Plant of the Future: Unlocking Full Facility Insights with AI

Good morning. If you're just joining us, thanks for being here to see everything that's new with Spot and Orbit. Some quick housekeeping before we get started. We'll be sending out a recording of today's presentation following the webinar. So if you have any questions during the webinar, you can use the q and a button. We'll be going through some of those both in the middle of the presentation today and at the end. So go ahead and start asking them early. We'll jump on them as soon as we can. And with that, I'm going to turn it over to our presenters. Amanda, would you like to introduce yourself? Sure. Hi, everyone. It's great to be here. I'm Amanda Gonano, and I'm the associate director of product for the Orbit platform. Gordon, do you wanna go next? Yeah. Hey, everyone. My name is Gordon Finnie. I'm a software engineering manager here at Boston Dynamics, leading a team that helps customers get insights from data. Good morning, everyone. My name is Dan Zuba. I'm a key account manager here at Boston Dynamics. My main focus is helping, many of our strategic customers, see the value in Spot and ultimately helping them scale it across their enterprise. So again we're really excited about, this release and, webinar today. It's a little different from ones in the past because, this one is really focused on how can we accelerate our customers with the growth as well as the value that they're seeing out of Spot. So, just wanted to quickly, you know, highlight a few of the, really meaningful points that we we we found and and what some of our customers are, as well as, cover a few of the different inspections of what customers are seeing value today as we we talk about some of these new releases. In terms of just review, it's a, again, really exciting time as we now have had over fifteen hundred robots deployed all across the world. This continues to grow at a at a pretty rapid place, and we continue to see excitement around this. Inspection points, we're we're averaging, over half a million a quarter, and each quarter over quarter that continues to grow with over two point seven million inspection points in total. And right now, we're we're we're just over a hundred and fifty in industrial inspection customers, you know, all over the the globe, whether it's in the US, in Europe, Brazil, as well as the APAC region. Next slide, please. So one of the things I wanted to to really start with is is, you know, how are our customers seeing value? This this may be review for some of you, or this may be the first time you've actually, you know, joined a webinar. But some of the things that we found for our our customers that focus on on what is readily available, on the truck. They see a significant amount of value, very quickly. And so that's really where we we start and focus on these four inspection types that you're seeing here on the screen. Whether it's using our PTZ camera to collect visual data, that we have some customers that are leveraging, different types of machine learning models, whether it's ones that we offer or whether they're building their own, own, models in house or using the the thermal, capabilities on that PTZ camera, looking for different types of pumps and motors and bearings that are overheating, you know, ultimately trying to to get insights on their factory, that they currently either aren't getting today or not getting as frequently enough, as well as using the the the fluke sensor that we we have here, where we're listening for the different types of compressed air and gas leaks. And then as a result of our our last big release, last year, we we we recently did update that to to include the capabilities around acoustic vibration analysis. So when our customers really focus on what is available and what we know works, they're they're able to move very quickly. And then it really changes the discussion in terms of not just what does this one robot look like, but how do we ultimately scale. Right? And this is a big component of what this this big release is is all about. When we look at this, this is where our customers are seeing value, and this is this is what's exciting. In all of the releases leading up to this day, it's really been how can we, you know, figure out where our customers are gonna find value. And it's been it's it's coming to be very clear in these four buckets, that we see across all of our customers. So whether it's focusing on reducing the amount of downtime, whether it's the energy savings, you know, from listening for those leaks that we're we're currently seeing from from all the compressed air, whether it's the time savings of not having to have your technicians focus on these, rounds and readings but can focus on a much more meaningful task, or whether it's ultimately reducing the the the the, breakdown reduction. All of this story has really been put together, to provide a significant amount of value for our customers. On average, most of them are seeing an ROI in less than two years, and we have a handful of customers that are starting to see this value actually less than twelve months, and that's what's really exciting for us. The one other thing that I would call out that's not on here, and this is kind of kinda tease up around a lot of what this this big release is focused on. And I always mention this when I'm talking talking to my customers is there's really a fifth bucket, and it's around around the data that you're capturing. Many of our customers that are taking this data that they are collecting on a consistent, repeatable day basis and integrating it into their systems and starting to use that to ultimately move from descriptive analytics to predictive and prescriptive analytics, and how they can start to focus, you know, much more of their routine maintenance, around and much more scheduled as around to, what they've always done in the past. And all of this story puts together to show that this robot continues to be more and more valuable. Go on. Next slide. So whether you've been with us from the beginning all the way back, September of twenty nineteen, so here we are, you know, coming up on six years of having this this robot, you know, released out there in into the wild. And in the beginning, you know, we just knew that this was gonna be an amazing data capture device, but we didn't quite know exactly where where it would ultimately go. And as you see through these releases along the way, whether it was back in twenty twenty when we first introduced, the ability for the robot to to traverse auto autonomously, whether it was adding the spot camera as a payload or the, the thermal capabilities that were increased or the acoustic ability to be able to to to now start to listen to things. Each one of these has set up from a a position to where now we really feel that we have a valuable product. Clearly, customers are seeing this. But now we've really have moved from a world where customers are looking, oh, can I start to get one robot, and what does this look like? To where we have multiple customers that have that have now, purchased and deployed, whether it's ten, twenty, thirty, even fifty and above. And so all of that, those insights that we've been been been gaining from our customers along this journey has really put this all together and has really been, kind of the, the impetus to this this release here because it's all focused on how can we now enable customers to, one, scale with the enterprise features that, the team is gonna talk about here. And two, we're not finished with the value that the robot is providing. So how can we now use the power of AI in adding that to the robot to ultimately exponentially increase the the the value through a variety of different inspections? So with that, I'll hand it over to Amanda to, walk through some of those features. Thanks, Dan. Like Dan just said and as we saw across that timeline, we didn't start out with Orbit and Spot together. Orbit was launched last year in twenty twenty four, really predominantly focused on fleet management and data analysis so that our industrial customers could get value and use this software to be able to enable their robots to do what they needed to do on-site. Today, we're continuing to invest in those similar themes. Now we're not thinking about fleet management just at a single site with Orbit. But as Dan said, how are you expanding that across a robotics fleet through your entire enterprise? We're increasing the types of data analysis you can do in Orbit with the new site We know that many of you, when you walk onto your site floor, don't just see a single asset or a single line. You're really seeing all the different visual information that's coming together to help you understand, is there a bunch of debris piled up somewhere? Does my site look like it's operating as efficiently as it could be? And so we're putting that additional visual data, catalog and information into Orbit so that you can monitor that remotely and also get all that rich data out and into your third party systems. We're gonna talk about the specific features that Thanks, Amanda. So Thanks, Amanda. So when we think about Spot in Orbit, we think about discrete inspection points across a user's site. They're capturing data about individual assets. But what like Amanda said, we've kind of learned through these deployments that capturing the environments in which the assets are located and how those change over time is equally as important. And that's why in five point zero, we're introducing Site View, which is a historical, immersive, and navigable view of your site to help you get insights into evolving site conditions. So let's take a look. So site maps now have site view where you can choose a waypoint to drop into your site. And when you do this, you end up inside of a three sixty image taken by Spot during a site view mission. In 5.0, any mission can be converted into a site view mission scheduled to start collecting panels. So you can look around. You can also click on the floor to navigate to other places that Spot has been and find points of interest such as this door. And because this, mission was captured on a schedule, we have a time series at each of these locations. We can go back and find anomalous conditions like when that door was open. So site view will be available out of the box in 5.0 and provides a new comprehensive view of your site and how it changes. You can go to the next slide. So we just use site view to remotely navigate our site and and, find this anomalous condition, which was the door being opened. Next step is to author inspection of that door and track it over time to make sure that doesn't happen again. And customers are scaling today by finding many use cases just like this. And at that point, what becomes important is helping make the inspection authoring process seamless. And that's why in 5.0, we're making it even easier to add your next one hundred or one thousand inspections by supporting inspection authoring directly from site view, meaning that you don't need to be on the factory floor to author. So let's use site view authoring to track that external door directly. So here in the inspection editor, we can also drop into site view. And when we do, we can look around and what we see are our inspection points and the exact location where they're being captured. So here we have a thermal. We also have a PTZ, and we can use this as an intuitive way to find inspections and edit them. We can also use the split screen blueprint on the bottom to quickly navigate around our site. Here we're jumping back to the door. We're finding it. And at this point, adding inspection is as easy as clicking this plus button and clicking create PTZ. Now we can position our frame and zoom in to make sure that door is where we want it to be in this image. And once we give it a good name, this is an inspection which can be scheduled and repeated by Spot. So while we're at it, we can kind of notice, you know, down here by this column, there's a puddle. And it's easy for us to quickly add another PTZ inspection, which will capture this image over time and maybe we want to track how that develops. So in addition to PTZs, thermal inspections are also supported. So up here we have this vent, which we may want to track the thermal conditions of, And we can author a thermal inspection in the same way, attach thresholds to it and start getting alerts. So authoring in Site View is also launching in 5.0. And we think this is a great way to increase the speed at which you scale up your inspections. Go to the next slide. This is this is an exciting one. So this is AI visual inspections. So like we just did, other customers have thousands of repeated discrete inspection points. They're collecting pictures of doors, puddles, and and other points of interest. And at that point, extracting insights from that data becomes critical. And wouldn't be it'd be great if Orbit itself could tell you whether the door is open rather than a human needing to do the review. And in 5.0 Orbit can, We're introducing AI visual inspections, which use a large foundation model running locally on your instance to answer questions about images. So let's use AI VI or AI visual inspections to inspect whether that door is open and alert when it is. Go to the next slide. So we're starting from that PTZ we just authored. And if we scroll down to the bottom, we have a new option here to add a triggered AI inspection. And when we do that, we get a new inspection that will run anytime that image is captured, and and it has a simple question field. Our question is, is the door open? We're setting our answer type here to yes or no, so the model always knows to respond with those two values. We're going to alert when the answer is yes. Now the robot hasn't captured this door yet because we just altered it, but I wanna validate that this works. So to do so, I'm going to upload a couple images that I took of the door open and closed. We upload those as test cases, and we're going to run them. And we see that the model is responding correctly, yes or no. And we also see that we get an alert when the answer is yes. So in addition to yes-no answer types, we also can extract numerical information from screens like this digital temperature gauge. So here I have already set up an AI inspection to extract that information by asking a question about the region of that screen where the temperature value is. I have my answer type set to numeric, and I want to alert when that rises above a dangerous value. So I quickly set that up with a greater than seven hundred thirty six value, run my test cases, and you see that the model is reading from the region I want and alerting when that value goes above the threshold. So AIVI will be available on 5.0, and we think this is an intuitive way to help you get insights from data by leveraging these state of the art models. And that's it for me. As Gordon just went through, Orbit AI visual inspections really enables you to ask even more questions about your site. These are just some of the examples that we've seen customers who are already using this today find value. There's a huge amount of opportunity in environmental health and safety features. Is the door open? Those fire extinguishers, they tend to get up and walk away. You wanna know when they have when the things that are really critical to health and safety has changed. Debris pile up at the end of a line or boxes piling up in an area that needs to move. Security risks like doors being open or text reading are all things that are really common. But we also know there's so much more to explore. The amazing thing about this technology is that there's an infinite number of questions that could be asked. And we know that every site is unique and specific. So as you all go out and you start to try these features with Spot and Orbit, we're excited for you to continue to learn in more use cases and also really make this AI accessible to your team. As Gordon went through, it's very easy and intuitive to write these questions directly into orbit. And very quickly, you now have robotics and AI happening directly on your site floor. As we've gone through these features, we've talked about them on an individual site level. But so many of our customers are scaling beyond one site, two sites to really thinking about how robotics is applied across their entire enterprise with Spot and Orbit. So in order to make that simpler, we now allow you to connect multiple Orbit instances into a single enterprise view so that your site in Massachusetts and your site in California can be connected into a single place where you can see every Orbit site across your enterprise as well as every connected robot that you have. You can easily click into a single individual robot and look at to see what that robot can see, what they're inspecting at any given time. And then all of the rich insights about site health and robotic health are pulled into this enterprise dashboard so you can understand how your investment in robotics is working across individual sites. This is focused primarily for our enterprise admin users so that they can see all the rich information that they need to about individual sites. Individual sites still have access to their Orbit so that they can continue to expand Spot's capabilities on that local site and become their own experts in the data that Spot is collecting at that individual site. Great. I think with that, we're gonna break for a few questions. We've had a lot coming in about AI inspections and site view, so I'm gonna go through a few of those before we continue with the presentation. So first up, we've got a few questions about training the AI model and how that works. Courtney, I know you just touched on it, but would you mind going into a little bit more detail about how the the model is trained and what kind of special teaching it requires? Yeah. So we chose to use, an off the shelf pretrained model that has built in a lot of knowledge about various, visual circumstances. So right out of the box, we already have knowledge about, puddles, about doors being open or closed, about all the objects that Amanda pointed out from fire extinguishers to other things. And over time, this model we will work to improve this model. But at the moment, it does not require any special teaching. Out of the box, it can start answering these questions. Yeah. We've got one asking about, like, the test images and documentation that we use. Can you say a little bit more about how we give Orbit some images that it can use to validate the model? Yeah. So when you, are editing an AI inspection, you'll immediately see if the robot has taken that picture before. You'll immediately see test cases populated, which you can use to test your model. You can also select any image that the robot has taken, to use as a test case. But like I sort of demonstrated, you also can upload your own because a lot of times these anomalous conditions don't appear, and you don't want them to appear. So uploading an image is is your best route there. So this test case obviously is a critical part of offering these inspections, so we've invested effort to make that flexible. Great. Had a few questions about the different types of things that Spot can inspect with AI inspections. We gave a few examples here. But, Amanda, maybe could you, you know, talk to some of the things that we know work really well today and, what options you have if there are other types of inspections that you'd like to try at your facility? Sure. Like Gordon said, we are focused on sort of the yes no questions that you can answer, is an object present or not, as well as numerical questions that could be answered. Right? Simple counting, something that's easy to do for the human eye to or reading a screen to report back a number. When we are working with customers on-site, we often see this broken down and most successful when you are able to bring something into a, a question that is object presence or not. So we talked a lot about the environmental health and safety aspects. But things like, is a dumpster full? Right? Or is a particular piece of equipment in the right area are all kinds of inspections that we've been able to see before. Looking at puddles under a piece of equipment, looking at gas, or oil buildup under a piece of equipment are all different examples of things that we've seen over and over again at different sites. We've also seen things that are really site specific. So, you know, perhaps there's a material that's really specific to your site, whether it's a little metal bead or perhaps it's debris that ends up at the end of a line. That's another great question that we see lots of people employing too. So it is going to depend a little bit about your site and what's important to you. But our team is happy to help support questions that you might have. And we have a a great question and example, table on our support documentation too to help you get started when you're first getting deployed as well. Great. Kind of along those lines, we showed some examples of that water forming on the ground. Amanda, can you elaborate a little bit more on how, we could help qualify what's going on there? Someone's asking, you know, can you tell if that's just a puddle there or if it's actively dripping from the pipes, for example? Yeah. The way that it works today is that this is a discreet inspection. So you would set up in a particular area where you know a puddle may occur. When Spot goes to take that individual photo, we'll be asking a question about whatever it is that the photo comes back. So if it's on the floor and you're seeing the puddle continue to increase because that's where the inspection is put, that's kind of the question that you can frame it as. If there's a way visually to be able to see if a puddle is overfilling or if there's running water, that's another opportunity for you to be able to frame that question too. It really depends on the image that you're taking and the question that you're asking to be able to get exactly what you need based on the information that's most important to your site. I think we'll just do one more before we keep moving on, and we'll answer some more questions at the end. So if we didn't get your to yours already, stay tuned for that. But, Dan, I'm gonna, present this one to you since you were speaking about how customers are finding value with Spot. Someone is looking for a business case comparing robotic conversations you conversations you've had? Yeah. Absolutely. That's a that's a that's a great question. And and I would I would say we have a pretty defined process in how we help customers, put together that business case. So a lot of times when, this might be their their first journey with a with a quadruped robot and they're they're looking at, you know, we we know we want this technology. We we we know it's the future, but we don't quite understand, necessarily where the clear cut value is and and how we ultimately go through through deploying a robot like this. So part of our process this process actually is working, with our team. We we have sales engineers, as well as sales managers that will meet with you to discuss the different types of inspections that you're currently doing, the frequency. We have different types of tools that can help quantify the value of of different types of anomalies that could be present. We even go as far as as as going on-site and doing different types of technical evaluations, to help you build out what your business case could look like and to make sure that we've identified the right place, to deploy the robot to get started so that the, value is derived very quickly and that the project is set up successfully. You know, I would definitely recommend to anyone who, this is new to to reach out to us directly on on our on our website. And, we'd love to walk you through what that process is to get you from, the the discovery phase to actually having a successfully deployed robot and starting to plan out what expansion looks like throughout your enterprise. So with that, I'll give it back to you, Amanda, to talk about a few of the other things that are new in this release. Perfect. So we've talked a lot about our kind of visual inspection features, but we've also invested in a lot of the Orbit infrastructure across the board. One of the things that we've invested in is making another option for you to deploy Orbit. So today, there are three different ways that you can choose to deploy Orbit. The first is with a U rack mounted application that you install on premise that's been available since the beginning of Orbit. We also have an Orbit cloud version, which is hosted by Boston Dynamics, the SOC two type two certified. It is entirely managed by us, so there's minimal effort on your team to get this up and deployed. The third option that we are releasing as part of this five dot o features is a virtual machine version of Orbit. That's an OVA file type available today for VMware, Hyper v, and Azure. Across all three of these deployment types, there's an option for whatever level of IT and security that your team needs to be able to get started right away with Spot and Orbit. A number of additional features we've included are really focused on allowing you to be able to customize Orbit workflows that make the most sense for your team. One of those things is the ability to create custom user roles so that your data experts can get the information they need out of Orbit and your most advanced robot drivers and robot wranglers can also get all of the insights that they need. All of these custom user roles can tie into any native identity provider that you may already have to connect with Orbit and can all be set up by Orbit admin directly in the software. Next, we've been really focused on improving the way that you can control privacy for your teams as you roll out these new visual inspection features. It's really important that Spot and Orbit are a tool for your team that they see as a way to gather data and get the insights they need without fearing for how that data may be captured or how their images may be captured. We enable you to optionally turn on face blurring for any of the images that Spot captures through PTZ or through those site view images that we were showing earlier. You're able to turn that on directly in orbit, and it's up to you and your team to be able to turn on and use this feature as needed. Next, we're making it faster for you to be able to update software on the robot from orbit. So after five dot o, anytime we have a new software released, you just need to upload it to orbit and then you can push it out to the robot without actually having to physically be near the robot or any of the payloads, making it faster for you to get all these rich updates into your facility and out across your entire enterprise. Next, another way to make deployment even faster is now the ability to automatically set thermal thresholds. So when you're going to set up an initial thermal inspection, you can take a small sample of data to help you determine what the right threshold for a piece of equipment should be. This minimizes the time spent manually identifying and applying thresholds and gets you deployed faster and quicker with a little bit of intelligence built in. Finally, we have a beta code less work order integration feature. We know it's really important for you to get this data out of orbit and into your third party systems. We wanna make sure that it's easy for you to pull that into a work order management system or where ever this information needs to go so that the right person gets the right data and can take action as soon as possible. You don't always have the resources to be able to use APIs to build those connections into third party systems. And so this enables you and your team with minimal effort to be able to push that data into an external system. And we encourage you to reach out to our team if you are interested in using this beta feature. This is just a summary of all of the things that are in the five dot o release, but there's even more smaller features all available in our release notes on our website right now. We've talked about a number of rich and exciting new things that are coming out in five dot o. But as Dan showed in that timeline side, we've been investing in different capabilities for Spot and over orbit for years. And all of that is really so that you can have real value out of robotic inspections. And with this release, we're expanding not just for the maintenance team, but thinking about safety, environmental, and health teams who also are starting to get value out of this new data. We're creating features that make it really easy for you to get deployed so that you can see that value out of the box right away and then take that initial success at your first site and scale it across new facilities. And finally, we're moving beyond just a single asset world to gain full site visibility in a real site health context so that you can understand what's happening not just at one site, but at multiple sites across the country or across the globe. Spot and Orbit are just the beginning. Orbit is the intelligent automation platform that connects the physical world to the digital world for all of our Boston Dynamics robots. We're really excited to be able to bring you the most advanced robotics hardware in the world now connected to AI powered software that's gonna bring you more and more insights, and we're excited to continue on this robotic journey with all of you across all of our different products in the future. And with that, I'm gonna pass it back to Mikaela for a little bit more question and answer. Yeah. Thank you, everyone. We have a lot of questions, so we'll see how many we can get through today. And if we don't cover them live, someone will be reaching out to follow-up. So don't worry if, you don't hear your question today. Gordon, I'm gonna ask you one first. We have a question about Spot's best security qualities. Can you talk a little bit about Orbit and Spot and how we ensure that these are secure? Yeah. Definitely. So I'd say there's Orbit, there's Spot, but there's also Boston Dynamics as a whole. And in each of these, components, we'd really focus on, you know, secure data formats, secure data access patterns, secure protocols and wire formats. And all of those details are written up in a white paper that's available, on our website. And I would also say, recently sort of speaking to the organization as a whole, we achieved SOC two certification for our cloud orbit system, which is a third party test more or less, and kind of speaks to having secure systems that, you know, are secure over a period of time. So we we have that certification, and you can look at that white paper for details. Thank you. We have a good one here. What are the common challenges plans face when implementing AI solutions, and how can those challenges be mitigated? Maybe Dan and Amanda, you wanna tag team that one? Amanda, you want to start or you want me to start? I'm sorry. I can start. One of the things that we've historically heard is that it's really exciting to collect all this rich data, but you don't always have access at a maintenance team to a large data science team to be able to build your own models or build your AI on top of it. It also can be intimidated sometimes to bring something brand new into a facility where you're piloting tons of new technology. And And so one of the things that we've tried to do is make it really easy to get started with tools that you already have and put the AI directly into the software without needing to do any training or startup to get started. You can just drop in and start using it. So it doesn't feel like such of a burden for your team to have to start to build to get going. Yeah. And and what I would add to that is is kind of the the expectations around where AI works really well and and where it doesn't. You know, there's a whole host of of hype out there on on artificial intelligence and and the value that it brings. And so it's really important to make sure that you're you're focused on the right use cases, right off the bat. And you're using it, you know, in in my opinion, as a as a nice, parallel path while you're deploying the robot with with the the use cases that we know provide value. With this, AI BI, you know, we we we truly believe it's gonna open up extreme value, you know, for our customers, and it's managing those right expectations to be able to to to to go along that that journey. And I would also say, you know, just to echo the point, it is intimidating for for the customers that, you know, this is this is all new and this is, you know, they're they're really excited about it, but they don't really know where to start. Well well, here is a a tool that you can get started very, very quickly, you know, as Amanda mentioned without a team of data scientists. And then on the flip end for for our customers that have, you know, teams of data scientists that are are constantly asking, how can I get more and more data to create the models that we ultimately want? This is just gonna be an additional toolset that they're gonna be very excited about. Awesome. We obviously talked a lot in this presentation about AI inspections and safety and quality and some of the other things besides maintenance that Spot can cover. But we have some people who would like to hear some more about, Spot's predictive maintenance capabilities. Dan, you touched on it at the beginning, but do you wanna go into a little bit more detail about the types of inspection Spot can do on, industrial equipment? Yeah. Yeah. I mean, I I I think, you know, if you look at it, most customers right now have some type of, PM rounds that are are scheduled. And sometimes they're hit in terms of, the frequency that they want, or sometimes those people are just busy doing, you know, fixing, the mountain of of other problems. So it's it's really, you know, I would say important to start to be able to collect that consistent data regularly. You know, again, whether it falls into those buckets that that that I talked about a little bit earlier. Right? Taking visual inspections, taking thermal readings of pumps and bearings and motors and steam traps, listening for the different types of of anomalies, you know, that might be present in the facility, and starting to have that as a as a as a baseline so that they can think about how do we want to set this up going forward? Do we do we even need to to change the filter every year on a particular piece of equipment? That's what we've always done. But perhaps it actually says that we need to to change that out more frequently, to ultimately reduce those unplanned downtimes, which could be could be quite quite costly. So it's really it's really around moving from this this, you know, kind of task list around somebody needs to go around and do the things at some point in time when they're available to now not even have it to worry about that. The robot's gonna go do the thing that that we we we tell it to do. We're gonna teach it what to do. It'll consistently collect that data. And whether we're watching the temperature over time or whether we're collecting visual data, to do some of the features that, you know, Gordon had walked through. All of that together is gonna help customers move from describing what's happening in their environment to ultimately predicting what's happening in their environment. Awesome. Had a few different questions about integrations, SAP, other systems. Amanda, Gordon, do one of you wanna talk to how integrations with third party systems work? Sure. I'm happy I'm happy to start and then, Gordon, you can you can feel free to jump in. So we we today have three options that you can use to integrate with third party systems. One is the Orbit APIs, which are available publicly, and you can go to on our support site to see all the different, Orbit API SDK and ways that you can work with that. We also have webhooks that allow you to push information from Orbit into your other systems. That's something that we see a lot of our customers use when they wanna be able to say that as soon as an alert occurs, send that information into another system so that a work order can be created. And then the third is the new beta system that we talked about earlier, which is more of a codeless work order integration system, where in orbit, you can identify the types of information you want to push into a system and connect that directly with a work order system based on kind of a secure connection. We enable three different ways for you to do that because we know that everyone's team is different in the ways that they wanna be able to connect and get this information out of orbit. And we also know it's really important that this information doesn't sit in one place and become another data silo. So we really wanna be able to make sure that regardless of where your, team falls kind of on that spectrum of integration, that you're able to get that data out and into other systems. Great. We've got one about site view. So, Gordon, I'm gonna ask you this one. For site view missions, how first of all, how can you create a site view mission? What are our options there? And then, if you're creating the new AI inspections, are those at a precise location or is it looking across the facility for those visual anomalies? Sure. So on the SiteView side, as I said, any existing mission can be turned into a site view mission. So if you have a route that you like that you want to turn into, a series of panels, you can copy that and then turn that into a site view mission by checking a box. Once you've done that, that mission can be scheduled. And when it's executed, there's another parameter that you have control over, which is the cadence at which those panels are taken. And you can choose anything from every one waypoint to every five. And kinda bridging into that next question, AI inspections take place at specific waypoints. So, they do not, you know, ask questions about the site as a whole or change at the moment. So when you're authoring the AI inspection, you're choosing an individual waypoint, which you've already guaranteed as captured, with a PTZ inspection. And then, every time that PTZ inspection is executed, the AI question will also be great. I'm going to wrap us up with one more question about everyone's favorite topic. That's Atlas. Dan, maybe you can take this one. Are you planning to use Atlas for industrial inspection in the future or, like, just generally where where are we headed with our Atlas platform? Yes. That's, that is one of the most exciting, you know, or one of the the the topics that we do get asked, I would say daily. So, you know, with regards to Atlas, we've we've publicly stated that we are going to be, you know, offering our our first, proof of concept, with Hyundai here in in q three. And there's gonna be, you know, we're really excited about, you know, moving to that that kind of stage of this journey, and that's really gonna open up a a whole lot more. So, you know, there's there's a lot to come on that. But but here's the reality. Atlas is is definitely gonna be an amazing, you you know, humanoid that is gonna gonna gonna change the world. But quite quite candidly, you you can't start with something like a humanoid in your environment. There is a very defined process that we've learned over over the years on on how do you how do you go from not having any time to to have stationary assets to these assets around, like a quadruped that will literally get up and walk around your facility and collect data? How do you get your, teams that are working alongside, the robot to be comfortable with that? How do we get the IT teams and the the health and safety teams and everybody, used to this world around around what it's like to have robots, you know, inside of our our our environment. So, you know, the customers that begin this journey around around what Spot can do are the ones that are gonna be set up for success for when something like Atlas is available, to to the broader broader markets. You know, I use the quote from from one of the customers that I work very, very closely closely with. You can't go from zero to humanoid, and and that really is the truth. There's a a lot of lessons learned around deploying Spot. So, you know, again, this whole whole release that we've been talking about has has really been focused on how do we help our customers scale, because that's the reality. Customers are no longer looking to to deploy their first robot. Customers are looking to deploy deploy fleets of robots, around the globe. Amazing. On that note, we're just about out of time, so I'm gonna wrap us up. If we didn't get to your question, we'll follow-up, after this, to answer those, and you can catch the recording after this too. But you can also learn more on our website. We have a blog, release notes up about all of these five Dotto features, and you can always get in touch with our friendly sales team like Dan, on our website as well. So thank you everyone for joining, and we'll look forward to seeing you at the next one.