Using Automation to Unlock the True Potential of DPC | Clothing Tech | PI Apparel

So I, uh, I'll just introduce myself a little bit while people are coming in. Uh, my experience and background is in software development, mainly with regard to, um, uh, making production processes more efficient. So my— most of my experience has been in the automotive industry, um, some in the aerospace industry. I've been to many, uh, Embraer, Boeing, um Airbus factories and dozens and dozens of car factories around the world over the years. About— tell you a little bit about ourselves. So we are a joint venture partner with Hexagon AB, which is a Swedish technology firm, probably have never heard of. There are about 25,000 employees, so it's a decent-sized company selling technology to a variety of industries, but none to the fashion industry. And so About 6 years ago, we started looking at the fashion industry and said, what good could we do? How could we help? What improvements could we potentially make? And so we did a little work and talked to a number of brands and investigated quite a bit and decided to invest in clothing tech. And so you see in electronics, I've been to— Foxconn factories as well with Hexagon. There's a— where they make iPhones and the watch. This watch uses some Hexagon technology in the production process. So anyways, that's a little bit about me. We decided to take a fresh look at the process of creating clothing and how we go about it. And what we saw was pretty interesting. So we focused primarily on the digital product creation because seemed like the new thing at the time. It's now 2024, so it's, uh, 5 years later. And of course, most of you are already using some kind of 3D in your process, some kind of digital in your process. Um, and some of the things that we originally had goals for and, and whatnot are part of digital product creation. And so we'll talk a little bit about this, trying to do things. Um, our goal was to take what is done now in a what I call expert-to-expert, um, kind of way of working and to automate and to simplify in that process where possible and improve the speed. So, some of the things we looked at when we were looking at the fashion industry, and I did most of the research and investigation. So, I met with brands, I talked with— I did quite a bit of discussion, some technology providers as well. The industry is very late to 3D compared to any other industry. Most industries started converting their processes and digitizing their processes in the 1980s. So 3D CAD software became available in the 1980s. And most automotive, aerospace, electronics, civil engineering, architecture, construction, all of these industries started using CAD in the 1980s, and mainly in 3D. So very slow to adopt the new changes because the product is a little harder. Typical 3D CAD systems use a set set of mathematics that works quite well. I'm a mathematician, so this is something of interest to me. It may not be much to you, but the set of mathematics works quite well for rigid body shapes like car components. It doesn't work so well with draping and flexible components. And so it was— the technology out of the box didn't work for this industry. And there wasn't a lot of investment and a lot of pull from the industry either to try to make that transition. Almost nothing is digital. Now when I say digital, I don't mean a document. A document is a digital item, but it's not digital information. So a tech pack is not digital. A tech pack is a digital document, whether it's a PDF or Excel spreadsheet, but it is not, the data in it is not understandable. The way you can think about it is could a robot read this? Could somebody, a computer read this? And the answer is no, it takes an expert. This is expert to expert communication. Of communication, which is very pervasive in the fashion industry. You have an expert that knows about pattern making. You have an expert that knows about design. You have an expert that knows about technical design. You have an expert— and they communicate to each other through these documents. But it's very human-readable in almost all cases. The information in them— which stitch do I use? Which thread and what color of thread? And which— I challenge you to go to any tech pack you have and be able to pull out that information about a specific garment to be able to say, this is the detail that I need for this step of manufacturing. It doesn't matter if you have another expert reading it because they can interpret it and say, there's lots of room up front. Please come up. There's room in the middle here. I know it's a little small room. Just have to get cozy. Yeah, there's right in the front. I don't bite.

I do.

Oh, well, don't sit next to her. Leave that chair empty. So digital information is short supply in the industry. There's a lot of data, but it's not all connected and it's not all accessible by computers. So this is— another thing is because it's expert to expert, you're communicating from one expert to another expert who's communicating with another expert in the process of the kind of serial process of designing and developing a product. Those are all somewhat by nature inefficient because everybody has their own language and their own symbols and their own way of describing things, and the other person may not understand them exactly the same even though they're experts as well at the factory and the supply chain, or maybe a supply, another supply technology supplier. All of these communication is typically still done in a, what I would say, inefficient way because it's the data is, because it's not digital information, you can't communicate on a on a solid foundation. You have to do it based on— and the other thing we noticed, one of the biggest things is there's almost no automation anywhere to be found. It's all manual. Everything the industry does is manual. Now, that's not true of the supply chain. Fabrics are very automated. There's some chairs in the middle if you can squeeze in. Come around here and you can fill in a few more spots. When you're weaving fabric, that's very automated. Knitting machines, making lots of very fast knitting, that's all very automated. Thread, converting fibers into thread, also very automated. But generally outside of that part of the industry, it's all very manual. Design, hand sketching or sketching on a computer. Technical design, hand creating of tech packs. Now yes, there's some tools that make it a little bit easier, but you're still, a lot of it's done by hand. Pattern making, also a manual process by a person who has to think, ah, the dimension of the waist should be 30 inches. Okay, that means this piece needs to be this and this piece needs to be this. And they make the adjustments manually, maybe using a computer, but it's still a very manual process. Very little automation to be found anywhere in the assembly. All of that, the selling, everything is very manual. Compare it, again, this is coming from a non-native speaker. Insider's view of the industry. There's a very large lack of automation. So one of the things we notice also is that 3D, what we call in 3D, is not actually— it's 3D visualization. It's not actually 3D design. We take 2D patterns that have already been designed and we visualize them, sew them together, drape them, fold them, place them, you know, and layer them, etc., and work with them and visualize them in 3D. Very useful tool, no doubt. But there's very little automation in that process, so it's something you have to do over and over again for each new garment. And the 3D systems, you're not actually— most of the other industries that are using 3D are actually designing in 3D, which is a little different than designing in 2D and then visualizing in 3D, which is by nature slightly less efficient than the other. So what we've seen is that— and you may relate to some of these— is that there's a lot of training involved of all because it is kind of a complex process when you're taking a 2D pattern and trying to visualize it. And it's also because it's so manual. You can't take repetitive steps like creating a welt pocket, creating that welt pocket and just doing that same set of steps over and over again. You have to do it each time on each new garment separate and new. And so the communication and the process, the actual series of steps that go from idea to production is still relatively unchanged. And I've— we actually talked to very many brands where the calendar has changed because of digital product creation. They may have changed a little bit, like 13 months to 12 months, but in a substantial way, like 13 months to 2 months, I've not heard anybody that's been able to make that big of a change in their design process because of the the current 3D visualization tools. And this is really where we focus 3D visualization typically. I know every brand's a little bit different. But forecasting, you have the ideation, and you have the technical design and pattern making. All of that happens typically before you get to 3D. Now sometimes you mix it up. You do a little bit of 3D early on to validate the concept. And then you might do— full 3D with better rendering and more detailed later on when you're wanting to create the virtual sample. What we see is potentially a better use case for 3D is from ideation all the way through to manufacturing. So you're actually designing in 3D from the very beginning. You don't start with a tech pack. You start with the— just create the 3D. It's faster to create the 3D than it is to create the tech pack. And it is faster to create the 3D than it is to send it off to a patternmaker and do all of that. So let me explain how that can be done differently. So automation is a big part of it. There's 3 things that we tried to introduce to the industry in this process of looking at it, again, from an outsider's perspective. One is that we have to define projects digitally. If you change your pattern, does your tech pack change? No, they're totally different. If you change your tech pack, pattern doesn't change either. They're two separate documents. They're not connected really very— they're not connected really in any way. There's two separate documents. If you change your 3D output, your 3D asset, does that change your production pattern? Generally, no, the answer when I ask this question. It's almost always no. They're separate things. If you define your products digitally, then everything is connected. The pattern, the 3D asset, the tech pack, all of the details of assembly, everything is connected into one. So if you're changing one, you're changing them all. There's only one garment. There's not three views of the garment or four views of the garment. The second thing is there's a very serial process. And I think that's why why nobody's really made substantial changes to their calendar because of the advent of DPC. The main reason is because you're still going through the same process, same steps. You have design, you have technical design, you have pattern making, and you have visualization, whether it's 3D physical samples or 3D virtual samples. And then you have iterations through that process. And I know everybody's a little bit different, but Generally, that seems to be a fairly normal workflow. Very serial. And our approach is, well, how can we make that so you could decouple those steps and work in parallel? So the technical designer could work on the rules of the garment. This is how we do collars and collar stands. This is how we do these details that probably a designer doesn't care about the detail level so much. But somebody has to care about it, right? I mean, you, you have to produce a product, so you need to have the details to be able to know how to sew it together. So being able to decouple some of these steps was another important goal of, uh, what we, uh, wanted to be able to provide a solution for. And then of course the automation part of it, because there's so much that is repetitive. If you go to a store now, if you went to a store 10 years ago, What would you see different? Well, of course, styles have changed, colors, fabrics, a lot of things have changed, but the basic shapes that we're selling and producing as products aren't substantially different. There's some staples that are always there, like jeans and t-shirts and whatnot. But there's, and there's of course a lot of subtlety and differences that matter in small ways, but we're still doing a lot of the things the same, but yet it's still a very manual process. And so why can't we automate some of these things that are the same every time? Every time, or almost the same. Why can't we automate that? That was, again, our thinking here. So if we get to the design process, so there's 3 areas we thought automation could be useful. And what do we need in order to automate those? One is the design process itself, the constructions of the garment, having finishes that are sort of assets now that I can apply to different garments. I don't have to use that same finish just on one garment. If I want facing, Is it really that difference from this garment to that garment? Well, the shape of the pieces is different, but the assembly, topstitching, all of those things could be the same or could be very similar. And so a lot of the steps of the process— because, as Fiona, the details of a garment, the shape of the garment, the pieces that make up the shape of the shirt that I'm wearing are not that dissimilar from another shirt. The detail is in the finish where the difference is. Where one brand differentiates themselves from another is that this one is, you know, the way it's topstitched, the way that it's layered, the way that the interfacing is put into it, all of those things are a little bit different. And so we wanted to be able to automate that process so you could set up your rules about what your brand does and what they like to do and be able to generate. So for us, the tech pack is an output and the pattern is an output of that. But through this this automated design process. Marketing imagery is another thing we've talked about. Some people say, no, in order to get photorealism, you have to take pictures. Well, that's fine if you need that level of photorealism. You could also use AI. AI has a tendency to make up stuff. So you tend to— it's a different garment than the one you're actually producing. So you're selling a garment, but that's not what you're shipping to them. And that creates a little bit of an issue. So you have some problems with AI in that regard in terms of photo-realistic. It looks nice, but it's not necessarily what the product is. It's hard to get the product visualized correctly. But you can do quite a bit with automation to what I call gaming technology, which is the rendering, photo-realism, ray tracing, all of these technologies that are used together to make a pretty close rendering. It's not photo-real. It's not real. But it's not far from it. It's pretty close. It's getting better all the time. And of course garment assembly is another area that automation could be used for. And so these are the areas we focused on. So other industries, as we talked about, I'm not gonna spend a lot of time here 'cause these are industries that you, you're familiar with. They use tons of automation. You go to a car factory. I've been to a, a engine factory for Volkswagen where they produce 3 million engines a year and there were like 200 employees in the entire, entire factory. The engines are big, heavy, massive things with lots of parts in them. Lots of— 200 people, the whole factory. And you go to a garment factory and you'll see 20,000 people and a lot of sewing machines, but not a lot of automation. So it's really very different. So I want to talk a little bit about definition, garment definition, because when we talk about definition, we're talking about the 3D typically as a visualization of the garment, but not necessarily defining the garment. What do you send the factory to be produced is not the 3D. You say, make this for me. You send a pattern, or they generate the pattern, you approve the pattern, and you have a tech pack that you agree on, the details of how the garment is to be produced. And you have samples. You go back and forth, and you mark them up and say, OK, this is wrong. Fix this, this, or this. And then eventually, you come to a meeting of the minds, and they can start production. Our case, we wanted to be able to say, OK, no, the 3D asset is not just a rendering of the garment, but it's actually a definition of the garment. It defines the seam allowance. It defines the stitching. It defines the relationships. It defines the layering. It defines the trims. It defines the order of assembly. How does this welt pocket get produced? Well, it's these set of steps that produce it. And so this is what I call my digital definition, which is very different. It looks the same. If you look at this and say, well, that doesn't even look all that impressive. It's kind of washed out. And it's not all that interesting. I mean, CLO could do that, right? I mean, browser, sure. You could do that in any 3D software, visualization software. You could create a garment like that. But this is not just a visualization. It's the full definition of the product, which is every detail, every stitch, every aspect of the product is embedded in this 3D. So this is why we like to present it. It's like the finishes, all the folding, the 2D pattern, the stitches, they're all interconnected into one set of— this is product definition, digital product definition, which is not— you don't find that today in the industry at all. I've never seen anybody that's defining their products in any way rigorously like this where everything is defined digitally. If you wanna auto-manufacture, automate manufacture, you need this information. You need it in a digital format that says, okay, this is the first step you have to do. This is how it should be assembled. Then you have to sew it. Then you have to fold it open, press it, top stitch it, whatever the steps are. These are important. You can't just, it has to be defined. And so you think, oh, that's a lot of work to define it, but actually it isn't. It's actually some ways simpler to do it that way. So the definition, you have the single source of truth. It's really all of the things are one, so you don't have a change where the pattern is different than the tech pack, or the tech pack is different than the 3D, uh, visualization and so forth. So the 3D representation is as true to life as you can make it because it's based off the same information. And of course the manufacturing of it is, um, it makes it production ready because you have you have all the details that are necessary to actually produce that garment. This is what we talked about with serial. You have this process and what we've— I'm not going to go into the details because you're all familiar with this process. What we wanted to do is make it so that it's more parallel. And how do you do that? Well, you do have to have a pattern. You have to have a pattern to make a garment. You've got to cut something. There has to be a shape, and you're going to cut that shape. And you have to have some information about how to sew it together, which is more of the tech pack. How does it get attached and assembled? What are the details of the finishes? So you need those two. You need that thing. You need a sample, right? So all of those things are necessary, but how do we decouple them? Well, we took the idea that, OK, most garments don't start out as a brand new garment. They start out as a derivative of an existing garment. Garment, right? You take a garment and you start with something similar. So I have a garment from last season. It's kind of close, a little different. Or maybe I have one from 5 years ago that I want to bring back. Or maybe I have just a basic, you know, block that is a pair of jeans and I want to adjust that. So the designer doesn't need to wait on the patternmaker because you're starting already with something that is already visual. Already, you can see on the screen in 3D. You have that block. We automatically use AI to help us automatically assemble the blocks. So they're already— you just drag them onto the avatar. You don't have to make stitch relationships or do any of those things. The tech pack— the tech pack for us is an output. You're defining the garment as you go, as you say, let's change that length here. Let's change the seams, the stitches. I gather onto the edge of the sleeve and make a little bit of smocking right here in this part of the garment and so forth. As you do that, you're defining the tech pack. You're interacting with the garment. So the tech pack is actually produced as you go, and you can export it at any time because it's just a human-readable version of the digital definition of the product. The digital, that big circle with all of the data in it, That is the product, and the tech pack is just converting that into something an expert can read, somebody in a factory can read. Pattern, you start with a pattern. You're adjusting the pattern as you go. You may not know you're doing pattern making because you're just shaping the silhouette of the garment. You're changing the length of the garment. You're changing the, maybe adding extra seams, like I want a center front seam and I want to change the fabric on both sides. I want red on this side and blue on that side. Whatever you're doing, you're doing pattern making. You just don't know you're doing pattern making because it's built into the software. We do what we call parametric CAD, which is very used in other industries, which is you change the shape of the CAD by parameters. I want the waist to be 27 inches, not 26 inches. It will change the shape of the pattern pieces to meet that requirement. And so the sampling— you always have a 3D. You're always looking at the garment in 3D. So there's never a case where you need to go and make a sample. You're starting with a sample. Again, it's not the same garment. Maybe you start with a t-shirt. You make it into some other kind of top. You added— you changed the finishes. You changed the shape of the openings. You just do that by drawing on it. And you always have a sample. So you can see how the tech pack, the technical designs of the details, if the technical designer has prepared the finishes, here's how we do plackets in our company, here's how we do cuffs. Okay, if you've prepared that in advance, our approach is to reuse those. They can be used across all garments. You don't have to, you're not doing a specific garment, you're doing a cuff, and the cuff becomes an asset that you can then drag onto a garment. I want to see this same garment with cuffs now. This type of a placket with it and so forth. So you can see how they can decouple these things, which means that you can go in parallel. You don't have to all work at the same speed and wait for each other as you go through the process, which dramatically changes the time where it might take a few months to go through all of those steps. You can do it in, in a matter of days if you're in a hurry and weeks if you're not. So some of the things that we talk about, just to summarize a few things, and then I'll show you a video. We're running out of time. But drag and drop finishes, customizable assets, all of those kinds of things. It's the pattern making that we do in the background. Let me just go through the concept. So you start with an idea, sketch. However you want it. It could be a competitor's garment. It could be a combination of garments you've seen. I like this about this one. I like this about that one. I want this about this one. And you want to put them together. So you have an idea. Then you have a basic block. We're going to use this block to create those 3 garments. And we're going to iterate on those things. And then once you're done, you have all the details that you can export. Is this washed out, or is that me? It's like I can't hardly read that. But you have— trust me, there's a tech back in there somewhere. So let me see if that'll start playing. There we go. So this is not quite real time. We did cut out some of the regeneration time. But it's, I would say, about 80% accurate in terms of— we sped it up a little bit. But basically, we're going to start with a simple garment. Drag it on the avatar. And this is real time. It takes that fast. It just sewed it together and put it, you don't have to, no stitch relationships, none of that. She's gonna remove the sleeves 'cause she wants to, we're making one of those three, her idea was to create one of those garments. And so she's gonna go through and change the opening. This one had binding when it came in and she's gonna make it a definitive, a facing. So she's putting facing on the 3— the neck opening, the 2 sleeve openings. And so she's going through and just— that's how easy it is. I want facing here. That's all you have to do. So there's facing inside of each of those holes, fabric sewn, you know, stitched in. All the stitching relationships were defined. And she's going to change the shape of that opening. So you have like a cutting tool that allows you with Bezier curves to to kind of go through and just adjust where you want the edge of the garment to be. Your patternmaking here, this is our version of patternmaking. You're not actually going into 2D and doing the, you can do that, but normally it is really that fast. It generates the garment very quickly. And then she's gonna, I think, add some, I can't, oh yeah, she added a polo. Placket. So the collar, all of that was resized to fit this garment. It was also, the slit was cut, the pieces were generated, folded, placed, topstitched, and the buttons, all of that was done automatically. Same here. We did a cut, spread, gathered it back in, created the, and she's gonna, I think, change the fabric of part of it, or maybe that's the first garment. Her first one, yeah. So this is basically how you're using the software. You have to have technical design because you need to have— these finishes don't come from nowhere. They have to be created. We create a lot of them. They're part of the software. But everybody has their own specific way of wanting to do things. There's so many ways to do just a simple placket. I went to a store, and just in one store, I was able to count like 15 different versions of plackets in one store. All slightly different. And the details matter. So we wanted to make sure you could have all your details matter. So we're going through the process now. This is the second garment. We took off the collar, just did a facing on the neck. And now I think she's going to change the fabric on the lower part, the gathered part. I don't know the right terminology for fabric, but basically you can see how it's She's adding a print to that. So regenerate it. So we have another one. And you can see how easy it is and how automated it is. I don't have to go through— in any other software, you'd have to do all of those steps manually. So for that gather, you'd have to cut the garment front and back. Then you'd have to spread those pieces front and back and know how to do that, because a lot of designers wouldn't know how to do that. That wouldn't be something they would be comfortable doing. And then you'd have to sew that back together, create that relationship to it and gather it back in and then change the fabric. So you can see how simpler it is just to predefine the finishes and drag them onto the garment. And then you can iterate in a matter of, I mean, this is like a 7-minute video or a little less and we've created 3 garments. And not only have we created 3 garments, but everything about those garments is defined, as well as we can export as a tech pack. We have the pattern we can export as a DXF. We have costing information. So just changing the length here. We have costing information, because we nest and show you what the yield will be for that fabric. If you put the cost of the fabric in, we'll tell you how much each of the sizes costs. And not a lot of— so what you have is you have 3 different skill sets blended in in a totally different way. You have a patternmaker still, you need blocks, you have to start with something. You have a technical designer who's focused on the details of the garment. How should seams be done? How should finishes be done? And you have a designer who's mostly focused on the visual. That doesn't look so good. Let's try this. Let's change this, let's take that off, let's adjust this, let's delete that, let's add this other thing. What about this? Oh, let's try that. And all of those things are done. You see immediately the feedback. You don't have to wait for a month to see the idea come to life in 3D. It's all done in seconds. So to, to us, this is what we call automation, and also about The fourth dimension is time because you can't— it's hard to move at the speed without this level of integration automation. So all of these things are all connected. So we're exporting the DXF file now for that garment. I think she's going to create the tech pack here in a minute. Yeah, so these are the data that goes into the tech pack and so forth. So I'm going to stop here because I think we're running out of time, and I want to to just open for questions because I see a lot of you looking at me and I think, ah, well, I have a question about that. So we're a small room. You can just ask your question or we can put Slido up if you prefer that. Any, uh, any thoughts? What's your feedback so far? Too good to be true? Yes.

You had that crew neck and —put in a placket. You can drag that placket in there. But then that placket, if I want to make adjustments to that placket or the collar, it goes to a finer after?

Yeah, so you have properties. You adjust that. So if I want to change the width, there's a property there. I can change the width of the placket, the length of the placket, the number of buttons on the placket.

And then after you make those changes, it'll take you back to a 2D pattern, or you can just validate it?

No, it'll just regenerate with a new width. If you wanna look at that new pattern. Yeah, you can go in and look at the pattern, yeah.

And what finally is my issue sometimes between 3D and 2D is that you make it pretty 3D, but then the pattern, I look at it, I'm like, it doesn't look right. How accurate is the correction?

Yeah, so we have tried to automate and use a lot of tools to create a very producible pattern, and we believe the pattern is quite good, but you know, If we need to make additional improvements, that is certainly a possibility to continue to evolve it. The pattern is, in the end, we want the garment to fit right and to be the correct shape on the person and be comfortable and also be optimal. And so we've tried to incorporate all the rules, like for balancing and also rules for curvature and things like that that we know, OK, you can't do that in a pattern making. So we try to incorporate all those rules into our expert-generated pattern.

[Speaker:AUDIENCE] Is the system connected to a measurement chart with grading?

[Speaker:EDWARD SINGER] Yeah, so the measurements come out automatically as well. You have reusable grade rules that you just drag onto the garment. So if I have a bust measurement, you just drag that on there. If I have a waist measurement, I just drag it on and say, I want to see what the waist is. And it'll measure it for you. Yeah, you have a question? Yeah.

Does this connect up with the fabric properties?

Yes. Yeah, we import all the fabric properties we could find. So that's mostly— most people have told us they have fabric in CLO format and GLB. They have a GLB export file format, U3M, which is BrowserWorks' format. And then we have our own format. We use off-the-shelf equipment, but it's the same stretching and bending. So the patterns you would expect to get different patterns out depending on the fabric? Yes. Well, it's going to show you what the pattern is going to produce based on the fabric properties. So it's going to drape accurately according to the fabric properties.

Yes, another question? What is the relationship then with this new process between your different types of designers? So starting with garment, then technical, then visual designer? Yes. Do those roles get eliminated or do they change?

So possibly one of them gets eliminated. But the first one is the pattern maker. Yes, you still need a base pattern to start with a block. You don't maybe need as many as you need specific patterns because you will need fewer blocks than you will need specific patterns. So there's probably less pattern-making work overall in this new process. The second thing is technical designer. Technical design, instead of creating tech packs, because those are automatically produced. Technical designers are mostly creating— basically creating the rules. What is allowable? How should we do this kind of finish? So if you have a pocket integrated somewhere into your seam and it's a specific construction method, the technical designer has to explain those rules to our system so we know, OK, this is the order. This is how it's done. And that way, the designer will use that in a correct way. So there's still a very important technical design role in this. And then the sample maker is the one that potentially could be eliminated because you always have a sample on the screen. You never don't have a 3D. It's always in 3D. You never saw a 2D screen here. There's never a 2D— you can go into 2D mode, but we don't need to very often because you can just look at the garment in 3D. Sorry, you had another question?

Yeah. So just to clarify then, your—

Software would replace CLO and Browzwear. So we don't like to think about us really as something totally different. We see ourselves as much more closer to the ideation phase. So you can use our software to ideate. And if the pattern isn't as accurate as you'd like, you don't have to use the pattern. But it gets you close. It has full costing in it. So you know, OK, this is a producible garment. This is something that we like. This is close. And then you can do your normal process afterwards, but you can funnel down the ideas much quicker by having the ideas always in 3D and be able to see this. So the designer gets feedback within a day. Right now the designer might have an idea, but they wouldn't get feedback for maybe a few weeks or months depending on the speed of the process. And so this way you get feedback really quickly. And if you still wanna use CLO to do the fine visualization, get the, you know, stitching exactly right for the 3D rendering and all of that. We render well, but I would say that's not our strength. The rendering is good, but it's not great. It's just good. It's good enough for what we want it to do, which is to— is the idea a viable idea? That's the main question right now. Yes, I know we're out of time. So come up to me afterwards if you have more questions. Sorry, they need to have another meeting in here. Thank you.