How to Improve Plant Reliability with Mobile Robots

- Thank you for joining today's webinar, How to Improve Plant Reliability with Mobile Robots. My name is Vatche Arabian, application expert for Boston Dynamics. We've got some amazing stuff to talk about today. But before we get started, I just want to run through some quick housekeeping. We'll be sending out a recording of today's presentation following the webinar. If you have any questions during the webinar, please use the Q&A button below, and we'll be going through some of those towards the end of the presentation. We're also going to have a quick break right in the middle for Q&A. So be thinking of questions as we go through the content. To kick things off, let's do a quick round of intros. John, you want to start us off? - Yeah, absolutely. Hey, everyone. My name is John Weiler. I am a global senior sales manager here at Boston Dynamics. I'm excited to be with all of you today. And over the past couple of years, I've visited a lot of plants around the world with the robot, and I'm excited to share some of those experiences with all of you. Kick it over to Andrew. - Yeah, great to meet everybody. Andrew Wang. I'm the director of enterprise sales here at Boston Dynamics. Similar story to John. I've been in robotics and automation for the last 10, 11 years. So been able to see a lot, go to a lot of places, and experience a lot of different things with manufacturing and the industry. [INAUDIBLE] - Yeah, cool. Good to go. - All right. Yeah, so we'll just go right into it here. A lot of people-- everybody's on various stages of their journey of digitalization and industry 4.0. And throughout the thousands of conversations we've had throughout our customers, potential customers, industry leaders, we continually hear multiple key themes throughout the industry. Everybody needs to increase reliability and uptime. They're trying to meet some ESG goals. So environmental and sustainability goals. They're trying to solve labor shortages with automation. So we've really seen in the past five or six years a boom of automation, whether it's collaborative robots, whether it's quadruped robots, AMRs, those types of technologies. More recently, we're hearing that people are leveraging AI for predictability when it comes to their maintenance and, really, their production and their environments in general. And last, but most importantly, prioritizing safety when it comes to their operations, their people. Everybody wants to make sure everybody gets home safely. All right. Just setting the stage here. I mean, as many of you know, manufacturing sites are very complex. We've got aging infrastructure. Many facilities are 50, 60, 70-plus years old. You've got dynamic areas. You've got challenging environments. Anybody that's been near a cement plant or paper mill you've seen, it's just full of various obstacles, whether it's wet floors, whether it's leaks, steam, dust, muddy terrain. A lot of the times, it's just this old industrial manufacturing. It's gritty. It's great. But what can really even be automated there? And then you get exposed just about everything as well. There's also the elements of OSHA and MSHA, making sure that safety and compliance are being met there. So just a lot of things that go into manufacturing. So what does that have to do with Boston Dynamics at all? So this is where Spot comes into play. So Spot is Boston Dynamics' quadruped robot, which can operate in most any environment. So we like to say Spot likes dull, dirty, and dangerous. So for Spot, we have them operating in all types of industrial environments, whether it's a factory-- it could be a wood mill. We have inspector Spot for that. And he's here to help with industrial inspection. And really, it's acting as a roaming IoT device. So what we've done with Spot is we've outfitted it with multiple payloads to create what we call an EAM or Enterprise Asset Management package. And so that includes the Spot robot you see there. It includes a 360-degree 4K camera with 30x optical zoom. We've got a thermal camera there, an acoustic imager behind it with ultrasonic microphone capabilities. And then we call it an EAP2 LiDAR, which is like the brains behind it. You've got a self-charging dock and then Orbit, which is the software and fleet management behind it. And the way we like to think of it is it's going from forwards to backwards there. At the very front of Spot there, you see, we like to say it's the eyes, the ears, and then the brains of Spot. So with Spot, people are starting to adopt Spot, just to help go from a reactive maintenance to a predictive maintenance. I think a lot of people are trying to do that with their digitization and their industry 4.0 standards and initiatives. So Spot helps you automate the inspection of critical assets. We're constantly trying to make sure that planes don't go down. So this is just such an important piece of it. Spot's going to go around and gather acoustic, thermal, visual data very reliably too. It's going to do the same thing every single time. It's going to notify the maintenance teams, whether that's through the Orbit software or through your own management software. Spot is able to notify your employees and your facility of what needs to be done and what might show as an anomaly. And Spot can do all this navigating slippery surfaces, going up and down stairs and just operating in active work sites. Going into the inspections we have available today, you can see the five here. So laser scanning, most frequently people are using this to just help create digital twins. Whether it's a digital image of a kiln or maybe their facility, that's what they're utilizing laser scanning for today. With visual, we do everything from checking gauges to looking for puddles or doors open, those types of things throughout your facility. Going to thermal, obviously, we're checking for changes in temperature, whether it's a pump, whether it's a motor, bearings, kilns, whatever it might be. That's what Spot's doing for the thermal analysis. Acoustic leak detection-- so we're looking for things like compressed air leaks. We're looking for-- maybe you have hydrogen in your plant. So we can try and find hydrogen leaks or nitrogen leaks. So any sort of acoustic leak or air leak, compressed air leak, spots are very good at detecting those. And then acoustic vibrometry-- so everything from rollers to bearings. Most often frequency changes are the P3, P4 anomalies. So before anything becomes too serious, but just being able to detect something really quickly and early on there helps to prevent. So from here I'm just going to go ahead and hand it over to John. He's going to go a bit more in depth of what he's seen in the field there and in the industries. - Yeah, thanks, Andrew. So I serve a lot of different industries around the world. Building materials is one of the big ones. And oftentimes when I jump into a customer-facing conversation, one of the first things I hear is, hey, your marketing videos online, they don't always reflect our environment. Our environment's dirty. It's wet. It's indoors. It's outdoors. It's very vertical. There's a lot of unstructured terrain. So I wanted to peel the onion back here and show some footage from some of our integrations around the world. And this is pretty typical for what we're seeing. So basements of mills, where they're standing water and their steam and their safety risks that are around you, this is a great opportunity to put the robot in that environment. If you're walking alongside industrial dryers or ovens and it's really hot and you've got terrain building up in different areas of your facility, you don't have to worry about that. The robot's got the ability to navigate autonomously through those types of unstructured terrains and still collect the data repeatedly and consistently. And really vertical sites too. So we see this a lot in preheat towers. We see this in finish mills. We see this in wallboard manufacturers. We see this in sawmills, paper mills. I mean, it's really-- in every industry, there's going to be sets of stairs. And typically what I hear from customers is there is a time investment in having to walk up sets of stairs to go and collect data. The motors, the gearboxes, whatever assets are on those floors, they're important to get to. But we're just taking away time from people being able to actually fix things. So this is a great opportunity, again, have the robot walk up and down those stairs and collect that data. And we'll go to the next slide here. And so expanding on that. So it's one thing to see some videos. I think it's another thing to continue to show. These are the environments that the robots are going into on a daily basis. And so basements of facilities where there's just dull, dirty, dark, dangerous opportunities, again, where we can remove the person from that exposure and allow them to do higher value tasks with their time. There's also a lot of hydraulic oil on these floors. There's a lot of indoor and outdoor environments. There's a lot of dust in the air. And it's just-- moving equipment too, you can't really see in that bottom left photo, but there's front loaders coming in and out of that site on a recurring basis. So you just don't want people walking through these areas. You'd rather put a robot in these situations. When you're inside of a really dusty and hot environment, you can only spend so much time in there. You've got to get suited up often. And it's, again, just great opportunity when customers come to me and they say, hey, where are the different types of environments we can put this robot? I typically will pull up examples like this to say, hey, how does this match the environments that your people are operating in? What do they have to do in these areas? And how can we get that to transition over to the robot so those people can be doing higher value tasks? And this is like-- it's one thing to put the robot in these areas. I think it's another thing that whenever I walk these sites, it's really fascinating because I walked about 43 plants last year with Spot. A lot of these are pulled from some of those site visits. And oftentimes I'll get on a visit. And there's a couple people there that they know everything about the plant. They've been there for maybe 30 years. And then the rest of the group is relatively newer, or they're in this middle phase where they're getting started and maybe they're in charge of a new asset group. And as we walk the site, the guy that's been there or the gal that's been there for 30-plus years, they hear something. And they go, hey, that sounds off. And then someone peels off of that visit. Or they put their hand on a motor just at random. But for them, it's something that they've done for decades. And they go, hey, this feels hot. This is the type of experience that our customers get to see through the eyes and ears of the robot and the sensors that live on it. So as most people walk through a plant, it's really hard to tell by sound if something's leaking because it's a really noisy environment. And there's also frequency ranges that our ears just aren't going to hear. You can't really tell by putting your hand on a motor if a bearing's vibrating. And it's really hard to tell about hotspots unless you pull out a thermal imaging device. That picture on the right here, you're looking at a vacuum chamber, an oven for one of our customers. And you can see the differential there in the regions. It's kind of small. But what we're showing is that a hotspot is forming. And there's almost a 350 degrees delta between these two panels. And that's early detection. That's giving people time to plan and to schedule and then, like Andrew said earlier, to move from reactive to proactive. Top left is a bearing that was moving inside of a motor. And then in the middle there, you're seeing some air leaks, some compressed air and gas leaks that are coming out of a customer site. So one of the more exciting applications that customers have shared back with me recently is there's kilns operating in all different types of environments. So you've got kilns used in different ways across different sites. When I visit a lot of cement manufacturers, this is a key component of their facility. It's one of the most capital-intensive pieces, and it's the main processing element of their entire process. And then when you visit paper mills and you visit sawmills, they're using kilns as well. And so some of the applications that have been shared back with me is, hey, people have to remove this coating and they have to go in there and they have to take measurements. And they take measurements in different areas because humans are variable. We're not robots. We're not going to do the exact same thing every single time. So now we've introduced some variability to the way that we're collecting the data. But there's also a lot of safety risks with going inside of a kiln. And so you remove some of the guesswork. And you remove that person from the safety hazard. And I like to say-- customers talk to me about industry 4.0 and using fixed sensors and how does this complement with Spot and it being a roaming IoT device? At the end of the day, the reality is you're either guessing, or you're knowing. And that's something that's really resonated with a lot of my customers is they hear that you can either be collecting data and knowing what's going on. Or you can be doing a little bit of guesswork. And so when you're able to put the robot inside of a kiln, you know down to as much deep data as you want. In terms of a scan to understand the measurements inside of that kiln as an example. The other thing that customers say to me is, hey, I've been using the thermal camera. I've been using the ultrasonic sensor. And we've got all these missions running around our site. Can we go ahead and convert those missions over to autonomous scanning missions? The answer is yes. The way that you do that is in Orbit. And so we'll go ahead and share a little bit more here about Orbit. Orbit's your window into Spot's world. This is the software that comes with the robot. And this is something that all of our customers use in industrial inspection around the world. At a very high level, this is what allows you to do real-time insights from all of the data that you collect. This is what allows you to manage your fleet. And this is what allows you to take those insights and do some trend analysis. And then this is really critical piece. We use our API, our Orbit API, to connect into your business systems, your MES, your CMMS. This is what allows you to do things work order generation, data integration, reporting, and a slew of other things. A lot of customers say, hey, I need to take the data from the dog and put it into my data lake. Orbit's the channel to be able to do that through. And there's a lot more to Orbit too. So we're going to play a quick video here. And we'll talk through some of these different things that you can do with the software. And this is a big piece of what we're doing here in the future as well. I don't know if the-- pulling up here. So I'll go ahead and keep talking. But the key features that weren't on that last slide is that you have the ability to manage your entire fleet. So if the video pulls up here, what you would see is you'd see a screen. And you'd be able to see the map of Orbit. And you'd be able to see the Spots walking through that area of the map. You can assume control of those robots, so you can teleoperate the robots from anywhere with a keyboard and an internet connection if you have access to the Orbit software. So it's your eyes and ears on the ground like I have customers right now that they set up remote mission control centers around the world. We'll take an example here in the States. And they have facilities, global facilities. Well, some of these facilities are going to be pretty lean staffed. And some of that's by design. Some of that's also just by the constraints of the world that we live in today. And so what they've told me is, hey, by design, in this example, we have a really lean-staffed group, but we want to be able to have eyes and ears on the ground. So they use our Orbit software with the robot to go do all of their checks. And if they see something, they can assume control of the robot and start teleoperating it. And they can go back to that asset through Orbit, and they can take additional data. And then they can decide, hey, does this warrant me sending a work order to have a human go and actually verify or fix this? I had an example where a customer logged into Orbit because they got an anomaly the night before. They teleoperated the robot from their house, instead of having to drive three hours in on a Saturday morning. And they actually prevented a pretty significant unplanned downtime event. And so all of your scheduling, all of your teleoperation, all of your fleet control it's all done through Orbit. And the other thing too is when I say we keep adding to it, there's going to be some exciting updates. I would say keep an eye on our YouTube the marketing team is going to be releasing. This is just the beginning for where we're going with Orbit. There's a lot of functionality and features coming in the future. - Oh, there we go. Now the video. [LAUGHS] - Here's the video. This is cool because there's probably stuff that I missed. Every time I talk about Orbit and I leave the room, I'm like, oh, I could have said this. Yeah. So outside of your scheduling, you can do all your mission planning. This is where you can get your display alerts. This is where you can get your email alerts. Here's an example of an email alert. And so customers ask me, how can I go ahead and get alerted of the anomalies? Again, that happens through Orbit. This is also where you can do some of that trend and that analysis. So you can plug in all of your historical data on an asset. You can click on that asset in the map and see all of the scans that the robot's done. You can connect it to all of your data systems. It's available on-prem or in the cloud. And this is really one of the most exciting features that brings the solution full circle. Glad that played. - Yeah, me too. [LAUGHS] Yeah, let's take a quick break. We have so many questions. We're definitely not going to get to all of them in this break. And I've tried to handpick a couple that popped up, a couple quick ones. In your earlier video, there was a blue light on the robot, and someone asked what that was. So that's a customers integration of a safety light system. John, do you want to talk to maybe our AV system and then how customers maybe have brought in their own things to match their own requirements? - Yeah, it's a great question. So one of the things that we're doing at Boston Dynamics is we're constantly updating and upgrading the hardware and the software here. And so with the largest fully integrated global fleet around the world, we've got a lot of customers that share feedback with us. One of the pieces of feedback over those years was, hey, we need our team on the ground to be able to understand what the robot's doing without having to interact with the operators or understand meaning. Are there visual and audible things that we can look at and talk to people about so they have a good human-robot interaction? So we incorporated lights and sounds into the robot as part of that feedback loop. And so there's a light now that indicates the robot in normal operating procedure. There's a light that allows operators to understand if it's collecting data. And then there's lights and sounds for when the robot goes up and down the stairs. And this is the same type of interaction as where if you hear a forklift coming around and there's that blue light. Or maybe you hear them honking as they come around the corner. The people on these sites, because they're dynamic, it just keeps their head up, keeps them a little bit more alert. And it increases that experience. And I'll just add one point too. People have said to us about stairs and the robot. So one of the cool things now that the robot does with this system is that it will actually look up and take a picture of the stairs before it actually approaches to make sure that there's no people on the stairs. So these little nuances come out from customer feedback. It's a great example of that AV system. - Yeah, I think maybe some of the digital twin scanning might have spawned this question, but can this integrate with Matterport scanners? And I think the official payload that we're working with is the Leica BLC ARC. And there's other payloads that people have used in the past. But most of our customers are definitely gravitating to that because of our partnership. But I don't know if John or Andrew, you have more to add there. - Yeah, I can speak to that. I mean, in general, Spot is a general purpose robot platform itself. So technically, the answer is yes. But when it comes to BD-certified payloads, the one that we're offering right now is the Leica BLC. - Yep, and there's another question too of, do we have to use your payloads? And as Andrew said, there's an API. It's an open platform. You can integrate any sensor that you want on there. But all that customization, all that built-in inspections in Orbit, all that's going to work with those BD-certified payloads. What is the minimum temperature that Spot can work in? We have industrial freezers that could be an area for inspections. - I can take that one. So the robot spec, that negative 20 Celsius, I think it's about 4 degrees, negative 4 degrees Fahrenheit. I'm not good on the math. And then the upper range is 55 Celsius. So it's about 135 to 140 Fahrenheit. And that's the temperature range that the robot's going to be living in. So the battery is spec at 90 minutes. Most of the customers will run it for an hour, charge it for an hour. And so can you go in colder or hotter environments? Sure. You got to work with our team to help go through that integration process so we can mitigate risk. But the spec is based on the robot living in that environment. And it touches on the point earlier of this is another example where our customers came back to us over the years and said, hey, we need you to increase the upper temperature range. And so we introduced a new heat sink into the robot. And the last thing I'll say is I'll just plug on that last question. We have an open API. So this is a really big thing that customers bring up to me in the site is if they have the wherewithal to be able to communicate and work with software engineers and talk to the API and their systems or hardware as an API as well, that's one point. And I would just plug the Boston Dynamics Consulting Group. This is something that we've spun up. If we have more of a project than a product, that's a great opportunity to work with that group. - Cool. Can Spot navigate through interior building, scanning, et cetera without any remote oversight from an operator? Can you set it up to scan building 1, room 10, for example, every Saturday without human intervention? So that speaks to what we were demonstrating in that video with Orbit on scheduling. Do you want to talk a little bit more about that? Maybe Andrew. - Yeah, I mean, I think you did a good job covering it. But as you may have seen from the video that was shown there, you're able to schedule Spot to do various inspections at various times. So you can really set it up to do timed missions or just have variable missions in that manner. So absolutely, the whole reason to have Spot is to have that automation in place. When it comes to spot actually working in your environments, again, dull, dirty, dangerous. But it gets really dull just seeing the robot dog walk around because nobody has to be doing anything with it. So it should be alleviating some of the pains of having somebody go and do these inspections for you. - Exactly. There are a couple of off-topic questions that I just wanted to call out. So there were a lot of questions about use for nuclear decommissioning, that sort of thing. We're not really covering that today in this content, but we've got tons of content about that on the website. So we definitely urge you to check that out. And we're happy to follow up and learn a little bit more about your application because Spot is used for many applications. And we're really focusing on one today. And then to that end, someone asked, seems BD has pivoted from security applications to industrial applications. Is that a safe assumption? It is not. Again, Spot is used in many different ways where we've got a great partner that works on a lot of those security applications. We've got a great webinar from past September with a customer about that application. So I'd definitely invite you to check that content out, and we're happy to talk more about that. Feel free to reach out following the webinar on those applications. Just to get us back on topic. We started this Q&A break on what is that blue light and that customer feedback, I guess. Just to segue back in, what's new? People are looking at our roadmap and what's coming and what's been coming. Is that a good way to get back into our material here? - Let's do it. - Let's do it. All right. And of course, there's 80 open questions now in the Q&A. We'll have time at the end. So stay tuned. Thank you. - Yeah, great segue here. So John had alluded to it earlier about how we determine our product roadmap and how we develop and make our solution better. So Spot has been rapidly developing. Spot was really only commercially available in the last 5, 5 and 1/2 years. So September 2019 is when Spot commercially became available. And then May 2020 is when we had autonomous navigation. We had the Spot cam payload. So going to the thermal and optical zoom. But then we started expanding more and more features. So you can see the growth here. And on top of security, on top of nuclear, we also saw the growth of industrial inspection with such a huge piece and a value add for a lot of these customers. So we started developing a lot more things like acoustic leak inspection, crosswalk actions. So all of this is just to say BD really focuses our product roadmap and what we're developing to create the full solution. Really, I talk about that EAM, Enterprise Asset Management, package before. What we're constantly building and developing is what the customer needs to continue seeing more and more value out of Spot. And the customers are the ones driving the roadmap. So going into the next slides, you'll see some more of the advancements and just the robot itself. So somebody had asked about stairs earlier. So by default-- again, John had mentioned this-- Spot is able to detect stairways. And then he's able to look up and see a look up or down and see, is there somebody on the staircase already or something in the staircase already that would put me or the operator in a difficult position where I'm going up or down the stairs? So Spot can wait until that obstacle is cleared. And it adjusts skate to handle them. We get a lot of questions of, well, I have graded stairs or open-riser stairs. Can spot handle those? I mean, you can see from these videos that the answer is yes. So really, any OSHA-certified stairs, Spot is able to traverse and navigate those. I'm sure there's a lot of questions about it. But if you have any specific questions, feel free just to reach out. And we can answer them one-on-one here. Going into slippery floors. You can see here the videos left and right. So the left is the before. The right is afterwards. Spot has been subject to all sorts of different environments. So ice, mud, beer, water, soap, all kinds of slippery products. So we, again, through the rapid development utilizing reinforcement learning, utilizing machine learning, we've gotten to the point now where Spot can automatically adjust to slippery floors, making sure that it's consistently balancing. On the left one there, it looks like a baby deer that had just been born. And on the right there, you can see just how well Spot has learned to navigate these areas. So keeping three points of contact on the ground at all points, smaller step heights, making sure that it has a lighter step and a lighter gait there. And of course, we even changed up the foot tread just to make sure that we're able to operate in all these environments. Some other cool things is just moving object detection. This goes back into us operating more and more and more environments. So you can see here, Spot's able to detect things. So if you have a cart or a trolley that's maybe run away or even just somebody who's pushing it and for some reason, it doesn't notice spot there, Spot will go ahead and stop and detect that moving object and just make sure that there's no collision there. It's able to do crosswalk identification. And the video that John showed-- or maybe the picture John showed earlier was where a lot of front loaders are going back and forth. So we go into places where there's a lot of pit equipment or material handling equipment. And we know that the human interaction with material handling equipment is usually the human will stop, look both ways. and then cross once it's safe to go. And Spot does that the exact same way and mimics that exact same thing, where they'll go ahead and stop at a crosswalk, make sure that nothing's coming from either direction, and then continue on once it's safe to be deemed so. And then we had already touched on the audiovisual warning systems where you can see that Spot has various colors there. The reason we adopted blue as the standard was because that's what the usual mobile equipment is. So any type of material handling, any type of forklift that you see out there usually has a blue light strapped on it. Now, just to show the direction of travel. So Spot can mimic that as well. But we've also got a few other color palettes, whether it's green, amber, white, red. And then the auditory system is just to make sure that people can hear it. So crosswalks, things like that Spot will be doing giving the signals. And what's really cool as well is as we've been in more real-world environments. We've seen more and more different types of hazards we've had to avoid whether it's just an extension cord draped across somewhere, ladders. It's in the Spot's path, random carts. Really, anything that you can think of, Spot has been subject to at this point. So we've just-- again, through reinforcement learning, through machine learning, we've gotten better and better and better at navigating around things. So Spot is seeing the world around it. And it's providing a visual semantic context to its navigation system. So it's just able to identify whatever might be in its way and find ways around it when navigating autonomously. All this to say that, we have been out there. And we've seen it. Spot delivers value at scale. So we've got over 1,500 robots in the world operating in 150-plus industrial customers. So we're up to 500,000 inspections per quarter at this point. We've got 2.7 million inspections. This is speaking towards the Spot is not a science project. It's not something that you just get to-- it's not something just to have around and be a cool solution. I think we've all seen projects not work out. Or I like to call them robot graveyards. You get one robot. And you just get to see it, and then it just sits in the corner until somebody wants to come play with it. Spot is just delivering real-world value at this point. And customers are scaling and continuing to scale with Spot as this industrial inspection tool and this asset that's being utilized around the world. - All right, John, want to talk about Ash Grove and some case studies and bring us home? - Yeah, absolutely. Thanks, Andrew. Ash Grove is a special customer of mine. It's been a really exciting experience with them. So they're part of the CRH Global Building Materials Manufacturing Company. And when you first visit a cement plant or a plant that's similar in terms of process, you immediately get pulled into what I've heard customers refer to as a pretty antagonistic environment. You've got really hot areas. You've got dangerous areas. You've got cement dust. So people spray down themselves. They spray down their cars. And electronics in the past have-- it's been a reliability concern. And then they're also very vertical. They're very spread out. There's dozens of kilometers of conveyance at any of these sites. They're indoors, outdoors. There's a lot of moving equipment, depending on what's happening at that site. So Ash Grove had to tackle all of these challenges with our team right out of the gate. And so we started working with them last year. We deployed and integrated the robot. And towards the end of last year and after the first 120 days of running, the robots now started to do some pretty spectacular things for them. So they're utilizing all of Spot's sensors and payloads on the robot. They're collecting thermography across lots of areas in their mill. They're doing acoustic inspections, where they have found air leaks and gas leaks. And they're also using acoustic vibrometry. So they've already detected a failing bearing in the finish mill that could have brought down the finish mill. They had a thermal event that was pretty significant as well. And the outcome, what we hear back from the customers, hey, this is improving safety across the entire site and it's freeing up skilled teams. And the other thing too is Andrew showed a slide about the hazard avoidance and the semantic understanding. I'll plug here the AI visual inspections that we're doing and starting to do at scale. They're finding value. Ash Grove is finding value from being able to understand what it's taking a picture of and if there's a change in that environment. And so when we mentioned things like MSHA and OSHA, this isn't just an Ash Grove conversation. This is all of our customers where they bring up safety audits and they bring up opportunities to tell us if things are happening in their site that shouldn't be happening. Can the robot take a picture and understand? The answer is yes. And so it's really cool to see this stuff come out in a real-life case study. And so we wanted to talk about more than just Ash Grove and the cement. And I'll plug the video again. Andy and the team, they are phenomenal to work with. But there's a video online. And there's a blog post as well. If we shift, this is a paper and packaging customer. And so when we travel around to these plants, it's a team. So there's solutions engineers that I work with on a daily basis that come with us to these sites. And this is a global paper customer. So they've got over a dozen mills in North America. And they've got dozens more around the world and about 100 converting sites. So this is a customer that we recently visited. And we talk about-- when we come and visit-- so how do we get to where we are today? You might be sitting in the audience. And you say, hey, I saw the Ash Grove video. It's really compelling. How do I go to on that journey with Boston Dynamics? This is the next couple slides is we're going to walk you through from start to finish how we walk a customer through that process. There's a technical evaluation, and there's an integration process. But at a high level, when we come on site and we come with our team, we're going to spend a day with you and the department leads, the decision makers. We're going to have the robot. And we're going to go into those areas that we've mapped out together virtually before we show up on site. For paper and packaging, oftentimes, what they tell me is it's the paper machines. It's the highest downtime cost per hour. It's the choke point. If that goes down, the pulp mill, the converting, the packaging, everything else is held to the paper machines or the tissue machines. So they're a really critical asset. And they're very capital intensive. They often tell us that the machine itself is going to have fixed IoT sensors. But the basement, all of the surrounding critical infrastructure, those are still being covered manually by routes and rounds. The mezzanine floors are still being covered by routes and rounds, which are hot and require stairs to go walk and take thermals on motors. And so we zoom out of where we want to start. And then we say-- we look at a pulp mill, for instance. Incredibly high density of assets in a pulp mill. It does not make economic sense to sensorize an instrument every single asset. So they cover the majority of these assets still manually. Great opportunity for the robot to go in there and collect a lot of data. Utilities and converting, this follows that trail of if we're looking at a fully integrated mill, these are some of the areas that continue to pop up from my customers. So if you're not in a paper and packaging mill, just think about with me here, where is the areas where you would want to start at your site and then expansion opportunities from there? You can go to the next slide. So this technical evaluation, it's a big conversation for a customer. When I show up on site, sometimes I'm showing up. And it's not as much now, Thanks to the folks in marketing. But some people have never seen this robot before. And so a lot of my job is coming on site and helping talk through a lot of, hey, here's my mill or here's my plant, here's the problems that we're facing. How can we integrate this technology? And how can we solve these problems? And what that boils down to is, what are the types of inspections that we want to take? What are the things that people are doing today? What are all the things that people don't have time to do? What are all the things that we want to be doing but we're just simply not doing for a variety of reasons? We document all of those. We also dig a little bit deeper with the customer to say, hey, why do we want to start here? Why is this inspection critical to you? This is where we can pull out a lot of the benefit to the organization. Customers will say to me after I leave a site, I need help to communicate the value of everything we learned today up the food chain. I need to be able to show leadership what we're doing here and why we're doing it. And we need to be able to use site-level data to help propel that conversation. That's a collaboration, and it starts with this technical evaluation. And so when we leave this site-- this site, in particular, they've got thousands of inspections for Spot. And they continue to grow. So we go and visit them, and then we integrate the robot. And so you saw in the last slide, there's paper finishing. There's utilities. There's pulping. There's chemical recovery. Not on this picture, but there's typically converting and packaging, utility rooms. So we break down those rooms. And we figure out, OK, where do we want to put the docks? A lot of customers are using multiple docking stations because their sites are 70, 80 acres, maybe indoor or outdoor, both. They're not going to be able to capture all the data they want with a single dock or a single dog. And so we map out, where do where do we want to create these missions? What are the areas we want to get to? And then we break it down within each of those areas to, what are the types of inspections? And so often hear from customers, hey, everything that we're doing thermal on, we should be doing vibration on as well. So let's go ahead and start adding up all of these inspection types. The AI visual on the gauge as well-- so customers that if you're still reading a gauge manually and it hasn't been instrumented, that's something that can be automated very easily with the robot. If there's housekeeping items like puddles on the floor, hydraulic oil, not just a housekeeping item, but if I'm at a sawmill and we're in a pump room and we're pumping hydraulic oil into the press, if 500-degree hydraulic oil starts to form in one of those pump rooms, that's a pretty big deal. And so does a person have time to go into that pump room every single day x number of times a day? No. Can we take a picture with Spot and understand if a puddle is forming? Absolutely. It's a great example. Same goes for the gas leaks and the steam leaks. Plant managers often tell me, my two most expensive utilities at a mill are air and steam. So let's put all of that energy back into your process with Spot. And go on to the next slide. I talked a little bit about it, but the technical evaluation is an artifact that we hand the customer. And it talks a lot about the missions, the routes, the starting point, the expansion opportunities. But really, what matters to the customer is, how do I justify this to my peers? And so we map all of that back to the impact and benefit. This is a collaboration between Boston Dynamics and the customer. We have to work with you and the historical data. One of the things that I found out throughout this process is if that's a challenge, then you're probably still living in that guessing bucket. And you're not necessarily knowing. And that's a great opportunity for us to shift from that reactive state to the proactive state. And we can actually start quantifying things with data. But we need to get a data collection device in there as soon as possible to start creating that baseline and that benchmark on these assets. If we have that data, great. Customers can tell me, hey, this is how much this area is going to cost if it goes down. These are my choke points. These are my bad actors. These are the things that people are constantly having to fix. We take all of those conversations. And we put them into a business justification. Regardless of the industry that I'm serving or the site that I'm visiting or the team that I'm working with, these are the most common buckets that customers come back to me with. We're reducing unplanned downtime events. We're being able to catch things early and often. So we give an early detection system to the boots on the ground, which allows them to repair things versus replace things. There's time savings in all of this. And then there's energy savings as well. I think one of the most impactful things that I've seen is sustainability in general is just a very-- it's a very critical in a lot of the industries I serve. So steel and cement, for instance. Carbon emissions is a very high-level conversation, a global conversation. How do we reduce those? It comes up in every single site visit I'm in. And so recently, when we were looking back at one of our customers where they've gone from site 1 to, I think, they're going into 20 or so sites, but they've deployed in about 10 of them. They started tracking the carbon back. And we've saved them millions of pounds in carbon. And so you can make an impact on the gross line yield. You make an impact on the dollars and cents. But when you hear stuff about sustainability and safety, these are things that we like to really talk about because they're things that customers bring to us. The last point I'll make on here-- and this is an important point for Boston Dynamics-- is that the dog is really like the gateway. So the dog's been out for five years. The dog allows you to understand and introduce a new species into the workplace. You're going to create SOPs. You're going to have a bunch of safety conversations, IT networking, digital back-end conversations about, how do we get an autonomous workforce really integrated with our human workforce? What does this open the door to in the future? It opens the door to working with Stretch, our warehouse platform. It opens the door to working with Atlas, our bipedal robot. This is where I talk to a lot of my really large enterprise customers is you can get immediate value today with Spot. You can set up a program today with Spot that is going to open the doorway to you to work with Boston Dynamics and future platforms as Orbit, our software, is robot agnostic. All right. I get really excited when I talk about this stuff. And so the seamless site integration, this is the last piece of the puzzles. The other thing that customers bring to me very often is, hey, this is great. We can justify the robot. We've got. Well under a two-year payback. On average, most of my customers are between 14 to 18 months ROI. There's going to be several that are in that sub-one year category. But then the next question shifts to a more tactical one of like, OK, so how do we actually walk through these steps together? How do I go from where I am today to having robots operating autonomously across not just 1 site, but maybe 5 or 10 sites? This is the playbook at a very high level. I often ask the customer, when's your timeline for implementation? When do you want dogs walking around your facility collecting data 24/7? Great, let's work backwards from that date. We did our tech eval. We've got our business justification. Now we can kick off the project with our integration team. There's a lot of change management conversations that are happening along the way. We're training operators on how to work with the hardware and the software. And our customer success team is mapping not just the site, but the corporate team and making sure everyone stays aligned throughout the project. The robot arrives on site. We start integrating that site. We run a site acceptance testing period. This is the customer's way of keeping Boston Dynamics and our team accountable, and then we hand it off into operations. The great thing about what you're seeing on the screen is it's a great tool to copy and paste across your network. And we're just going to wrap up here on one of my favorite slides. But putting a ribbon on everything that we've talked about is we want to-- customers at a very high level-- the dog is innovative. They talk to me about how amazing it is when it shows up on site. The reality is what it's doing around the world is it's avoiding unplanned downtime. It's approving a plan efficiency. And it's increasing plant safety. And it's doing that by autonomously and consistently and reliably collecting data for the customer on the things that they care about. And so now that unlocks the customer's ability to really do more with the data and do more with what we are really good at. So we can go around and we can fix things. We can problem solve. These are all things that are high-level tasks. And we can do that because we get some time back based on all the data that the dog's capturing. So thank you. I don't know, Vetche, Andrew, if you want to add anything to this. - No, I think we've got a ton of questions. So I definitely want to leave time to go through some of those. I think your mention of Atlas spawned some questions that I wanted to just jump in front of. So the first question was, is Spot the only commercially available robot? So from Boston Dynamics, the commercial robots that are available right now are Spot and Stretch. So Stretch is a warehouse robot. You can learn about that on our website. We're not going to be going too deeply into Stretch right now. And Atlas, right now we're focused on applications that support our primary investor Hyundai. So that is not commercially available today. But that's something we'll be talking about down the road. But basically, a lot of the feedback that John was alluding to is customers are trying to get ready for this truly mobile robotic future. And they see all of the challenges and the uphill battle that is in a lot of these older sites. And they're solving a lot of those problems now with what's available in the market now, which is Spot is there. We have tons of customers that are giving us that feedback of value. So that's what John was talking about. To that end, I guess, a couple slides back, we talked about integration. And there was a customer that came in of, are all sales of Spot engineered? Or are folks ever able to receive Spot and set it up themselves? And I've got an anecdote here, but I don't know if John or Andrew you want to give one, but I could start it off. Maybe you can add, but my favorite one is Michelin. So Michelin is another great case study. I'm going to keep plugging our case studies Michelin talks about in their video how they bought Spot; got training here on site, here at our facility; and then basically went back to their plant and got up and running all on their own. And the robot was running for months before they finally reached out and said, we'd love more help and then got us to come down. But for them, it was really important to prove out the technology without our sales engineering. But then after that time realized they could take it to a different level, just because our team has so much experience. What's the line that you kept saying, John? We have we've integrated more of these than anyone. So people recognize that and want to hit the ground running and are leveraging our services there. But if you guys want to add. - I can just add a tidbit, and then I'll kick it over to Andrew. I would just say, the way that most of my customers think about the integration services is an insurance policy for success. They have accountable metrics to key performance indicators that we set with them. It's essentially ensuring that the project's going to be successful. When you introduce new technology, walking technology into a site, there are eyes and ears on it. And you will have attention. And it's a good thing to how you use that attention early on. So our integration team is world class. Thousands of integrations and 40-plus countries. And so they come on site. And they work with your team and collaboration to make sure that the project gets off on the right start. It's just really key when we invest in a new technology early on, being able to show that it's working and it's successful. It's the way that I've seen customers think about the integration services. - Yeah, not much more to add to that besides, it's the easiest way to get started, but not the only way to get started. I'll say that. Yeah. - Good follow-on. How long does it take to train Spot in a facility typically? I don't know what-- - I mean, technically, you can get Spot up and running in minutes. It's really not that difficult. But our team, our integration team does have specific tools that they use and best practices, the likes of Michelin, unfortunately, didn't have right away. So you can have an autonomous walk running, again, within minutes. To be a little bit more value add, it really just depends on the facility and site by site. But our integrations usually take about a couple of weeks, a few weeks, just really depending on how many assets we're going to be looking at, what we're doing, and the area that we have to traverse. As you can imagine, it takes a lot longer to cover 3 and 1/2 million square feet as it does to cover 20,000 square feet. So there's a lot of variability in that answer. - Where do you store all the data collected from the thermal sensor, scanners, et cetera? Cloud or standalone hardware? - So we offer both or Orbit and the industrial inspection solution. So if we think about the hardware or the software, the training, the services, we typically talk about things as from a solution standpoint. And so specifically to the software, it depends on the customer's requirements. Some people want a physical server on site. Some people want to utilize our cloud. And there's some exciting stuff coming in the future to continue to expand the offering that we can give our customers. But that's the most common answer is we look to them to say, hey, how do you want to store that data? How can we fit within your infosec requirements? And then we collaborate with you on meeting those requirements. That's what I've seen across a variety of sites. - Great. What are routine maintenance requirements for Spot? What parts require scheduled replacement, that sort of thing? - It depends. It depends on the environment. So this week I was at a sawmill. And the requirements for a sawmill, if the environment's extremely dusty and the motors can't breathe because of the dust, then we're going to probably speak to that customer slightly differently than if we're in a semiconductor fab, where everything's clean, well lit, or a pharmaceutical plant. So some of this has to do with environmental risk and what are some of these standard SOPs that the customers operate in within their environment. But within Spot specific, the maintenance routine is extremely minimal. You've got five stereo vision cameras and lenses on the robot. We want to wipe those down to make sure that the robot's eyes and ears are all being able to see and listen. We want to make sure that we keep the robot in an area where its motors can breathe or its fans can breathe. So that gets back to the dust. The foot treads-- anything with a tread is going to wear over time. So think about the more that you run every day, the more you're going to have to replace your sneakers. Very similar situation with Spot. In the cement world specifically, I'll say a tidbit that a lot of my customers do is they'll just use compressed air, and they'll blow the robot down. So they'll either-- they'll make a little dog house, or they'll get the Spot station directly from us if they want that. And that keeps some of the debris from just landing on the docking station and the dock when it's charging. And then on top of that, they might blow it down a couple times a week. - Great. So since we're talking about dust, there were a couple of questions about Spot's tolerance to wood dust. And, obviously, you talked about some examples. But any more to add there? I think this customer has got a particle board plant. - I don't know, Andrew, what you would-- do you want to add something? Or do you want me to take this one? - Why don't you take it? And then if I have anything to add, I'll add. - I would just say we-- so the shortest answer is that we haven't had any kind of hardware reliability issues across any of the verticals that I serve, which wood products manufacturing and OSB particle dust. So the robots are water resistant and dust resistant. And then we work with the customer on the site to be able to mitigate risk if those exist. And some of this comes down to the planning of the routes, the rounds, the missions, the inspections of the robot. What vantage point do we want to collect data from? What room do we want to collect data from? How much time do we want to spend in these rooms if they're very antagonistic? All of these things are part of that conversation with the customer to figure out what's the best way that we can provide the highest level of performance. I'd also plug Spot Care. Right. Spot Care comes with the robot. And it's one of those-- it's a premium white glove service. And this is where-- if the robot is operating under normal conditions, this is where the customer can work with our support team to figure out if something was to happen. But those are the types of conversations we're having with customers. - Great. This one's for Andrew. How can we use robotics in mining applications? - Yeah so that's a good question. So we've got quite a bit of inquiries mining and what we actually do. So within mining operations, we target the process and the stockpile and loadout areas. So everything from underground. So like postblast inspection with digital twin or survey. Hazardous area entries, we can utilize Spot for. The crushing load-- so a lot of conveyance. You have miles and miles of conveyance that run from underground, above ground. There's a lot of dust particulate in there. So people will utilize Spot to do inspections of those conveyors. And then within the process plant themselves for ore refining. So there's a lot of rotating equipment thermography and vibrometry there. So we'll take a look in there. So the middle area there within the mining operations is where we generally-- you see Spot being utilized. - Great. Do you have any case studies of Spot being used for asset integrity inspection? We actually do. We've got a great case study of the Hamburg Bridge. It's on our website. Definitely check that out. Using a scanner to look at the integrity of the concrete in that bridge. It's a great one. Can Spot perform IR scans to electrical panels with IR windows? That's an interesting one. I've personally seen that. I think we actually tested that at CRH, John, if you remember. But yeah, I think height might be a challenge for some of those. But if the IR windows are low and the things you want to look at are at the right robot level and you can angle the camera through them, you could definitely perform that inspection. - I just add one thing to that. - Yeah, of course. - It just made me think of something earlier. So we talked a little bit about Orbit. We talked about making a map. Customers approach this in a lot of different ways. And I'll tie this together, I promise. But like MCC cabinets and taking thermal audits-- so there's changing thermal rules that customers share back with us about moving something from you should be doing this to you shall be doing this, for example. So these different organizations that provide regulatory conversations back to manufacturing plants will tell us, hey, thermal audits are changing. And this is now a requirement. It's going to deploy in the next year. How do we get in front of it? It's a great application for Spot. So once a year, people can take down their cabinets, open them up. And the robot, you can teach it a mission to be able to deploy x number of times a year to take thermal scans on all of those cabinets once they're open. And now the person does not have to spend all day doing it. Or now you don't have to hire an outside contractor to do that. The reason why I brought up the map is because when a customer maps out their entire facility, they can turn certain areas of that map on and off at their will. If they're doing construction in that part and they don't want the dog autonomously roaming because the environments change, great. Turn those missions off. If they have missions that they only want to run during downtime, planned downtime, or they only have missions that they want to run maybe once a year or twice a year to do something like thermal audits with all the cabinets open, that's another example of how we can capture that data. - Yeah, great. I think we're at time, but I want to close out with one more question/statement. So there was a question about what are some of the planned future capabilities of Spot. And I think, John, you alluded to some of the stuff that we've been working on. And we'll definitely have a lot more to talk about in the next few months as related to our next software release, 5.0. So we really appreciate everyone's time here. Thank you so much. There's definitely a ton of questions we didn't get to, but we will be following up through email. If you are interested-- or I'd see one question coming in about can Spot monitor an industrial laundry plant? And then there's another one about water treatment facilities. Yes, any of these industrial areas, Spot is great. And so definitely raise your hand. We're happy to have an additional conversation with you and find out a little bit more about your facility and make some recommendations. So thank you all for joining. Thank you, Andrew. Thank you John. And we'll be sending out a recording in the next few days. Thank you so much. - Thanks, everyone. - Thanks, everyone.