Epic Games x Tapestry: Scaling Fashion with Real-Time 3D

Thank you for joining us. We're gonna begin because we're on a strict time frame apparently, um, and we want to have a quick chat, but we also want to get some questions from you guys. Um, the kind of the plan is that I'm a huge fan of Brandon, and I think that Brandon's, um, not necessarily a massive fan of mine, but he's quite a big fan of the technology that I work very closely with. And so what we thought would be hopefully informative, mildly entertaining, was for us to have a kind of a joint interview, a conversation. So very quick introductions. My name is Sally-Anne Houghton. I work at Epic Games. I specialize in fashion and retail, which means that I help the industry at large, whether it's a brand, whether it's a startup, whether it's a creative studio. I help them understand how they can leverage our technology. I also lead partnerships because we're not the only technology, nor do we we want to be, and so help with integrations so that we can develop workflows and pipelines. Brandon, would you like to introduce yourself?

Yeah, thank you, Sally-Anne. I'm a big fan of Unreal Engine. My name is Brandon Keeney. I'm the Senior Manager of Digital Product Creation at Tapestry. My primary role is to focus on innovation and implementation of new tools within our brands and product categories. Also, I feel there was something I wanted to say there, but to create, make, and sell our products.

So just so you know, the visuals that we've— we decided not to have slides or clickers. So we've just got some visuals behind us. These are Brandon's visuals, i.e., this is the work that his team have been doing with Unreal Engine and other tools. So we'll weave it into the conversation. But if you have any questions about them, we can answer questions about them a little bit a little bit later. Do you want to begin?

Yeah, for sure. So my first question is, how does Epic see the future of digital fashion evolving across fashion, retail, gaming, film, and other industries?

Which is a huge question to start with. So I'm going to do the shortest version of that question, otherwise we'll be here for many days.

I think—

so fundamentally, we— feel, have always felt, that digital fashion is a huge, incredibly important part of digital spaces. We are, whether you're aware it or not, the biggest unknown fashion company that there is. Our business model is selling fashion. Our players come to our ecosystem to buy fashion and to begin to build digital identities within that space. And we enable creators to be able to build characters and be able to build narratives in real-time environments. And so digital fashion to us is super, super important. And then when we get very specific about the industry, as in fashion and retail, we very much want to be part of the digital development of the industry. We feel that real-time is a huge unlock for the industry to be able to continue to progress with its existing digital efforts and designs. And we feel that we can be a big part of that. And with the help of integration, from other tools that we can begin to build those pipelines and unlock all of that opportunity. So that was my short answer. And what about you? What do you think should be happening within fashion and retail with concern to 3D, 3D tools, real-time? Because you're using the tools. So where— why do you invest in those tools?

Yeah, so I think real-time is really important. Real-time, typically it requires really good optimization of your 3D assets in order to make real-time really effective. You're able to cut rendering time substantially because how it looks in the Unreal, it looks like in the final output. So it's really nice because it's cutting out the time to create the renders. But also, you can be interactive with it. So you can really do— kind of anything you want. And that's one of the ways that I've sold Unreal Engine too internally is that Unreal Engine is the tool to build whatever you want. It's a creative's sandbox. So you can build a configurator, a showroom, an experience. On this slide right now, that visual actually only took me 5 minutes to make. So it's a really quick visual. And that's real time too. So it wasn't even a render. So it just looks like that.

Awesome. They're really beautiful.

Yeah, so my next question actually is, what's one feature that you wish fashion brands knew about in Unreal Engine but most still don't?

OK, so I think where I'm going to struggle with that question is one. I'm going to— so I'll just say a couple very quickly. MetaHumans is now in Engine. So MetaHumans are sophisticated avatars. They have configurable bodies. They're now in Engine. And some people don't know that. So I think that knowing that they are and going in and playing with them is a huge unlock with their bodies that can now achieve the measurements that we want across the industry. We also have some really awesome motion graphics tools. And I know that doesn't necessarily come as a natural, oh, that's what I've been missing my whole life. But I think as you're using your 3D to create more and more interesting content, when we move past just that still image or a piece of linear content, we can start to build the next layer. Like, you know, what is the future of creative content and creating content to share across your platforms? And so having tools that allows you to once again very easily incorporate and build more sophisticated media. I think that that— and a lot of people don't know that's in there.

That's fun that you said that, actually, because I was actually talking to Daniel from my team that one of the next things I want to look at with Unreal is the motion media side of things because I've been seeing a lot of really cool YouTube videos. There's a particular artist who did a whole class of motion media design because that's a space that we're also kind of interested in because I find that, A lot of 3D artists in the fashion space, they don't communicate how powerful a 3D asset can be. It can be more than just a standstill image. It can be a store pop-up. So we really integrate our 3D models for many things. They can be a popsicle, a cake, a bounce house. We do it for a lot of things. It's even been a hot air balloon. So we use our 3D models for everything.

Another one, another one, a quick one is we have Chaos Flesh now. Once again, this might not— feel intuitive, but Chaos Flesh is, um, ML for flesh. It comes out of the character build space and wanting, uh, flesh and muscles to look really realistic and to simulate correctly. Now imagine applying that to leather goods. You know, it's a simulator. Applying that to leather goods and different materials, materials that you've potentially struggled with previously, and being able to get a much more realistic dynamic movement. I'm excited to see how people use that in a very unconventional fashion retail accessories way?

Yeah, we work closely with our photo studio as well to get the setup of what the visuals should look like. So having the real-time feedback is super helpful because they can really be like, oh, I feel like that shadow is kind of harsh. One of the cool features in Twinmotion that I've actually never seen another tool really do is you can really soften the specific shadow. —which I thought was so interesting.

So that kind of takes me to my next question to you. How are you seeing— like, I know that your use of real-time is, let's say, developed. But how are you seeing real-time change workflows at the moment with your team, but maybe across other teams as well?

Yeah, so— Right at the moment, we're really looking into the virtual showroom space. I know a lot of companies have kind of done that in the past. And a lot of the times, they haven't really been successful. We also have done it in the past, and they weren't really that successful either. But what's different about this route this time and how we're including Twinmotion as a big part of it is that we're doing a lot of work to really upskill our designers into 3D. And the way that we're doing that is they're not making the 3D models, but they're making the visuals, so the visualization. So they're doing it in Substance Painter to create the final render for us that we can use our assets for design-approved showrooms with Twinmotion. But also, they can be used for all of these other use cases that we use them for, like the cakes and the popsicles and stuff like that. And they'll be design-approved already because the designers made it. It's like the same work that— Coach has done and, well, pretty much every brand has done in the past with Illustrator. When before, when people were sketching on napkins and they had to flip to Illustrator.

And look, I think, like, let's be real here. Like, things, a lot of the things that we want to do, we're trying to do, they're not easy. It's hard. Like, what are the frictions that you're seeing?

Yeah, definitely. I think having the skill set is kind of hard to come by. I find that it's, especially in New York, it's a little challenging to come across somebody that has Unreal experience specifically. But I find with tools like Twinmotion, if you haven't seen it yet, it's very accessible, easy. It's Unreal for the beginner, but it's advanced. You're getting a really advanced real-time visual that you can navigate around, make instant changes. You can even create a configurator and adjust your materials if you're doing a review and they want to see it in 10 different colors. You can create all the colorways and just toggle between them. Do you want to go? Oh, yeah, yeah. So my next question is, what role do you see Unreal playing in the future of virtual fashion?

Wow. I think that it— First of all, I think it— there needs to be integration, integration, integration. So as a foundational piece of software, almost like as this kind of mothership with integrations across the board into other iconic industry stakeholder technology, I see it as this, you know, this incredible connective tissue. That allows for us to get closer and closer and closer to this full end-to-end 3D digital workflow pipeline that we've talked about. It kind of goes back to the make once and then deliver it wherever you want. But that cannot be done without integrations. It cannot be done without plugins, without building pipelines with people like Browsware and Cloe, which we— and Adobe, which we do, and we want to do more and more and more. And it can't be done without the, you know, the help of the industry, without feedback, without teams building their workflows, their existing workflows, and turning them into 3D workflows and finding out what the problems are and then sharing those problems. And, you know, and there being a fluidity. So I, you know, I see it as an opportunity for us to build in 3D and then be able to interact with that build all the way through to a point when you don't want to. Like, if you don't want it to be a 3D content anymore, but fundamentally, it can be all the way to the end. And it's only really Unreal that can do that.

Yeah. No, yeah, I agree. And then my next question is, What Unreal native tools or workflows are being developed specifically for fashion?

Well, it's a good overlap. So we have a great relationship with CLO. We have a great relationship with Browzwear. We have a great relationship with Adobe. And as a result, we're learning how those tools are being used upstream and trying to remove clicks. It's as simple as you shouldn't have to have 5 steps when we can have a simple preset and have 1 step. And so having those relationships and learning what the frictions are and then working through them like a to-do list. So I think it, and that's all about having brilliant relationships. So we, you know, we love working with the BrazeAway guys, we love working with the Clo guys, you know, and Adobe. And also teams that are using those tools and want to do more with the projects that those tools allow for and push them further into a real-time space. So yes, and also on our side, we're really, really keen to continue to integrate MetaHumans. So MetaHumans, I think, can be the body that we all want in the future, something like this. This thread that can go all the way from the authoring tools all the way through to content of all different shapes and sizes, and having a digital garment pipeline. And we have the beginnings of a digital garment pipeline with USD that does take you all the way to Fortnite if you want it to. But there's still a lot of— work to be done. Like, it— you know, there are still frictions within that pipeline. We, you know, we want garments to flow really, really easily into real-time environments, into gaming environments, into e-commerce environments. You know, wherever you want to see them, we want that flow to be easy. But we've got to work with, you know, what is simulation parity with simulation that everyone's using? What is— what— how do we keep things with and without bodies? How do we move things? How do we make sure that the size doesn't matter when we know that in most places the size does matter. So yeah, there's lots to be done, but it's, you know, this fluidity that we're working towards.

Yeah, amazing.

It's my turn, isn't it? Okay, let's see. So what do you think? So everyone's here because they're using 3D, I'm assuming, or at least 3D curious. How do you think people can scale? Um, people are like— I feel as if this room is filled with 3D authors in that, you know, tools are being used to create a 3D asset, and we're all talking about the opportunity thereafter, or we are. And there's a lot of AI at the conference, and, and, you know, how we continue to, to leverage digital. But how, from your perspective, are there any tips or watchouts when it comes to scaling? Because I'm also seeing that a lot of people have been mass to scale, like success here, like there is success, which is awesome. Yeah. How can you help from your experience?

No, yeah, I think that's a great question. And it also ties back to what we're working towards with getting the designers upskilled into 3D. I think the way to kind of bring, upscale the assets is to not leave them behind, but bring them on the journey. Yeah. I don't think we'll ever want like a giant DPC team, but we'll want, you know, designers to be upskilled into 3D and always like, have that skill set on their end so that that way they can build the assets and they're design approved. They can be used for so many things.

And in-house— this is a rogue question, so I apologize. But I feel like it's come up in a few conversations that I've had over the last couple of days. How do you deal with people that are non-believers? Oh, yeah.

Well, so it's interesting because actually at Tapestry, we kind of have a little bit of the opposite. A lot of teams are begging us. We're at this point where a lot of teams are begging us for assets. So scale actually is our biggest challenge. We don't have enough assets to provide the needs that are needed for the business. So that's why we're turning to our designers to upskill them into 3D.

Wow. So I know that you were at Unreal Fest weeks ago. Oh, yeah, I loved it. And you saw a lot of the updates. And you saw the MetaHuman updates. How do you think— because I feel as if MetaHumans has been in a hard proposition for a while because although they're awesome, you couldn't really do what you needed to with bodies. How do you see the updates affecting how avatars can be leveraged with apparel in the apparel, digital apparel space?

Yeah, for sure. And conveniently, actually, we're on this slide. So this slide actually is a kind of a funny story because the avatar on the left I sat in a meeting with, I don't know, 12, 13 people and spent— I think they spent $1,000 on that avatar. And nobody liked the avatar. They thought she looked a little crazy. With MetaHumans, you get more of like a— it just feels more realistic. It's an avatar, so you're still going to get a little bit of the uncanny effect. But I think it will help bridge the gap between teams that don't understand avatars and help get a better visual. And also, before with MetaHumans, because we were dabbling in it, before 5.6, you only had the 6 body types at first. And now with 5.6, they have all of the measurements. So you can go in, and you can type in your measurements and get the perfect body that you can then use the Clo Live Sync to upload the ready-to-wear onto the avatars. And then you can get a real-time render that you don't even need to wait for. It's just there.

Okay, um, so this is, this is my final rogue question, which basically means that Brandon has no idea I'm about to ask it to him. Um, I feel as if AI has been a big subject here, um, and so I, I feel like we can't, like, we can't have this chat without bringing it into the conversation. Um, and this might be a bit unfair, so you can push it back at me if you want. Um, How do you feel real-time and AI can live together? Yes. What does that mean for you?

This is such a fun question, actually, because I was talking with somebody yesterday actually about this. I see a future where graphic— because my background was actually in gaming. So I'm like, when I think of like the gaming industry as well, I see a future where AI is kind of the graphics. If that makes sense. So you get these— you can create these really incredible visuals, kind of. And I see it needs the real-time technology in order for that to be successful because right now AI is kind of— I don't want to say wonky, but you get differentiations in each render. So you need something that's consistent. So I could see a future potentially where AI is doing that heavy work to produce a more— creative and visually appealing outcome, if that makes sense.

Yeah, totally. And from our perspective, we feel like there's a huge place for AI in the real-time space. We are firm believers in 3D, in the pureness of 3D assets and 3D environments, because that's what you need if you want to be able to fully interact and move around a 3D environment in real time. But we very much want to empower developers, empower creatives, and we want to continue to democratize tools. And AI can do that. You know, it can help us go from a code environment to a no-code environment. It can help us go from a complicated click environment to something that's much more prompt-oriented. So that's, you know, we very much see there being huge opportunity within that build space so you can build quicker and have an, you know, have that build, have that 3D entity quicker, and hopefully open the doors to more and more and more people so that you don't have that skill set miss and that you don't have as many objections from team members. But also as well, that final filter at the end. Unreal is a great renderer and can make stuff look great. Oh yeah, as a final image. But sometimes you want to do that and you want to do it at double speed. And I think AI can be a really great final filter within environments for content where, you know, you, you know that everything you need is in that scene, but you want a really quick final, final render, final look. And I think AI can really help to take something that, that looks it's 70— no, 97% there, but there's a little bit of uncanny valley. And you know what? Customers don't like uncanny valley. And AI can help take it to that final few percent, so it's everything that you want from that image.

Yeah. And actually tying into that, something that I kind of took away from Unreal Fest when I was there was that y'all had this AI chatbot within UEFN now. I think it is a chatbot. And so I'm thinking kind of like, what is the future of that chatbot to build these virtual experiences with your brand assets inside something like Fortnite or maybe even just Unreal Engine if you all brought that feature to Unreal as well? 100%.

And I think that without giving away too much, I think that that suggestion wouldn't be a mad suggestion in the— yeah, so creators within the creative tools in Fortnite, there is a chat box so you can basically ask ask the AI what you want to build, you know, the what you want to build in the game, and, and it facilitates. It either guides you through the quick steps or it's able to, to build that, that, you know, that level, that feature for you. Um, and I think that we'll definitely look at that for other parts of the ecosystem. Um, we also had real fun with AI characters, so we had Darth Vader in there, and, and you could go and you could chat to Darth Vader, and Darth Vader knew everything about his life, weirdly. And, um, so If you're a Star Wars buff, that's an incredibly engaging experience. So you can imagine how applying that to a brand and you want to build a brand, you want to build engagement with a brand, you want to build engagement with your legacy and your products. It's, you know, it's a really exciting way to be able to use AI through characters, through interactions with customers.

Yeah, I do have another question for you, but I'm not sure. Should we? Yeah, yeah. How can fashion brands best leverage MetaHuman customization for virtual try-ons?

So I am very, very excited to see how MetaHuman can fit into the digital fit space. I'm sure you have all seen— I feel as if I've seen lots of different companies trying to solve the problem, and most people are nearly there. And I think that MetaHuman can help get us all closer. To, you know, build pipelines that are from the authoring all the way through to a digital fit space that represents customers, that represents body shapes. So I really think, and I'm very excited to help support, be part of that next push. We now have the body diversity that we need. So how can we apply that to e-commerce and digital fit? I think that's super exciting. So I'm excited to see— I know there's some POCs, I know people are exploring it, I know companies are integrating. So I hopefully next time we all come together, there's, there's something we can all look at and point at and see that, that progression. So it would be really great to get something that can be white-labeled and used across the industry, uh, and, and, you know, something that works because it's easy. And at the moment I feel like that digital fit space is hidden at the back of a website and you have to find it and you have to work really hard and it's jingly jangly. So something that just works. And I think MetaHumans can help with that.

Yeah. I also think too of like the fun implications with MetaHumans as well, because Coach brand has a virtual kind of like influencer. So think of two of like you can, you can really with MetaHumans, it's like it automatically rigs everything and you can insert any shape as long as it has a face. So you can insert a shark in there, and it'll do the whole MetaHuman rig on the shark. So it's really cool. So you could build these influencers, dress them in your brand, and do all kinds of cool and crazy stuff with it as well.

We have 3 minutes. Does anyone have any questions? I do. Oh.

It's actually quite relevant to what you guys were talking about before. So you guys were bringing up a lot of the uncanny valley, a lot of trying to integrate getting people to do those try-ons, what I've noticed is that the consumer is kind of hesitant because they're hesitant towards these 3D spaces. They consider it a little strange. But however, as we've seen, you know, through Fortnite, people are so willing to buy the skin that relates to their character. They love seeing Darth Vader because it's that human connection. So I'm curious on what you guys think about making that marketing to allow people to feel connected to this metahuman, to make it a little less uncanny valley, a little more this is my idol girl who I really love. Just encourage that. I'm also curious on your thoughts on mixing real life AR experiences, maybe like pop-ups where it's virtual and in person.

You want to get this? Yeah, I'll talk first, and then maybe you can finish that one too because we also do gaming with one of our brands. And we find that a lot of Gen Z are seeing their— or in younger are seeing kind of like their avatars as an extension of themselves. So like having your brand represented, you know, maybe it's the first time they're interacting with your brand. And so they find something that they really resonate with. They have it on their character and then maybe they'll want it in real life and they'll have affinity towards the brand.

I think, I think co-creation has an important part of, of, of that space in that some people don't want an avatar that looks like themselves. Some people want an avatar that represents their body. Some people want something that's totally fantastical. And so I think allowing there to be a really intuitive ease of creating the avatar that you want, and if that is in a, you know, an e-commerce space and it's about clothes looking the way clothes are going to look, then it has to be easy and it has to be intuitive. But I also think that it has to be— there has to be a diverse output And, you know, and I think the reason why video games win and why Fortnite wins with concern to that, you know, that purchase moment is that you're already whoever you want to be. Like, this— it's such an ease. You don't have to take a photo and then put in your measurements. You know, it's like you're already there. You're already the character you want to be. It's about layering on that. And it's— and there's an experience around it and there's a reason to do it. So on e-commerce, the reason is that you want those clothes and so tap into the goal. In gaming, the reason is that you're hanging out with your friends and you're about to go and have fun. And so, you know, you therefore have the steps and you make it intuitive to get to that goal. So it's— I could talk for longer about that, but I will stop. But thank you so much for your question. I've got 10 seconds. Anything else? All good? Awesome, awesome. Thank you, thank you so much, everyone.