Radically Transform Your Revit Electrical Modeling with AI-Powered Solutions

SUNNY WATTS: Welcome to our course today. We are so excited to be with you and present a little bit about how AI will transform your Revit electric modeling.

So here's a quick agenda of what we're going to cover today. First, we'll do some introductions. We'll explain a little bit about who we are and why we're here. We're going to then talk a little bit about why AI is so important for construction. Then we get to walk you through the ins and outs of how our AI platform works. And finally, we're going to give you a look into where AI is currently at and where it's going in the future.

AARON SZYMANSKI: OK, so before we introduce ourselves, we'll introduce our company. So we're Augmenta, and we are a software company. We're trying to build a new generation of design software for the construction industry. Now, this is software that builds on top of Revit and the foundation of Revit. And what it does is it actually automates the design of core building systems, like your MEP systems, the things that, as we all know, are behind the drywall. People don't see them when the building is done. But that's really what drives the cost and complexity of building projects.

My name is Aaron Szymanski. I'm a co-founder of Augmenta and the chief product officer. A bit about my background. So I originally was working in industrial design. I used to design phones and tablets at BlackBerry. I then moved into digital product design where I worked on a number of projects. But the most interesting, I think, for this presentation is I worked and led the foundational design work for Autodesk's first generative design program, which was called Project Dreamcatcher. Now I'm leading both product and design here at Augmenta.

SUNNY WATTS: And I am Sunny Watts. I have been an Autodesk trainer for a number of years, and I've had the pleasure of being able to bring new technologies to companies. And I'm really excited to be able to help bring AI into the construction industry.

AARON SZYMANSKI: OK. So before we go into showing how our technology works, let's talk about why we're even here in the first place. So a lot of the people on our team don't come from the world of construction. I mean, Sunny does, but I don't originally, and a number of the engineers on our team don't as well. So why have we chosen to focus our efforts here? Why do we think that AI really matters for construction?

So to start off, if we can go ahead. Construction really needs this. So construction is the second largest industry on the planet. It's probably the most important in terms of the impact that it can have on human life in general. But this is an industry that is facing a number of really significant challenges.

These are, of course, challenges that make it unique compared to other industries that it's often compared to, like manufacturing. But even taking that into consideration, looking at it objectively, construction really is not a very efficient industry. The average construction project sees a huge amount of rework and material waste. The labor shortage is only getting worse. Materials are only getting more and more expensive. All this is leading to even more and more risk for effectively every party involved.

So there's obviously an opportunity here, but where is it that AI should be applied to help out with this?

So we believe that the design process is really where we should be focusing. This is really the key to solving many of these challenges. It's this design process that translates the initial requirements at the start of a project into the actual reality of that project. It's what drives all the decisions around how a building is designed, how it functions, how it's going to look, how it's going to perform. And then within it, how the various systems in it are going to be routed, modeled, procured, and installed.

And even today, the process of designing buildings is essentially still manual. There's a lot of inefficiencies and mistakes in that process that end up leading then to rework and waste and risk in the field.

So let's look at that design process a little bit more closely. Because even today's design process, even what we consider to be the current state of the art, doesn't yet do enough to solve these problems. And of course, BIM is a really big deal. The introduction of BIM was a huge improvement over doing everything entirely in CAD and 2D. Even before that, doing everything with pen and paper and pencil.

But today, you still have on any given project dozens of engineers, modelers. They're spending weeks or even months toiling over a building design. They're working off various assumptions and best guesses, and they're trying to do their best to create these very accurate digital maps, quote unquote, that we call BIM models. And the value of that work and the value of the model is without question. It's super valuable, but the process of producing them is still too slow and expensive.

And this is not anyone's fault. It's just that building projects are just getting so complex, timelines are getting so tight, and they just involve so many parties that we're really at the limits of what human beings can grapple with on their own. So improving design is probably the most powerful change that we can make and one of really the biggest levers that we have to reduce time, reduce uncertainty, rework, and waste. And these are things that we have to do today.

So let's talk about AI. AI, of course, is on everyone's agenda, everyone's mind for very good reason. But one thing that we want to start to dive into is that in fact, there's critical differences between different forms of AI. There's different types, and they have different pros and cons.

So for example, something like ChatGPT, which is, of course, on everybody's mind, is very good at processing and then making sense of and interpolating between huge amounts of information. And this allows it, for example, to process text in a really interesting and surprisingly coherent way.

But because of that, or despite that, it still produces inconsistent and uncertain output that you can't really trust. If you have played with ChatGPT yourself, you'll realize it's very good at some things, but in others it sort of falls short, and it doesn't seem to be able to tell the difference between those two. So now try to take that approach and apply it to construction, the work that you're doing. Can you really trust something like that?

But there's other approaches. There's other tools or technologies, for example, like AlphaFold. So AlphaFold is a Google DeepMind project. It's an AI system. It's used to predict protein structures, how proteins will fold. And it's already doing very valuable work in the medical field. It's a system that from the ground up has actually been built to produce designs or predictions that you can actually act on. You can go and actually try these out and apply them to solve real-world problems.

So what we're trying to do is build AlphaFold for buildings. It's called the Augmented Construction Platform, or ACP for short. And in this workflow, what you're doing is as an engineer, as a modeler, you're not drawing designs manually. You're instead defining requirements of the problem that you want to solve, and then you are using our system to generate detailed, coordinated, efficient designs in hours instead of days or weeks.

So it saves a ton of time, reduces a lot of risk, and allows you to work with a level of insight and certainty that you can't really using a traditional process.

This is a very hard problem to solve. We'll talk a bit about some of the things that we've run into and learned along the way. It's not producing a poem or a pretty picture, as you can imagine, but the value of this is that we really can give people superpowers. We can let them do things that they really could never do before.

So our very first product is built for electrical engineers and the VDC teams inside electrical contractors. It automates the routing and the modeling of electrical raceway systems. And of course, we have even within this product a very long and ambitious roadmap. We want to do it all. But right now, where we're focused with the first version is really taking you from a model that has no conduit populated to something that is routed and clash free that you can bring into your first or second coordination meeting and doing that very quickly.

So it takes in a set of data points. Sunny is going to walk you through what that actually looks like in practice. And then it then produces not just one but a number of different designs that you can compare and contrast. Effectively, what we're really trying to do is automate a lot of that upfront VDC process.

So what I'm going to do is pass things over to Sunny now to talk about and to show you how this actually works.

SUNNY WATTS: OK. Thank you so much, Aaron. So now that you know a little bit about why AI is so important for the construction process and you've been introduced to our platform a bit, I get to actually show you that platform and give you a high-level walkthrough of how it works.

So first, a little bit about the platform. So the Augmented Construction Platform, or ACP, leverages real-world electrical expertise, incorporating current codes and construction standards to efficiently route and optimize feeder and home run conduit configurations within your project.

So it does this by looking at that Revit model data to find space that it has available, so areas that aren't already taken up by other objects. And it then uses that information to create clash-free pathways to connect a source and a destination that can actually be supported within our project using our conduit.

So this process is created using studies, and that allows the user to adjust the input once and evaluate those inputs and suggest multiple solutions for the conduit to be routed. And because multiple studies can be run concurrently, users can easily explore how various constraints may impact their routing conduit almost simultaneously. So let's dive into what that actually looks like in practice.

All right. So the very first thing that we need to do is provide ACP with our various input data. So it can't run conduit if it doesn't know where everything is. And so we do that with the Revit model information as well as some schedule information. So we'll start by importing an electrical model from Revit to ACP using the Augmenta add-in inside of Revit. That electrical model data tells the application the locations of each element within the model that we're going to route the conduits between.

So I'm just going to go ahead and open up my Revit. This is a sample project that we've taken, and I'm just going to open up my data import. So this is a project that we've taken and modified in order to test and thus also showcase what ACP is capable of doing.

And in this particular project, there's actually three different levels. I've cut it down so that we're only looking at the two so it's a little bit easier to see what's going on. Inside of this project, we have various equipment. So we've got this switchboard here, we've got some panels, we've got some transformers. So everything that's highlighted in red to make it a little bit easier to see. That's all of our equipment.

And then we also have some devices throughout. So these little blue boxes that you're seeing, those are the final home run locations. So we're going to route from panel to panel. We're also going to route from the panels to those home run locations as well.

So how do I actually get this model information into ACP? Well, if you see in my Revit, you'll notice I have an Augmenta tab added to my Revit. And from there, I can launch the application. And that's going to actually connect the Revit model to that web-based platform, allowing me to work inside it.

So I've already launched it. I'm already in the project. So for today, I've created this AU sample project. And I have a couple of those studies that I mentioned. We're going to first look at just this study that's not quite complete yet.

So in here, you can see I have an opportunity to import in my design inputs. And I can either use a previously imported one or I can import a new one if I make changes to my model. And that's really the only time that I'm going to do that. So I can bring in that electrical model data.

So again, that allows me to locate where all of the equipment is. I also need a way to identify which equipment, a source, and a destination for each of those conduits that I'm going to run in my project. And I do that with my input schedule. So this is where I would bring in my input schedules.

So let's take a look at those input schedules. They're actually an Excel file. So if I open that up. So we have two input schedules. The first input schedule is called a conduit schedule. And in the conduit schedule, you can see I have a run ID number. That just identifies the row of information. And then within that row of information, I can specify my source and my destination. And I do that just based off of my panel name inside of my Revit project.

After I've specified my source and my destination name, I can also give it a system. That system information is just a text parameter, and it does a couple of things for me. One thing is when I push my information from ACP back into Revit, so when I bring in my conduit solution back into Revit, it'll pull that information in. I can use that to organize my schedule. I can use it to create view filters. All of that. I can also organize the heights of my conduit based off of my system. So this is an important piece of information.

The last piece of information I have in my conduit schedule is the feeder ID. And the feeder ID number, this number that you see here, actually corresponds with my second schedule. So my second schedule is my feeder schedule. And inside of that feeder schedule, each row here has a unique feeder ID, and that then identifies for me the number of conduit runs that I'm going to run between my source and destination, as well as the size of that conduit.

So for example, every time that I see this 400-311 in my conduit schedule, I'm going to get two runs of conduit that are three inches each. So with all of that information, now I know not just where my source and my destination is, but also what I need to connect between the two.

So in addition to my electrical model and my input schedules, I also have site geometry or background geometry. That background geometry is uploaded from the exact same Revit file that my electrical model is imported from. And it does a couple of things for me as well.

So the very first thing that it's going to do is actually establish supportability throughout my building. So where can I actually support a conduit rack from? Can I hang it from the floor above? Am I going to support it off of a wall? How am I actually going to build that rack? So it looks at architectural and structural elements to see where there are actual structures in place that I can support the rack off of.

The second thing that the background geometry does for me is it actually gives me all of the collision geometry. So that's going to show me where there's something in the way, something that I need to route my conduit around.

So by establishing all of those things, this is how we're able to create fully supportable and clash-free conduit throughout the entire building. And just like with my electrical model, my site geometry, I can choose to import each time or I can just import when I have adjustments. So that makes my process a little bit faster. I don't have to change my model every time that I want to look at different options.

[VIDEO PLAYBACK]

- [STUTTERING] Ah, input. More input.

- No problem.

[END PLAYBACK]

SUNNY WATTS: OK. So we can provide ACP with more input with user controls that allow us to constrain the conduit routing in the project. So this is really important, because even though ACP uses real-world electrical knowledge and the Revit model information, each project is unique and it has its own requirements and its own unique needs.

So there are two main user input controls. The first one is the routing controls. And those exist inside of our actual Revit model. They're things that we can control inside of that Revit environment. The second are design rules, and those will be specified directly in ACP. So I'm going to show you both of those. I'm going to start in the Revit project. So let's take a look at those design tools.

So in my Augmenta tab-- I'm going to open up my view here. So in my Augmenta tab, you can see I have this panel for routing tools. I'm going to focus on these three tools today. So we have a May Route and a Keep Out box. And you're seeing that here in my 3D view. So see these boxes that are kind of scattered throughout. They are just a generic model box that we can adjust the size of, and they tell ACP something really important.

So first of all, we have a May Route box. So we know that throughout our entire model, our Revit model, we're going to have a lot of obstructions. One of those obstructions, for example, is the floor. And in order for me to be able to route conduit through the floor, I need to have a hole. So this box actually allows for ACP to run through areas that it's placed over where there would normally be obstructions.

So again, in this example, it's where my floor is going through, and I need my conduit to be able to penetrate in that area. So I've put it right above my switchboard so that the conduits that need to go to the other levels can actually get to the other levels. So that's the May Route box. It basically punches a hole.

The Keep Out box is kind of the opposite of that. Essentially, it allows me to block out space that normally looks like it's big, wide, and open, where I might want to run conduit, but I know I really don't.

So for example, I'll place those over stairwells. So you can see I've got a bunch of stairwells here. I also have an elevator shaft. ACP loves shafts. Elevator shaft is not where we want conduits to run, so I'm going to place my Keep Out box over that.

The other thing that the Keep Out box actually allows me to do is place an element within that space. And the conduit for that single element, whether it's a panel or a device, can actually penetrate that space. None of the other conduit can but just the one conduit for that specific element. So it doesn't block that. So again, elevator shafts, I'll have a motor, and I want to be able to connect to it.

I'll also place it in areas like this lobby that extends between levels. Again, it looks nice, big, and open, but I do not want conduit there. So that's going to allow me to block that out.

Now, the third tool of the routing tools that I want to talk about are Preferred Zones. And you actually are not seeing them here in this 3D view. And that is because they are applied to spaces, and spaces aren't visible in 3D. So we're going to look at a 2D plan so that you can see that.

So here in my first-level floor plan, you can see I have a number of spaces that are colored green. And they are my preferred routing. Essentially, what I'm telling ACP is if you can route conduit in these spaces, that's where I want you to go first. Like, this is my preferred.

So you can see that even though I want my conduit to run in some of these other rooms, they don't necessarily have to be preferred routing. They're going to be routed to regardless. But this is really where I want my main rack to run. So I'm able to designate that and give that signal to ACP. So it's going to really prefer to run in there.

OK, so those are the user controls inside of Revit. Now I'm going to go back to ACP. So back to my study that we were looking at previously. And you'll see at the bottom I have these design rules. So the design rules allow me to specify specific constraints regarding some different settings.

So first, I have my support. So if you remember, I talked about how in the background geometry we're able to find the different surfaces that we can support off of. With my supports, I can actually say exactly which surfaces I want it to look at based off of what supports I've specced out for my project. I can give it clearances around various objects, or an additional buffer where needed. I can also specify my spacing between my conduit. And I love this because it's actually face to face instead of center to center. Just saying. I love that.

I also can designate a raceway height. And if you remember in the schedule when we created our different systems, I can actually do that by the system that's coming directly from that Excel file that I've imported into my project. I can decide that I want to also route through the walls.

So one of the things you'll notice in my Revit project is I don't have a May Route box every single time that I need to punch a hole in the wall. I'm actually going to let ACP decide where to penetrate those walls. And I can do that by switching on that little control.

I can also tell it that it can penetrate ceilings. So instead of having to create a May Route box in every ceiling, it can penetrate through that. We added that one in part because for some reason, electrical rooms have ceilings now. I don't know.

The other thing about our design rules is that they can apply to the entire site, like my raceway height is. Or it can apply to specific scopes. And I can control that with both scope boxes that I create in my project or spaces. So I have a lot of flexibility of where I'm applying what rules and which constraints.

OK. So keep in mind that our designs and our solutions will always be adjusted based off of all of those inputs that we've created.

So this is where it gets fun. Once the design inputs, background geometry or site geometry, and design rules are entered, the Augmenta application will analyze all of that information to generate multiple routing solutions based off of those inputs and constraints. Those are then available to be viewed and analyzed by the viewer-- or by the user, sorry-- prior to pushing any of those solutions back into Revit.

So inside of the platform there are actually three different ways that I can view my solutions. There is a table view, which shows me a bunch of metrics that I can look at. I've got a chart view, and that gives me a visual of two different metrics compared. And we'll look at all of those. And then I have the model viewer. So let's take a look at those. They're kind of fun.

So going back to my ACP. I'm going to go to a study that I've already generated against. And you can see that with each study, it returns 12 different solutions. And within those solutions, now within that table viewer, I can see all of those different metrics. So I can look at it and see how much conduit is each solution using. How many bends are there? Things like, did I actually route to everything I was trying to route to? What's the construction cost estimated at? How much time do we think this is going to take? And we can start to make that comparison.

If you want something a little bit more visual, you can go into the chart view. And you can see right now, it's looking at estimated cost versus the construction time. And that kind of makes sense, that trends upward. But it changes if I change it from estimated construction time to quantity of bends. That's a little different. Does that change which solution I like more?

And then finally, once I've decided which solutions I want to review, I can either click on that from here, or I can click on it from the table view, and I can go into the model viewer. So I'm just going to select one of the solutions here. I'm going to click on this, number six. And you can see it's generating here. It just takes a couple of seconds. It's pretty fast.

So once it's generated inside of my model viewer, I can start to analyze that. And there are several different viewing tools that we have in here. I can just zoom in to the model. I can fly through the model, which is kind of fun. I can fly through the whole thing. I can come all the way out of the model, and I can turn off the site geometry, that background model information, and look at it as a whole and start to analyze it that way.

I can even zoom in and turn that background geometry back on. So that makes it very easy.

Another thing that I really love is I can compare solution against solution. So I'm zoomed in to this area. Maybe I want to see number six versus number three, and it will allow me to do that. So lots of different ways that I can view, compare, and analyze that solution right inside of the viewer.

And once I am ready-- so once I've decided on that optimal conduit solution for the project, I can easily export that solution from ACP back into the Revit project. And I'm going to do that as system conduit. And then I can start to look at it a little bit deeper. I can do further analysis. I can start to edit it. I can start to detail it out. So I'm going to show you what that looks like as well.

So because I am in ACP connected through my Revit, I can literally just be in the solution I want to export and go up to this Export button to Revit as system conduit, and it's going to push it back into my Revit project. I've actually already done that, because it takes a few minutes. So I've already done that for us today.

And there it is. So there is our conduit solution back into our project. And if I click on it, you'll see it is just system conduit. So from there, I can do all of the same things with that conduit now that I could do with my normal system conduit. I can create view filters. I can create schedules. I can start modifying all of those different things.

OK. So at this present time, ACP content still requires some human adjustments to be ready for the final coordination and installation on site. So this is actually a snapshot of that same solution that I just showed you in that sample file. And that's on the left-hand side. And then on the right-hand side, that same solution has been modified by our partner company EngBIM for construction.

So you can see already how some of the ACP output on the left needed to be changed for that final installation requirements. Once that happens, now we can use that for our coordination. Also, we're ready to have it prefabbed or built on site.

Now Aaron is going to take us a little bit through where ACP is today and looking into the future.

AARON SZYMANSKI: OK. Thanks, Sunny. And thanks for that overview. So what I want to do is wrap up the presentation today with a couple of different topics that I'm sure you all find pretty interesting.

So, so far we've been talking a lot about the current state of the art, where things are today in terms of applying AI to the design process. Hopefully we've given you a good picture of that, warts and all of really where things are and what we're capable of today.

But I want to start off by then showing you what it actually then looks like to do this not on a sample project, but on a real-world project. What does it look like to apply ACP to a real project with a real schedule?

But then after that, I'm going to step back a bit. I want to try and give you a bit of a broader perspective. We'll talk about our own experiences as a team, as a company, trying to apply AI to this problem, what we've learned along the way, the challenges that we've run into and been able to overcome, and then try and wrap up with a feeling of the trajectory that this technology is on so you can start to think about what the future might look like.

All right. So to start off, let's talk about applying this project or this technology to a real project. So this technology is not theoretical. It's already being used. This is actually one of our most recent case studies with one of our early design partners, CNR Electric. So they used ACP to automatically populate over 11,000 feet of conduit on a elementary school project. And they're currently in the process of running it on yet another project as we speak.

So here's an example of some of those edits that Sunny mentioned. So Travis over at CNR used ACP to take an empty Revit model-- empty of conduit, that is-- and populate the vast majority of the conduit in that model, and then went and made some edits by hand.

So like Sunny noted before, we're still not producing 100% complete designs. So for example, you can see that the elevation of [INAUDIBLE] track was changed. Of course, we've got controls that you can apply to do that now. Later on in the project as well, additional conduit was added. So there's still a coordination process that happens even after we're doing that initial population for you.

So the photo on the right was after a couple of those rounds. So it's not a direct comparison of before and after, but still illustrates fairly well, I think, the state of the technology today and hopefully the value that it can deliver as well.

So with that, let's step back a little bit now and talk about what we've learned through this process of actually applying AI to construction. As it turns out, it is a very, very hard problem. It's really hard to use AI to accelerate even a fraction of the expertise that everyone on this call and the presentation has built up over your careers.

So first off, from a purely computational perspective, routing and coordinating conduit, probably even more than other systems, is really challenging. Our systems have to consider a large number of variables. So height and spacing requirements, different support types and requirements, hold point requirements, different varying band radii, and the effect that those will have on coordination and the availability of space in labor costs are a factor in all those decisions.

So when you're designing this kind of a system, every time you add a variable, you are increasing the range of possible solutions that you have to explore and filter through exponentially.

And what makes all of this even harder is that it's construction. We're building real buildings. The bar for acceptable quality is very high.

So think about applying AI to image generation. Think about back when Midjourney was still adding an extra finger or two to people's hands in the images. But that's fine. Even when it didn't work that well, the output was still really interesting, still really valuable in a lot of different applications.

But now imagine you're trying to model a rack or you're trying to coordinate a rack, and suddenly you notice that a bunch of your conduit have just disappeared halfway through or a bunch of new conduit have been added halfway through. Suddenly, using the tool starts to feel like a bit of a liability. You need to now go and spend the time to review and rethink the entire layout of your rack.

So what that means for us as someone trying to build software for this industry is that that threshold of what we internally call correctness is very high. It's very high in construction and in engineering. And this means that for us, it takes years of R&D to produce something that even starts to become useful. Even that minimum viable product takes a lot of work and thinking behind it.

And last is data is messy. And I'm sure that you all deal with this too in the Revit models that you work with. Of course, not the models that you create yourself, but of course, the ones that you get from everyone else. The ones you create are perfect.

So you have slab that's modeled as a ceiling element. You have dots placed as placeholder masses alongside actual ducts that are not placeholder masses and then are hidden in various length models. There's some creative use of Revit categories.

So all this stuff, generally speaking, it's hard enough for you to deal with on your own. But it's something that you can work around. You generally know the intent of those decisions in that model. But then think about the process of trying to deal with that programmatically, trying to interpret all that data while still preserving the intent that was behind it. So again, all this adds to the complexity of this problem.

But despite all those challenges, we are making really good progress. And I actually want to suggest that maybe this is one of the most important slides in this whole presentation. So in the top left, we see ACP's output on a hospital project. This was something that we ran with a design partner back in November of 2023.

And maybe it's a bit hard to see on the screen, but it barely worked at all. So you can even see, if you trace the conduit, it's leaving the electrical room, it's punching outside of the building, immediately looping back, kind of doing a bit of a staircase pattern before shooting down the length of the building, probably because it thought that was a more efficient way to arrive at some destination.

As a design, it's useless. It's spaghetti. And at that point in time, it was very much the case that using ACP on a project would probably take longer to do than to just simply model the whole thing yourself. And keep in mind, and you can see this in the model, that's with mechanical and plumbing hidden. It's not even trying to coordinate around anything other than the beams in that model.

But then look at where we got to in August of this year running on that same project. So now, we're producing fully coordinated routes. Mechanical and plumbing is fully turned on. It's fully coordinated against those models in a very dense hospital project.

Of course, it's still not perfect. The design still needs some cleanup and editing. But we're now at a point where we're averaging about 50% active hours producing and editing and ACP design compared to doing that same project manually. So you're about twice as productive. And this is only 10 months later.

So now think about what that trajectory looks like. Think about what this implies. Think about where we might be in another 10 months.

So in terms of where we want to be in the next 10 months, electrical is really just our first step. So right now, of course we're working on electrical. That's where we're focused. That's what we showed you today. We're going to continue to evolve that. But we're also in the design phases for our future products. We're going to be working on mechanical and plumbing.

And the vision is not just to automate the design of those systems the same way we are doing here for electrical, but to also automate the coordination of those systems. So everyone, every party is specifying those requirements, and we're producing those designs that are automatically designed and coordinated against each other to optimize for the overall project, to reduce the overall schedule, the overall inefficiency, or to rather maximize the efficiency of that system overall.

So for anyone listening on this call who is maybe from some of those other trades, we definitely want to talk to you. If you're interested in working with us, please reach out.

So hopefully you found that presentation interesting. And I want to pick up on what I closed the last slide with. And that is how can you try this out if it's something you want to try for yourself? So we've got a couple of ways to get involved. Our product right now is not yet ready for direct purchase, but we have different pathways that we've entered the market in and are ready to serve the market with.

So the first is if you're looking to apply ACP to get the benefits of AI, of ACP on a live project, really the best way to do this is through a partnership that we've launched with EngBIM. So Eng is the largest BIM consulting firm in the United States. We're working with them to deliver AI-powered BIM consulting services. So with this partnership, we're able to deliver really high quality models in significantly less time for your projects.

If you are a little bit more exploratory, you want to try this out yourself, see how it works and work with us to evolve it, then please reach out to us. There's some contact information there on the screen.

We have our design partner program. In this program, you will have access to the tool at no cost. We'll train you on it. You can try applying it to your own live projects. Of course, in return, all we ask for is your feedback as we continue to evolve the product.

And if you want to do both, that's absolutely an option for you as well. So please reach out if that's something that you're interested in. Otherwise, I'll pass it back to Sunny. And thank you.

SUNNY WATTS: OK. Thank you so much for being with us today. And that is the end of our presentation.