Markor's Mark Feng on the AI opportunity in design

29:30We're seeing designers starting to think that way. But I would say the consumer is ahead of everybody else. So how the consumers find information is changing and how they shape their decisions are changing. So with that comes challenges, but also tremendous opportunities. I would say that this year, the U.S. market has been a highlight across the globe. The Middle Eastern market continued to be strong. We're seeing great opportunities in the European market. The Asian market has been a lagger for us. But yeah, it's a very complex picture. It is a very complex picture. And the issue with the tariffs, I mean, yes, as we said earlier, there's the tariff exhaustion.

30:17And part of the exhaustion is just designers who are just every day getting 10, 15 emails from vendors saying, here's how we're raising prices or here's how we're going to – some vendors have decided we're just going to spell out exactly what the tariff is and we're going to call that whatever we're calling that, some kind of additional tax or markup or whatever you have. and everyone who from the maker's side who has had an asian supply chain has has tried to shift things turns out it didn't help to shift things to india it didn't didn't help to shift things to vietnam because vietnam got hit with the tariffs so for the massive amount of product that you're bringing into the states how have you navigated the price increases the decisions you've made about how much you can swallow as the manufacturer versus how much you have to pass along?

31:13So when the first wave of tariffs started to happen, a lot of our peers started to move out of China. But we've been kind of, you know, holding on, you know, for two reasons. One is we invest into technology. So our factories are a lot more automated than your average factory. So labor isn't as much of an issue. We also have a higher end product, right? So that requires that more technical production process and that craftsmanship. And also we have the Chinese domestic stores that we supply with. So our Chinese facilities have always been critical for our brands and our stores in China. But right before the tariffs happened, we actually started acquiring our own factories in Vietnam.

32:05So we kind of anticipated this happening. So we merely shifted production. Now, even that, right, is hard. Every time you give the same finished panel to two different factories, it's different colors, as you know. But at least it's our own production facilities, and we're able to have superior quality control. But yeah, it's definitely been a very complex problem to solve. this month i'm catching up with amber lewis about her new rug collection for laloy so when you first started working with laloy was there a learning curve an educational process that went on so the construction of rugs is a language i don't speak no matter what still to this day they're like they're like this is a hand i'm like where who know what's now like i don't care.

32:59I want it to look like this. Get me to, you guys are at A, I'm at Z. Just tell me how to get to Z. I think now with all these years under our belt, it's easier for me to say, yes, yes, no, no, absolutely no. Let's do this. The language is very easy to speak now. I also too have so many rugs that we've designed into that I love now filling the void. So where's the gap in the market? Where are we feeling like we don't have something that already looks like a version of this? How do we do this better now? It's nice to have all those rugs under our belt to say like, oh yeah, this really works. This is strong.

33:31Let's do this, but do it a little bit different and better. And now back to the show. And are you imagining that, I mean, the president is over in Japan right now and is ultimately, we think, going to meet with President Xi and we seem to be thinking that there's going to be some great deal ironed out and that much of this might go away or that relations will at least improve. Is that what you are imagining? And are we just going to see some much higher level of tariffs just be the norm going forward? You know, I can't predict what's going to happen with the tariffs. But one thing I would say, this market, we had the honor of hosting a delegation from China.

34:23And he's a district secretary from the capital of furniture industry in China. And it was a delegation of about seven government officials. So we hosted them. And the mayor of High Point, the CEO of High Point Market Authority, we all sat together and we had one of these signing of the friendship certificate. kid. So obviously, you know, we had speeches. Now the Chinese secretary, he's very, you know, first time in the U.S., first time in High Point. He was blown away. And he says, wow, we have so much to learn and we would love to have a greater partnership. But he didn't know what the mayor of High Point would say in response in light of all these geopolitical tension.

35:10So Mayor Jefferson, when it was his turn to speak, he said, look, our countries are fighting over this tariff stuff, This is me paraphrasing, not a direct quote. But the second thing that he said, but we don't care. You know, our cities, our people, right, have always loved each other. And our industry has always been a global industry. I mean, the High Point market has been, is over 100 years old, right? And High Point wouldn't exist without international participation. However this tariffs thing would go, I'm sure our industry and the people in our industry and our two cities will find a way to work together.

35:47And that was, you know, honestly, I became a little emotional too. We're worrying about the tariffs and whatever, but what we're forgetting is the people that actually work in our industry that we're all connected. And it's impossible to just sever the ties with a single stroke. Are you hoping or looking to do more manufacturing in the U.S.? I mean, part of the tariff, the whole idea of it is to try and push more business back to the U.S. Can you imagine doing more production here in the U.S.? Yes, that's definitely a possibility. We're talking to a lot of industry partners. And we've always had partners here in the U.S.

36:30as well. Roe is just a fantastic asset that we have. So, yeah, we're not only looking in the U.S., but we're looking in Europe. We're looking in South America. But we do have a pretty high standards in our partners because we also have our own factories. So that process for us sometimes is a little more demanding because our quality standards are high and our designs sometimes are hard to replicate. Speaking of Europe, you mentioned earlier showing in Milan and many people who have gone to Milan for Salone can't imagine a world that feels more different to them than coming to High Point. So those are almost like two different planets to some people, I think.

37:20And not too many High Point companies are also showing at Salone, but you are. And I'm curious, tell me about the thought behind that. Tell me how you're received and how it's gone for you. Yeah, it's been a dream since day one, actually, because we are kind of like this centaur, half-bred, half-east, half-west, global, right? And even the, like, okay, so you call it the most landlocked city in China, but it's right in the middle of the Silk Road. And my father, when he was growing up, he was learning Russian and he was, we were eating kebabs, right? So it's this fusion, this global fusion kind of culture that we've always had.

38:02So at the beginning of Caracol, we've always had a global vision in mind. So the hopes was, hey, we're going to do High Point, we're going to do Maison Object, we're going to do Solon. But as you know, priorities and investments, it never happened that way. So when it happened, we had a lot of traffic and it was a hit. And most people came in and said, oh my God, you guys are finally here. because they are already customers. They're coming in from Europe, you know, possible. Like, okay, now we don't have to come to High Point anymore. You know, now we can just see you guys there. Years ago at High Point, you would come and a new introduction would be six, 700 SKUs and there'd be all sorts of furniture pieces that you couldn't imagine were ever really going to go into production.

38:51But it was just so much throwing it all up against the wall and seeing what people responded to. And I wonder, this market felt much more contained. I didn't feel like there were huge introductions really anywhere that I saw, in part because surely that reflects the economic environment that we're in, but also perhaps a much greater discipline that companies are showing about what they really imagine making. Do you make a lot of prototypes that you aren't sure are really going to go into production, or how do you think about it? So it differs from brand to brand, but the short answer is we don't anymore.

39:29We used to absolutely do that, right? Whatever, it's like, oh, travel around the world. Take your inspiration from this country, that veneer, that architecture element. Man, my God, like this, right? A hit, right? And then you do that as like crickets or the worst thing is you do that. You win a pinnacle award and then a cricket moment, right? So nowadays our approach is we have data on these market. We have knowledge. I mean, we have these merchandising teams that gather, hey, very much the CPG approach, right? Like what's selling? Okay. What's selling in that market? What does the consumers want?

40:08What attribute about that product is selling? Because something, a white sofa may be selling, but is it because of the hardware? Is it because of this or that? So we gain that insight first. And then we draw a line for our designers. You can still travel. You can still get inspired and do your mood boards. But when you do that, you know what price point to hit. You know what trends there are. So we're putting them in a box and we're finding that they're as creative, if not more creative. Because if you give them an empty canvas versus giving them restrictions, we find that human creativity actually spikes when there's restrictions.

40:43So now we're able to, you know, have our cake and get it to. Our intros are probably 20 % as much as we were before. or things are turning and our market waters are more. And it's also comforting to people to say, hey, whatever you see here is going to get made. There's nothing more frustrating than seeing something at market and then, oh, it's dropped. Oh, it didn't get attention. So yeah, we've come a long way in that regard. Yeah, well, and I think that's all for the best. I mean, I think even though it was fun to see all of these pieces, to your point, nothing is more frustrating for a designer to leave market and then find out, Oh, no, you can't get that.

41:21Or the item gets discontinued just a few months after market because, of course, it didn't sell. And it's like I don't want people to feel, because when I say this to people that I talk about this topic, so that means your land is getting boring. No. The top seller of this market for Caraco is a sofa called Chrysalis. It's modeled after a cocoon that a butterfly. It looks outworldly. you know how we got there wasn't just because we stumbled upon a butterfly right it was through research was through data so it does work and it does promote pushing the boundary of the design too and is that uh is that part of what you learned in harvard business school mark is that education paying off for you in a big way or uh that specifically i actually learned from my dad the dad i didn't learn from school he's well and that's what i'm wondering so like i'm wondering what you learned in school versus what you learned from your father and how you brought those two schools of thought together.

42:22Yeah. So I think I'm just a weird person as you get to know me. I'm extremely nerdy, but I'm extremely artsy as well. Sometimes in a good way, sometimes in a bad way. So my father is my greatest mentor. He would tell me from a design point of view, We need to be Maya, you know, most advanced yet acceptable. All of these great designers design principles from a young age. And from Harvard Business School, it's kind of filling the gaps, right? Finance, you know, corporate strategy. But I would say the biggest thing that I've learned is leadership. 50 % or 60 % of my work is all people problems. It's all about the people that you have in the organization.

43:08You know, I give my people a ton of autonomy and I let them fail. You know, I let them challenge me. Like when I go into a meeting, you know, I don't want to be the person that provides the answer. I want people to prove me wrong. So I just had the privilege of learning that from the books. And also I had great professors like Clay Christensen, you know, who wrote The Innovator's Dilemma, who pinpointed disruptive innovation. I used to go to his office hours. Really? Yeah. That's such a seminal work. I mean, the innovator's family. Yeah, and he's a giant. He's 6 '4", and he's just the most gentle, the best.

43:51I go in one time and he goes, Mark, you have a brilliant mind. I said, oh my goodness. Wow. Coming from you. And then I was thinking, wait, does that tell that to all the students? He probably does, but it made me feel really good. But from that, right, like, you know, now I also have a startup called DecorX. So I'm thinking about disruption. I'm thinking about, you know, pushing the industry. So, yeah, definitely a lot of benefits from that. I find myself very lucky in that regard. So let's talk about DecorX and your startup that you have. Yeah, yeah. Long story short, I was back at campus for a reunion.

44:29A professor pulled me aside. He's like, Mark, check out Dolly. Here's a cat in space wearing spacesuits. I'm like, that's cool. And he's like, it's going to change your industry, Mark. And I said, okay, can you help me and tell me what to do? He's like, take my course online for$5 ,000. Great salesperson. So I did that. So you took his course online? I took his course. He gave me a discount, but I took his course. And I read a ton of books. I learned Python. I was in nerd mode. And I befriended an AI professor from Duke. And we started this project in our stores in China. It was basically take a photo of your room because we all have the cat drawings, the 3D modeling, the rendering that takes two weeks, then the consumers wait.

45:19So through that project, we were in the early days. This was before NanoBanana. We were able to train these AI models, fine-tune the Loras. Even earlier in the days, we used something called Dream Booth. We trained these models on what our furniture looked like. So we were able to take a photo and replace it with Markov furniture, Caracol, ARTO row. And it was working really well for our designers who worked in the stores. And I showed it at High Point. And a lot of our retailers and industry peers saw it, and they're like, hey, can we license it? and our interior designer clients saw it and be like, Mark, can I use your tool?

45:55So that's turned from this internal project into an actual company. And now we have products for retailers, whether you're a e-commerce retailer or you have physical stores. And we're also going to be launching a product for interior designers early next year. So currently what we do is in twofold. For retailers, as we try to build these agents and do these visualizations, we're finding that the fundamental problem isn't that you don't know how to build AI because that's actually easy. A little secret. It's very easy to build something that works 80 % of the time. But it's that 20 % in the end that hallucinates.

46:39That's really the problem. And the other thing is your data pipeline is everything. So all these retailers, they have 1 ,000 vendors. and the data they get from these vendors is crap, right? You have one solo and you have no product description. So we've trained an AI model to do that for you. You give us whatever you get from the vendors. We'll just do your product page for you with beautiful copy, all the right SEO, geo tags, but also other lifestyle photography. We'll take your brand standards, we'll take your art direction deck and we'll train the AI model on how you want your walls to look like, how you want your lighting to look like in batches.

47:19So that's been great. And then the product that we're designing for interior designers is, everybody's using ChatGPT, using Nano Banana, but these are kind of imperfect vehicles for the interior designer workflow. And what about my product library? What about pricing? What about this or that? Mood boards and colors. So we're building something that connects everything and our image models actually work a little better in Nano Banana in preserving multiple skills, the details in the furniture. So yeah, I mean, full nerd mode, and it's been very fun. Well, and with the idea that ultimately you want to build this tool that designers will use to help save them time, save them all of the rendering work, as you were saying earlier, that they used to have to do?

48:10For sure, for sure. I was on an AI panel yesterday and I talked about this kind of like the 80-20 rule. 80 % of the work that we all do, whether you're a designer or just any job is annoying, is time-consuming, is manual. So we're not going to try to replace your 20 % because I know people enjoy that. But that 80 % can be gone, especially with the help of AI. And, you know, what we've also found is that, you know, if you truly have a great interior design tool or, you know, a visualization tool, it's not going to help your best interior designers. Like, the best ones don't even need help. Like, they probably don't even need a rendering.

48:56Right. They'll just sell their projects, sell furniture left and right. But it's really the ones that, you know, your average ones or your young ones, you're able to lift them up and get them 80%, 70 % there. So that's what we've seen in all of our, you know, hundreds of in-store design associates. Like we can't make them all A's, but we've turned C's into B's. And that creates a big difference. Well, and for the companies that want to tie an AI, a generative AI model to their furniture business so that they can generate images of the furniture in situ so that people can go, great, I love that room.

49:39Can I just order all the pieces from you? How close are we to that model really working for you? And is that part of the end goal? Yeah, so the short answer is we're there. So for a lot of, yeah, we're already there. So for our own brands, a lot of our catalogs, for some brands in particular, they're 100 % AI. And you cannot tell the difference. We used to spend huge amounts of money renting out these houses, shipping them there and setting them up. But now we can do it with AI. And, you know, it's not like I know there's things out there in the marketplace, but they're kind of hard to use. Right.

50:21And, you know, if the first version doesn't work out, you know, you've got to make the adjustments. But every time you adjust it, it changes the whole thing. So our model kind of takes the opposite approach. We actually start by training the world, if you will. So we train the AI on the brand. And then we also have, you know, AI model that handles all of your SKUs. So we would ingest all of your SKUs and train the AI models on not only how it looks like, but the attribution, the material, the dimension. And then we just let AI do its work. And we're finding that it's working so much better. So do you find that to that point, if you, so if you train it very specifically on the product line, to your point, here are all the finishes, here are all the dimensions.

51:06does it no longer hallucinate in the same way and create pieces of furniture that aren't real or part of the line? No, yeah, so it doesn't do that. Now, there will still be 5 % of the time that it would hallucinate here and there. But we also have, so this is, sorry if I'm getting too nerdy, but we also have a GAN. So it stands for like Generative Adversarial Network. So basically it's two AI models competing. One is generating, one is evaluating, right? So after everything is generated, we have this quality control model. That's like, is this good or is this bad? So as they compete, they get better and better.

51:48So one thing that's unique to our AI workflow is the QC pipeline, if you will. So that we're able to do things in batches and try to avoid touch up the next step as much as possible. Because as soon as you start editing AI, it just never goes well. So you're hoping to take this to lots of other companies beyond your own and say, look, look at all these different things that we can do for you, to your point. Forget all those photo shoots. Forget all those renting houses and staging them and all of that for your catalog work. Yeah, so this is not just for my own brands. Obviously, we use our own brands to test and to train these models, but we're working with international e-commerce companies or retail companies that also sell the products of our competitors.

52:35And the beauty of that is now they're able to, for example, they used to only have three lifestyle photography per SKU, right? But now they can think about doing it for different geographies that they're in. It can be a catered lifestyle photography for Florida versus Aspen. It could be a whole new set of lifestyle photography for Christmas. It could be catered to the point that we can measure the IP address of your people and, you know, cater the lifestyle photography to that person. So sky is the limit. And a lot of big brands are thinking this way and we're working closely with our CMOs and CTOs.

53:16So who's going to be most disrupted by all of that? It's a hard question to ask. I think AI is on most people's minds that we talk to. So I think it's a good thing that we're democratizing creativity. But I do think the people who do not embrace this tool are going to be left behind. Now, these tools are not perfect, right? If we think about the last industrial revolution, when the internal combustion engine was first invented. These engines were not that great. And these cars, more importantly, the tools that connected the technology and the users were also not that great. So people still rode horses for decades.

53:59And they were able to perform the same job better on horses. But as these cars got better, you know, horses had no chance at the end of the day. So we do not know, you know, how fast this process will take, I'm sure your best horse riders in that way will still outperform your AI workflows in the short term. But just the ease, the expense, and the speed, you know, I predict it will happen much faster than the last industrial revolution. Yeah, yeah. No, I still prefer all of my telegrams to be hand-delivered on horseback. That's just how I prefer to communicate. Oh, that's the ultimate luxury now.

54:36Well, that's the thing. I mean, you know. But that's a great point though. I mean, it was remarkable to your point about the automobile, how long people resisted the automobile when it first came around. And it was noisy and it was spitting out smoke and a lot of people were leery. But AI, to your earlier point about how the consumer-driven market seems to be way ahead of where companies are with AI. And the furniture industry and the interiors industry in general tends to have a bit of a lag when it comes to technology, in part because they don't have the budgets to update their technology all the time, and it's understandable.

55:19And again, their focus is on creativity and design. They're not focused on the technology. Dennis, this is another thing that we covered on the panel yesterday. The question was, do you think AI will be good for creativity or bad for creativity? And the moderator was like, hey, look at our industry right now. Everybody's churning on the same thing. And so my answer was, well, it's not AI's fault. It's the algorithm's fault. The TikToks and the IGs, they're creating these echo chambers where you're only seeing what you like to see. So that's why we have 50 shades of ivory blue clay and all of that.

56:01And you know how I feel about that. I clearly, right? But the beauty of AI is like young people nowadays, Google isn't the first thing that they go to anymore, right? And the way that GEO is done is different from SEO, right? So AI is able to cut through algorithm and find you things. So I actually think this is great hopes for kind of our industry that's really, I will become more creative And then the other question was, do you think AI was more creative than the human mind? And my answer is no, right? Because if you look at how these models work, they're copied on how the human brain works, right?

56:48Like, for example, ChatGPT, they're transformers. It's not logic. How they come up with the sentences, they have enough data on what the next word will be. It's math, right, in a way. And then these diffusion models, they control the noise. So it's similar to how humans dream. So that's how they were able to create these outputs. But everything that these AI models do is backward looking. AI would never experience in real time, replicate the human experience. And how we get creative at the end of the day is, as we go through life, the curves we see, the magazines we read, the places. So I'm of the belief that we're never going to be truly creative because we're just remixing what we've ingested.

57:37I think biologically, it's impossible for us to come up with a truly innovative or non-existent idea, right? But AI will get us to do it faster, let us do it more, but it's not going to be at the tip of the spear. It's always going to be a human that starts with the idea. Well, that is certainly my hope. And we talk about AI often in the context of this fear of human replacement and all of that. But I wonder, Mark, as we wrap up, when you talk to your fellow high point furniture company CEOs and you talk to the high point market authority and you share and you and I have been on the panel together where we talked about AI and people were sort of, you know, wondering what it was all about.

58:25But what do they make of it all? And how much do you get a sense that they are eagerly getting on board with what AI can do to help their business in many of the ways that you've just described? You know, you can say that maybe we're in a little bit of an AI bubble or whatever, right? But like the C-suite people that I talk to, it's on everybody's mind. Everybody's looking at NVIDIA, right? Open AI. Just driven by the incentive of money. you have to think about it. And in that bracket, and then you start to quickly see, okay, who really knows their stuff or who like really wants to do it, not just for the investors, not just for optics, but actually wants to, you know, change the organization.

59:11In the early days, most of them wanted to do it just to do it. But now you'd be surprised, you know, your high point CEO, most of them is on their radar and they actually are thinking about layers, There's the work chart, different functions in their organization, how to change or how to improve efficiency. It can be a back office kind of thing or a front office kind of thing. So it's changed over the past year very rapidly as we're just bombarded with AI every day. I think it's on most people's minds now. It is. It's on most people's minds, whether or not I'm getting a sense of everyone, to your point, really folding it into their operations.

59:58I can't really tell, but I think it is remarkable to have the leader of such a huge set of furniture brands as yours also running an AI startup company and being as knowledgeable as you are about it. And so I wonder if a lot of people hearing this conversation will think, gee, I better hurry up and get on board because it is here. I mean, and you definitely see it in such a huge way. And sure, we probably are in a little bit of an AI bubble and we probably have all gotten a little carried away. And when you hear open AI sort of doing all these cross deals with companies, you wonder if it's gotten a little long in the tooth.

1:00:44but it is a revolutionary technology that is going to transform, as you were saying. It really is another industrial revolution. It's a lot of noise out there. I would just caution, there's a lot of companies out there too. Because we currently compete with a lot of VC-backed startups from Silicon Valley doing kind of similar things that we do. We're just puzzled by the amount of money that they've raised and the quality of their product. Like, what have you built with$50 million? So there's a lot of that because, frankly, I do believe it's a little bit of the bubble because most AI companies, they're building solutions first without knowing what the problem is.

1:01:28So how we've been able to scale with just a team is we're in it, right? We know what problems to solve. So it's a crazy, crazy world. Well, as you say, people have gotten a little carried away, but there are genuine uses for this technology in our industry, and you're putting many of them to work. And so we'll have to check back in with you and see the adoption rate after people get to hear this conversation. And hopefully some people will reach out to you wanting to know more. But I'm thrilled to get to talk to you. You're so kind. I know you must be running on fumes at this point at High Point, but you're so nice to take the time, and I'm always glad to talk with you, so I thank you very much.

1:02:17Thank you so much, Dennis. Always a pleasure. Thanks for listening. If you'd like to keep up with the latest design industry news, visit us online at businessofhome.com, where you can sign up for our newsletter, browse job listings, and join our BOH Insider community. for access to online workshops, a free print subscription, and much more. If you have a note for the podcast, drop us a line at podcast at businessofhome.com. If you're enjoying these conversations, please leave us a review on Apple Podcasts. It helps others to discover the show. This show was produced by Fred Nikolaos and edited by Michael Castaneda.

1:02:55I'm Dennis Scully. Thanks again for listening, and I'll see you next week.

1:03:04You