-to-day maintenance part of this and inspection part of this much simpler and more straightforward? At the end of the day, if you still have to send people to do all this manual work, then we haven't hit our target. The goal here is to alleviate that pain and gather that data in a recordable, definable way that then actually does lead to the insights that make the difference and allow us to fix things and prevent the line from going down or increase our yield or output at the end of the line. I think we're touching on this. And once you get to that site acceptance test, the robots deployed, how are we guiding customers to measure the success of the deployment? You talked about data being a key part of that. What success really look like? So that's the first conversation that I want to have with the customer before we even begin the project is, what does success look like for you? And we can give them examples of, hey, we've worked with customers in your industry in the past. This is what they wanted to see, and we can work with them to build what those success metrics look like. But that should be the first conversation before the project even starts is, what does success look like for you, customer? I think if you have a plan to act on the data and you have a plan to integrate that data into your existing system so that it's not living in its own little island. And then you have people that are trained and enabled to spend a few hours a week really monitoring the robot, if you will. It's not a huge lift, but we do need to have someone who cares and who is paying attention to this new tool. It's like any tool. You need to have a human that's leveraging that tool and then using dashboards to understand what the impact that the robot's having so that you can measure it against your own metrics, but also report it up to your leadership to say, hey, we introduced this new thing into our factory. We've never done this before. Here are the results we're getting. And here's where we are on track in terms of realizing the return on investment. The most common question I get when I'm staffing a trade show and talking to people who see our solution for the first time is, what's the ROI? How long does it take? And I'd love to hear your thoughts on that. I mean, most customers we work with for capital projects, they need to see a three, maybe two-year payback, which is fairly achievable with Spot. Now, if we talk about purely manufacturing, they're really focused on reducing downtime, like Caleb mentioned before, increasing their gross line yield. There are some customers where they just can't hire enough people for the jobs. Either they're really remote or, in Europe, aging workforce. And you look around and there's no one in there that's under the age of like 45. And that's a real problem. They don't have workers replacing that. So there's a non-dollar value metric there too. That is, we don't have enough people to do this job. We need something else to do this job. I think the key to achieving an ROI is just to follow the prescription that Adam's team lays out. We build a business case, we deploy, we train the team on site, we act on the data, we fix the things that are broken. If we find a bunch of broken stuff and you don't fix it, you don't realize ROI. And if you follow that prescription, we see our customers especially in industries like food and beverage, pulp and paper, cement semiconductor-- we see customers, if they follow that prescription, getting to a 12 to 18 month ROI. You talk to any operations person, SubTo year ROI is the target. You get close to a year, that's incredible. You do that over and over and over again to get to that metric. So follow the prescription that we lay out and manage it the way that we help you to do, and you'll get there. In the early days of selling this robot, there was so much interest, and we had very little idea of how to generate an ROI. Now, we just had an incredibly cool hardware platform, and it was very much like you buy it, and then it was up to you to figure it out. Now we've done this so many times that we've really honed in. So when we're speaking to a customer about how to get that value, it's coming from a place of having done the hard work to figure that out and get there. We're adamant to our customers that there's a certain way to do this to get value. And if you follow that plan, you'll definitely nail it. My favorite thing along this journey, because it's the front row seat that I've had, is the robot came out as a platform. We saw where customers were having success. We started to work with customers on that story, and then that story just got better and better and better. Starting from the early days with our published testimonials like GlobalFoundries to as recently as ST Engineering and in cargo these customers have found value in what the robot's doing. And it's been a privilege to see that come to life. I think this discussion has been great. I think we're ready to take some questions in just a few minutes here. But before we turn over to that Q&A, we've been talking about the path to success, introducing this type of robot. What advice would you have for folks watching that don't have a Spot today that are thinking about it? What's the last thing you want them to remember about how to be successful with Spot? Yeah, I guess something I'd want to circle back on is, as these customers start scaling their fleets of robots, they are going to need something similar to internal IT departments but for the robotics fleets. If you think about the days when Microsoft started selling computers and companies started using Windows computers to do all sorts of things, it wasn't manageable for Microsoft to try to service all that stuff. Companies had to build internal teams, IT departments to handle this stuff, and I think robotics are going to follow a similar path where the companies that are most successful are going to have these internal teams dedicated to robotics. And I'm going to give you two things because I struggle with following rules and instructions. The first thing is, have a plan for the data. So I've been very repetitive about this, but have a plan for the data. What do I mean when I say that? Where are you putting the data? Are you creating work tickets automatically? Do you have a data lake strategy and a semantic layer that understands that data lake? Can we push this data to a data lake where an AI would take it into account and look at the total picture of what's going on in your facility and draw out predictions and insights? And the second thing, I would say, is prepare your site. I think the sites that have the best adoption are the ones that bring Spot to the lunchroom and let people drive it. They host Q&As. They utilize our posters and site prep videos that we send ahead of time. They name the robot. They name the robot like having a naming competition, and dressing the robot or wrapping the robot in the company branding, that stuff goes a long way. I was at a site once where we waited for a shift change, and we went out where one shifts walking out and one shifts walking in, and we stood there for an hour and just interacted with people. Let them drive the robot answered questions like sharing what this robot is doing and not doing. Those are the types of things that they seem little, but they can make a huge difference in your ability to adopt this type of new bleeding edge technology. Awesome. All right. We'll be back in one minute. So stay tuned. And we're going to take some of your questions. And we're back. Please use the Q&A button to ask us questions. We've pulled a few for discussion here, but our moderators are in the chat and we'll get to as many of them as we can. To kick things off, folks, seeing this theme of folks being weary of new technology. And what are some things people do or you've seen work when deploying Spot on site? Yeah, so one of the things we've learned over our past 4 or 4 and 1/2 years of having a commercial product in Spot and working with customers is change management is a huge piece. The workers on site might be worried that Spot's trying to take their job. People might just be confused of what the technology is. So we really partner with our customers to make sure we help them with that change management. So we provide different informational posters, informational videos that they could play and, say, the cafeteria where everyone will be there, can watch the videos. We recommend people wrap Spot with their company branding, make it more friendly, maybe even a naming competition. That's been huge for a lot of our customers. So we really like to partner with them on that change management and just be transparent on why this robot's here and what it's going to be doing. One of the things I'll add to my previous comments was that same site where I went with the team during the shift change, we went outside and showed the robot, demonstrated the robot, got people to ask questions and drive the robot. That same site was one of our early sites where we deployed Spot. And about a month after, the robot was really doing its thing every day, during that first month, I would go in and I'd see people stopping their work and getting their camera phone out and taking a picture. And it was just, what is this thing that's in our environment? And there's certainly a cool factor to the robot and it wows people. But one of the signals that you know you've done a good job in change management is probably about a month in. If you walk into a factory and no one's paying attention to the robot anymore. The robot's just doing its job. And everyone understands, this is a co-worker, another machine doing its thing. That's when you know you've done a good job. I forget which side it was, but there was one site that I remember being at when the-- something happened, the robot had to come in for service and go back, and they told me people missed the robot. So even though it was become a part of the crew when it was gone, because they've humanized this robot a little bit and they missed it because it was their robots, their site's robots almost like their mascots. Very cool. Very cool to see. One question that we have touched on in other webinars, but how many Spots are out in the world? Thousands. I don't know if Caleb knows an exact number. I think we're past 2,000 at this point. Yeah. And we are also starting to see some exciting stuff where not only do we have customers that have scaled across their entire enterprise, they're coming back and adding more robots to the same sites to do other types of jobs. So, yeah, it's certainly been from the early days of going, this feels like a solution looking for a problem. How do we nail down the right value proposition to seeing customers buy 40, 50, 60 robots and deploy them out around the world? It's a good feeling. Yeah, I recently, I think it was a month ago an email went around that Spots, our fleet of Spots. And this is including the customer data that customers have opted in to share that data. That fleet size has walked 1,000,000km. It's [INAUDIBLE] it. And I think we're only getting data from probably half of the robots. How many assets are customers typically looking at a facility? Well, I can break it down to one robot. One robot generally covers-- It's a lot. I wouldn't call it assets, but let's say 2,500 inspections a week. So individual inspection points that it can cover a week about 2,500, maybe 3,000. I've seen one customer up to 3,500. But for them, all their inspections were very condensed, close together. Most customers, I would say, 2,000 to 3,000 inspections per week per Spot robot. Another related question. How long are most customers missions? 30, 40 minutes. I mean, Spot's battery life expect at 90 minutes. And most customers run a 60 minute on 60 minute off cycle. So you get about 12 hours of inspecting across the day. So obviously Spot's working around people. Can you talk about its safety system. They're all going to me. [LAUGHS] I'm happy to answer anything. So at the core, Spot has its perception cameras. And that's creating a 3 meter radius around it where it can see the obstacles around it. Anything above 30 centimeters, it's going to see as I think like a wall. Anything above 30 centimeters within that 3 meter range, it's going to see and it's going to avoid. That's its base obstacle avoidance. Now, when we're running Spot autonomously, especially in our customer facilities, we had a LiDAR on top. Now, that LiDAR has a 50 meter range and it's looking for everything, it's looking for forklifts coming, it's looking for vehicles coming, it's looking for people coming. And what it does when it sees those things is it stops, flashes its lights, sounds a buzzer and lets you know, hey, I'm waiting for you to move. Get out of the way, do your thing, and then I'm going to proceed. It's the safest behavior we could come up with. You don't want the robot trying to react to what you're going to do. It's not smart enough to do that. It doesn't have a brain. So the easiest way to have the robot operate around people is if people come within a certain range, just stop, let them know you see them and wait. We learned a lot in the early days about what constitutes a safe robot and an environment, and it's interesting because there are no safety standards, machine directives built for legged robots there. Everything was built for AMRs or AGVs and don't really candidly apply to Spot. So we had to work with customers to be open minded and iterative to get to where we could combine the robot's ability to sense traffic, humans. With physical things like lights, sounds that signify to the people in the environment that the robots up to something. It's going up some stairs. It's stopping to inspect something. It's coming down some stairs. Things like when you see the robot attempt to go downstairs, it will pitch its body and look to see, is there anybody on the stairs? It's not going to start going down the stairs if there's a human there. These are all things that we learned in the field, working with customers who were very collaborative and patient with us to introduce these features. Another physical feature, a simple thing, an E-stop on the back of the robot that's intentionally placed. I was watching a YouTube or an internet video last night. I think I was actually on Instagram scrolling and I got targeted with this video. There was a Chinese humanoid robot in a restaurant in Orlando that was doing a dance, and something went wrong. As we know, with robots, things can go wrong. And it got really close to a table, too close to a table and then kept doing its dance routinely. It was just slamming the table and everyone in the restaurant was panicking. And someone pointed out. The restaurant workers were grabbing it and trying to pull it away, but it was still flailing and someone pointed out something as simple as just an E-stop on the robot would have completely solved this problem. So we've learned a lot of things. When you rush a robot out to market, instead of going through this thoughtful, iterative process of introducing it into industry learning what safe looks, learning what the expectations are, you can run into some challenges, and we've fortunately just been really lucky to have great engineers, great safety people on our team and really smart customers who are willing to help us make sure this is the safest robot on the planet. And to add, robots aren't just running at customer facilities. They're in our hallways. If Caleb and I try to go to a meeting room, we're usually walking past a robot in the hallway. So these behaviors are always being tested around us as well. So it is very, very robust that even we let them run in our office. There's been folks asking about their specific industries and some of the industries people are curious about what seem challenging like mining and materials processing. What are some of the more challenging operational environments that Spot is deployed at today? Maybe that might be a good way to just answer that question. Yeah, you already touched on it. Cement and mining, which I actually used to be the sales engineer that covered those industries, are very dirty. They're not the best environments. I mean, that's why they want to get people out of those environments. Really, the only changes we make is we educate the customer on the additional maintenance activities they might have to do. Usually you have to clean the fan filters every month, every two months. If you're a cement customer, you're probably going to be cleaning those every four or five days. But I mean, it really only takes 15, 20 minutes. So the data spots gathering is valuable to you, which our customers in these industries it is then that's worth it. I can think of two sites that when I walked away, I was like, wow, that was wild. One was Fukushima. I got the chance to go to Japan. I drove into the Fukushima plant and seeing that up close and seeing what Spot's doing there, it was just unbelievable. And there's such an important mission there. Humans aren't going into these environments. It's incredibly dangerous, and we've been able to send Spots into these environments flawlessly for months and months. And that's been really cool to see. The other one that I think of is I went to a wallboard facility or sheetrock facility in Canada. That was like the dustiest place I've ever seen. And it was just inches of dust on everything. And it was like not a lot of people in that environment, not a desirable place to live and work as a human every day. Everyone's wearing respirators. And so that was one of those examples of, you can just see so clearly this was absolutely tailor-made for a robot. So we're getting to the end of our webinar here. But before I let you guys go, there's people in the chat that want to know about Atlas. And how are these robots going to work together? Do they work together? What's the future look like with Atlas in the mix with Spot? Yeah, I think everything will be driven by the use case, the most efficient and practical way to accomplish that task, and then the economics of putting that all together. So in other words, we now offer three robots, really, Atlas, Stretch, and spot. When our customers come to us and go, we have this challenge, we're going to look at that challenge and we're going to look at things like tack time and physical environment and how much space do we have and how quickly do we need to move and what are the safety constraints. And we're going to look at the economics to go, how do we make the business case work? And we're going to prescribe the robot that makes the most sense. And there are a lot of things that make a ton of sense on two legs. There are a lot of things you really need two arms and hands for. And then there are some things that we're going to go, no, this is much better done with a four-legged robot that's lower to the ground. So everything from form factor, footprint, height, safety mechanisms all the environment considerations and the economics are all going to play into that. Do you think that folks that have gone through the experience of deploying Spot will be in a better position to deploy a humanoid like Atlas on their site? I believe so. And one of the core foundational reasons for that is I've watched our company build, deploy, and train in-service and support a quadruped platform over the last six years. So there's so much we learned from that that puts us in a really unique position as a company. We're not a company with a cool humanoid and a lab trying to build all these business functions. We have them built. We have a world-class deployment team. We have a world-class sales engineering team, support service logistics. We know how to take robots to market, which I think gives us a huge advantage. Similarly, customers who have gone through this journey, they know how to introduce walking robots to their safety team. They know how to meet with their InfoSec team and figure out, what ports should we or shouldn't we open? What package should we or shouldn't we send? The practical things like, how do you talk to your union about introducing walking robots into a factory? All of that stuff is foundational stuff that you're going to have to do. And so our customers who have done all of this are already well ahead of the game. When we introduce humanoid, it's going to be quick. These robots are going to get on the floor in an operational way, much, much faster than they would have if they didn't go on the Spot journey. And I have two things to add to that. So Atlas is going to use the same platform Spot does, our orbit platform, our fleet manager, our data warehouse platform. So right there you have an advantage. You're already using orbit. You have it set up across your enterprise. There's advantage one. Going back to the change management, which we've talked about multiple times here. It can be very difficult to get Spot a four-legged, cute, yellow robot into a customer's facility and get people to accept it. If that's difficult, imagine how difficult it's going to be to get a humanoid form factor into customer facilities. Imagine how much more difficult it would be to jump right from nothing to a humanoid robot. If you've already done that process with the four-legged robot Spot, I think it's going to be a lot easier for Atlas. Awesome. Thank you both for your time today. And thank you all for joining us. Has been a great discussion. If you're interested in learning more about Spot Atlas, please head to our website and we'll see you next time. Thanks. Thank you.