She Builds with AI: Women Shaping the Future

Fixing Online Shopping: The Future of AI Fashion Search & Discovery - With Chang Liu of Plush

• Julia Lach • Season 1 • Episode 10

Share episode

What if your shopping search could think like you do: mood, context, and all the tiny details that make an outfit feel right? In this epidsde I sit down with Chang Liu, co‑founder and CEO of Plush, to unpack how she transformed a personal frustration with endless scrolling into an AI‑powered fashion discovery platform that feels like talking to a stylist who actually gets you.

Chang’s path runs from Yale to Wellington Management, then through analytics and product at DoorDash, where she helped scale DashPass and learned to ship minimal, lovable products. Those experiences shape Plush’s DNA: clear moats, real unit economics, and a relentless focus on clean, structured data. We explore the technology unlocks that finally make intuitive fashion search possible: multimodal models that see silhouette and fabric, instruction‑tuned LLMs that turn “elevated basics for a fall birthday party” into actionable parameters, and stronger embeddings that map vibes to precise results.

But this isn’t about handing your wardrobe to a faceless agent. Chang argues for curation with evidence: fewer, better picks and a clear why behind each one, trend cues, fabrication, brand identity, and styling insights. We talk co‑founder fit, rebuilding hybrid search from the ground up to 3x retention, and real user moments ranging from “upstage the bride, but don’t” to postpartum black‑tie comfort. The future she’s building preserves the joy of browsing while removing the grind of filtering.

If you’re curious about AI in fashion, personalization, or how product, data, and taste come together, this conversation delivers both strategy and heart. Subscribe, share with a friend who loves style and tech, and leave a review telling us your most specific search you’d want AI to nail. 

Links Chang/Plush:

Links Julia/aiflow/Consulting: 

SPEAKER_01:

Hello and welcome to another episode of She Builds With AI, Women Shaping the Future, the podcast where we celebrate women who are shaping the future with technology. I'm your host, Yulia Lach, and today I'm joined by a guest whose journey takes us from finance to data science, from San Francisco to the heart of fashion innovation. Chang Liu is the co-founder and CEO of Plush, an AI-powered fashion discovery platform designed to take the endless scrolling out of online shopping. Think of it as talking to a stylist who truly gets you. For example, if I type in lightweight flowy maxi dress with soft task colors up to 500 bucks in size small, plush will understand that nuanced request and find matches across different brands. Before launching Plush, Chang studied at Yale, worked in finance, and later became data science and product manager at DoorDash. Inspired by her own frustrations with shopping, she set out to build a smarter AI-driven way to find the right clothes. In this conversation, we'll dive into her founder journey, the aha moment behind plush, the role of AI in transforming fashion discovery, and her vision for the future of shopping. And of course, we'll end with a fun peek into some of the quirkiest searches people have tried on plush. You're going to love this one. Let's get right into it. A warm welcome, Chan. Thank you. Thank you, Yulia. Chang, your journey has been pretty unique from being born and raised in China to studying at Yale, starting in finance at Wellington Management, and later transitioning into tech roles. Can you walk us along your career path in more detail?

SPEAKER_00:

Yeah, definitely. Yeah, I recognize my career journey has been pretty unique because I've, you know, down, well, I guess five different rows so far. But I do think looking back, those transitions have been pretty intentional. So when I started my career, you know, I was fresh out of college. I knew I wanted to sort of dive into the world of business, but I didn't know exactly what I wanted to do. Finance was the path a lot of my classmates took, and I just thought that was a really good way to learn about different business models. So I started off as a bi-side analyst at Well into Management, which is a big institutional active asset manager based in Boston. So I was investing in a sector such as consumer fintech software. I was also investing across different business classes like launch for equity, eventually growth, high yield. So I learned a lot about what makes different business models tech. So while I was there, you know, we're just constantly really inspired by entrepreneurs coming to IPO meetings that were solving a lot of real world problems. And for me, I had always just felt like, wow, okay, investing is great, right? It's very intellectually stimulating, but you're kind of an observer, like looking at how other people do things. Like you're not actually out there solving your own problems. So for me, I had always just found online shopping really frustrating. Like I really wanted to solve that problem myself. So I just really inspired by entrepreneurs coming through for these IPO meetings. And one of them was actually Katrina Lake. So I met her in my last year there. She came through for the Stitch Fix IPO. And I just felt really inspired because she felt like both she and I are kind of just regular girls who just are really into fashion. And but she's actually out there changing the way people shop. I just saw that's actually so inspiring. So that's actually one of the pimp moments that made me think, okay, I wanted to learn what it's like to be an operator and hopefully one day can be like her and changing the way people shop. So after that, I decided to switch to tech. I knew San Francisco was where the action was and every store is, obviously. And so decided to move to San Francisco. So my first shopping tech was actually finance strategy S Square. So I knew I really, really liked FinTechs, and I thought I could put my finance skills to use. So I was on the corporate finance strategy team. I was looking at capital allocation. And so for me, that was great role in terms of helping me understand like corporate finance. But soon enough, I realized that if I wanted to one day be a funder myself, I needed to get closer into products. So I taught myself data science and I became the first analytics hire at a social shopping startup called DOT, D-O-T-E. They are a social shopping app. I joined the company and I really, really loved the experience. I just was very passionate about the problem they were trying to solve, which is the discovery problem in fashion. I remember that was July 4th weekend. We were up in Tahoe, the weather was beautiful. Everybody like wanted to like go to the beach and like go on a boat. And I was there, I still actually looking at the creator payback model, and I just didn't feel bad about it because I was just so passionate about the work I was doing. So I remember thinking back then, like, wow, this is so much fun. One day I really wanted to go to a company in the same space myself. So unfortunately, my experience was kind of short-lived. The company didn't do well, they had to shut down. So after that, I thought I first really wanted to still learn, like deepen my experience in data science, ideally in an e-commerce adjacent company. And also I knew I wanted to eventually transition to product because that's going to teach me a lot of practical experiences about how to take product from zero to one. So I wanted to transition from a data science row to a product row. So the data science row at DoorDash really, really checked my box. So I joined DoorDash, was the first data scientist on the Dash Pass team. So Dash Pass was the subscription program of DoorDash. So when I joined, the team was super small, like four people with tens of thousands of users. And I was the first data scientist. So when I joined, I really helped them think through okay, how do you do different growth experiments? What are like the different things you can test? How do you improve retention? How do you think about the economics when negotiating a partnership with Chase? So a ton of really, really exciting initiatives. So over the course of two years, I helped them scale basically over 100x to 10 million users. And I also learned a lot about data science during that process. I became head of data science. I grew a team. So really enjoyed it, but I also knew, you know, eventually I wanted to fund a company. So I still really wanted to transition to product. So it was hard really saying go back to that row, but I knew they like product was a better row for me. So I switched to the homepage recommendation at Discovery Team and became a PM there. So that was a really great role because a lot of the stuff I learned there were directly applicable to what I'm doing today at Blush. So, and then after two years on the product team, where we shipped a ton of products, actually, I would say if you open DoorDash today, probably 80% of the products were either created or improved by my team. And so that's a really tough moment today. Whenever I use DoorDash, I still feel just really, really proud and happy. And then after two years there, you know, with the advancement in a lot of AI technologies, I was like, okay, I feel like I've learned it enough and the technology for doing something I want is finally here. So I finally took the leap and became a founder myself, even though there were a ton of transitions, but I do think it was a very intentional step by step to try to get to where I eventually want to be.

SPEAKER_01:

Super unique and super exciting career path. Really, really interesting. And I'm super excited for you that you eventually took that leap into founding your own business. Really, really exciting. You said that during your investing days, you met many entrepreneurs, including Katrina Lake of Stitch Fix, who inspired you. Maybe there were a couple of others. What was it about those encounters or their stories in particular that sparked your entrepreneurial drive?

SPEAKER_00:

Yeah, I think that honestly, it was it was two things is just recognizing a lot of those entrepreneurs like started off just like I was, right? I think Katrina Lake actually had a very similar experience. She started, you know, finance consulting and similar to me. And she saw a problem, and she literally just started the solution from her dorm room at Howard HBA and just went going door to door trying to solve that problem. You know, initially people even thought, you know, her solution was ridiculous, right? And so she had a bit of a difficult start, and that's just super inspiring because, you know, just like me, you know, I am someone who loves online shopping, you know, who has ideas about how to make it better. But she actually just started like out there doing things, right? I just always admire people who are out there on the field. That's something that really inspired me. And I think another thing was just their passion for what they're doing. I think just observing up close while those entrepreneurs were deeply passionate about the problem they were solving. And that really resonated with me. I do think to this day, right, you gotta be really passionate and mission-driven about the issue you're solving because being a founder is really, really hard. And if you don't have this like innate sense of mission to ultimately solve this issue, to make the world better, I think that like you're not gonna last very long. Those things truly are what inspire me.

SPEAKER_01:

I can really resonate with those reasons as well. Being a founder is super hard, and having that passion, like that intrinsic motivation is going a really long way. Chang, you've worn many hats. Investor, data scientist, product manager, and now CEO. How have those different experiences prepared you for running plush? Are there skills or lessons from those roles that you find especially valuable as a founder?

SPEAKER_00:

Yeah, I think looking back, I definitely feel very lucky because I'm able to draw on experiences from all those roles that I have done in the past. So being an investor really taught me how do I develop patterns that separate lasting products from hype. It really taught me a lot of practical lessons about what makes a business great. You know, is it, you know, what makes I'm constantly asking myself, what is the mote of the business? What is something they do that other people just don't do that well, right? And sometimes what makes the business great is just their ability to generate cash flow. And sometimes it's because they they have a great unit economics, even though it was they were in a hard sector. So DoorDash is one of those businesses. And then sometimes a business is great because they have great leadership, right? Just a really great team with very strong institution power. And sometimes maybe it's because of regulation. And I think being able to think through what makes a business model tick in different industries just gave me a sense of like, okay, if I were to do build my own business one day in this particular sector that is like e-commerce and shopping and discovery, it taught me a lot about the kind of business I want to build. And then later, as a corporate finance analyst, I learned a lot about the practical skills of like, for example, how do you build a feyback model, right? It's not just theoretical. You have to really think through all the expenses that are happening here and identify what's truly recurring and what's truly variable, and like combine that with accounting, make sure your business actually, you know, runs. And I think that is like a very, very practical skill, a very nice skill to have. Being a data scientist really taught me just like how do you basically read between the lines and numbers and see what that's telling you about your product and trying to get measurable signals from those data so that you're not just going with your intuition, you're going with very concrete, measurable evidence. And that has been a really, really good skill to have, even today. A ton of times, you know, when I'm feeling like, wow, like there's so many things we could be building that plush, my first intuitions always go back to numbers, like hey, like when the numbers telling us. Because when you talk to customers and they can tell you all different things, right? And it's really hard to talk to enough people to get a representative sample. Numbers, they just never lie. That is a really good skill to have. And then I think lastly, being a product manager, I would say I learned the most from their job. So I learned a lot about how do you basically have empathy for the end user, right? Because as a funder, it's tempting to build, for example, what's easy or maybe like what's really good for revenue or something. But ultimately, as a consumer product, you need to truly like stand in the shoes of the user and think about is this going to make their life better? If you build something, can they truly understand what the product is for? Do they immediately know how to use it? Right. And I think from a process perspective, I learned a lot about how do you work with different people, you know, data scientists, design, user, research, right? How do you really bring down the strengths in each and like build a team and then build towards a great product in the end? And also how do you ship something really quickly? So one thing I learned while DoorDash is this idea of a minimal, minimal lovable product. So people always say, you know, MVP, MVP, ship quickly. But I think it's really important to think, okay, like how do we do the bare minimum, create a product that still is lovable by people? Because if you want to get signaled on why something works or not, you need to have some sort of basic things that make people really, really love it. So using that concept really helps me think about how do you ship products, how do you iterate fast and your feedback. And then other things, for example, how do you decide what to build when there are like so many competing priorities, how do you prioritize? Those are all like really, really good skills to have. And I think another thing, and then there's just like a lot of technical things I'm learning that are just directly translatable because essentially I was working on the same set of features, discovery, recommendation, and search. So a lot of practical things, for example, how do you build a green merchandising carousel as dynamic, right? Or how do you like recommendation and precision products, like when to rely on heuristics, when to build a dynamic model, like all those things, things I learned at Ordash and like apply every day to the process of building plush. And I would say, lastly, one really important takeaway was understanding the importance of having clean structured data at the foundation of your recommendation products. When I was at Nordash, oh, this is really interesting. Like I would say three years ago, we actually didn't even have the data as to what we're selling you at the product level. You know, whether we're selling a burger or a sandwich. Just like didn't have that kind of data. And then we basically got those data labeled. And then immediately after that, we're able to build a ton of like great presentation products. So that made me realize, okay, like having that really detailed understanding of data, right? Large-scale clean structured data at a foundation of your product is is really, really important. So that's why today, when we're building plush, like that lesson always stays with me. Like we gotta leverage all the events as the AI to make sure at our core we have this like large-scale foundational data that's very clean structured. I deeply believe that's really, really important for whatever product we build. It's really important for e-commerce. So yeah, really lucky to be able to draw upon experiences and from all these different roles that I have done in the past, even though I took many twists and turns to get here. But I definitely think those different rows were not wasted.

SPEAKER_01:

Definitely led you where you are, made you who you are today. Really, really cool. I think like as a founder, it's like particularly helpful to be like all rounder, like to benefit from a lot of different skills and experiences that you had along the way, which you can all pour into your business at last. Really, really cool. Talking about plush, what specific problem in fashion or online shopping were you trying to solve when you started plush? Was there a particular frustrating shopping experience or insight that made you go? There has to be a better way. I'm going to build it.

SPEAKER_00:

Yeah, absolutely. The fact that searching online shopping was just so bad that like you just have to scroll endlessly every time. Like, I've had so many experiences where I'm like, okay, I like I sort of know what I want, but I just couldn't find it. So many times where, for example, I tried to find a dress to a wedding. Actually, uh a white party the night before. And so we needed to wear something that's like a white dress. And I was imagining something that's like structured and chic, but like not too revealing. I remember this back in like 2017 or something. It was a really trendy to have those like root high slits on the side of your legs. And I was like, okay, that's just not mean. And when I was going through it on the website, I was like, okay, how do I just not see them? But everything I see is high slits or like exposed belly or like something that's just not structured enough. So I had always found that to be super, super frustrating. And then ultimately, I think that Moo Causes, you know, on our shopping wasn't built for how humans think. It was built for how databases think, right? So that caused two issues. One is these sort of filters and like search capability are just not nuanced enough. So going back to my example, you can't just simply say like not no size list or like no strong shoulders, right? Or like solid color. And you also search also doesn't understand those nuance criteria. So that and women's tastes are you know very nuanced. We're never just looking for just a white dress, right? There are actually a lot of those like preferences. So yeah, so I think that's one big issue. And the second issue is, you know, online shopping doesn't really understand how humans actually think. You can't really search in a way that you actually think about fashion. So for example, people usually say, I am I'm thinking a cozy sheet outfit for Thanksgiving without trying too hard, right? That's how they actually say, but you can't just type that into your search for you have to really think hard and translate that into maybe some sort of filters or plain language. And that just made stopping very frustrating and not intuitive. And with a lot of resonance in AI, I just really saw the opportunity to completely change that and transform it into something drastically more intuitive, personalized, and delightful. So that's really the problem I was going after.

SPEAKER_01:

Yeah, talking about new AI capabilities, you started developing plush around the time when new capabilities like multimodal models became available. How did these advancements make plush possible in a way that couldn't have been done just a few years ago?

SPEAKER_00:

Yeah, absolutely. You are absolutely right that, you know, I think plush wouldn't have been possible even just two and half years ago. So they're having a lot of capabilities that have made buildings and like plush possible. I would say number one is multimodal understanding of the model, right? In the past, models can deal with languages, but fashion is very visual, right? There are probably a million black lace dresses out there, but what makes them all different from each other is these subtle visual cues, right? Things like the fabric, how the fabric looks and feel, and the silhouette and like how revealing the neckline is, and like how the shape of the shoulder, like a million different things, right? So when the models are trying down these hundreds of millions of image and text pairs, they start to put together all these image style cues and text. That's when models can truly see fashion the way sort of humans do. That makes building a real foundational understanding of the fashion database possible. And I think the second thing is just the advancement of instruction tuned LOMs, right? Like these large language models can not only understand text, the complex cues behind the text much better than before, but they can also translate that into a set of very concrete parameters that our system can understand. So they can really advance the systematic capabilities on top of it. So today, when you type in something like elevated basics, the model can immediately understand. Like when you say that, like what does that mean? Well, it means maybe it's neutral colors, maybe it's structured fabrics, maybe it's not much logo, maybe it's high-quality fabrics, right? So all those contextual understanding, and we then translate those translated language, right, into a set of parameters our search can understand. So that capability makes it possible to build the kind of search that we're building today. And I was saying the third one really is I would say just like much better embedded models. You know, we have those embedded embedding vector embedding models before, but the arrival of large language models just make them a lot better. And nowadays we can easily get a representation of the search query. Hey, I want something that's modernistic for a wedding in a Tesca villa, you know, you can get a vector representation and maybe match it to like the embodies of our products. They can do that on the fly, and that's really, really powerful. Or if for using another example, right, when it comes to precisation, when you're trying to match people's favorite brands, you can easily get a brand representation in the vector space. So you can use large language models to summarize, okay, what's unique about this brand, and then you made the legendary embeddings of that, and that can be done really, really quickly. So that makes civilization possible, possible like never before. So those all have really exciting advances that just made plush possible today. They're really exciting. You really started at the right time at this day and age. Yeah, and uh, you know, the miners keep getting better. We we see that every day.

SPEAKER_01:

And you also take all these advancements into account every day. Uh it's just like moving so so fast. Yeah. Super exciting. Chiang, you co-founded plush with Ying Ju, who is now the CTO. How did you two come together for this venture? And what strength do each of you bring to the table in building this AI-powered fashion platform?

SPEAKER_00:

Yeah, absolutely. So, my co-founder Ying, we were actually both on YC Cop undermatched. It was a pretty much like the co-founder daily platform. You know, we just got lucky. She's one of the first people I met through the platform. I remember we, you know, met in person for brunch. Sometimes we see a person within minutes who just felt like they're fit. She was very fashionable that day, and she was wearing a bouteille bag really light. And also both, she and I are Chinese. We're both really tall. Like we're both like about 175. And, you know, immediately felt like this like affinity. And another thing about us is, you know, at the time I was a content creator, but a style influencer on Instagram, and she was a fashion photographer. So we also started should be some content together. We just felt like, you know, on fashion, we were a match. I can tell she's just, you know, passionate about fashion. I immediately realized, okay, this is gonna work because we brand a ton of complimentary skill sense, but at the same time, we can really talk on the same wavelength. So I do everything that's not poor like coding, right? I can do finance, I can do data science, I can do products, and I have maybe like top investors, and she does everything. Cody, she's the one that actually built the product, right? She actually worked on many different teams at Airbnb, and then she just brings with her a ton of experiences. You know, one thing I like about her is like she's not just someone who a lake holds, she actually is also really into the process of like building and funtoing a product, and she has such setting eyes, and the thing that has to do with the fact that like she is a photographer, but whenever we design a product, you know, after I design her was the designer, and she will look at it and she will always come up with like a really, really great feedback. I was like, wow, like I've never thought about that before. It's it's really great to sort of have a partner who can also like give a lot of feedback on the work I was doing. And I wasn't just only doing things not related to like a co-product, I will help her, for example, fun tune a prom. I will help her look at the data that we're grabling for the items in our inventory and help her think through how and make it make that process better. So even though, you know, I'm not an expert in her domain, she's not an expert expert in my domain, but we're able to give each other feedback that way. She has a great working style. I really like. She's very honest and direct, you know. It's like really great to be able to give each other immediate feedback like that. So yeah, we have to follow like a great working relationship.

SPEAKER_01:

Amazing, amazing. That sounds really exciting. Like you two sound really like the perfect match, complementing each other and you know, mashing each other's strength and weaknesses. Really, really cool. And yeah, you spend so much time together, no, with your co-founder, and it's so important that you get along and that work is just like fun, also.

SPEAKER_00:

Very your work husband, right? Exactly. So you know the mayor's break personality type. She has the same type as my actual husband. I think that might be one of the reasons why we're together so well. That is so interesting.

SPEAKER_01:

Look at that. That's so cool. All right, since launching Plush, how's the journey been for you so far? What kind of response have you seen from users? And are there any big wins, milestones, or feedback so far that you are particularly proud of?

SPEAKER_00:

Yeah, definitely. So we've been working on plush for about a year and a half. I think definitely, you know, similar to a lot of other founders, there have always been ups and downs. I remember just at the beginning, it was, you know, really just like trying to figure out what we're doing the right thing, were we seeing signos of product market fit. And when we immediately didn't see that, we were like, you know what, we gotta keep our heads down and work on the product, just search and make sure it's really, really good. So we actually had to sort of build our entire system from scratch, build our own sort of proprietary AI-powered hybrid search. And then after about that and and this few other making a few other improvements, we saw that retention of 3X from the beginning. So that was definitely a really prompt moment seeing that okay, like we are not getting the connect customer law we wanted. But then after being has done and working on the thing with that we're gonna do which is make search drastically better, we actually achieve our goal, and that's like a really, really prompt moment. Then I would say another, I think some of the other prompt moments had always been just like moments from moments from hearing directly from our customers, customer, you know, emailing us like this is like the first fashion tool that actually understands the way my brain thinks about fashion. Now it's such a validating moment. It makes everything work. There is always a ton of rejections you have to deal with, like hardships, moments you don't know what to do. Hearing them directly from customers, I'm always paying honest moments.

SPEAKER_01:

Yeah. So rewarding, really, really cool. Yeah, and that makes it all worth, right? As a CEO of an AI fashion startup, what exciting developments are you seeing right now in fashion tech? And if you imagine this shopping experience in five to ten years, what does it look like?

SPEAKER_00:

Yeah, I think that I think broadly this whole trend from that's this whole shift from transactional shopping to more contextual understanding, like mood-based discovery, that has been a trend, you know, fundamentally we've been very excited about at plush, right? And so I think it's it's still a learning curve for people because some people come to plush, they still say, Oh, I want like a brown meaty dress with so-and-so characteristics. It's kind of a bit of a learning curve for them to understand, oh, you know what? I can just say I want a structured and chic dress with ursy colors that's good for a Thanksgiving dinner, right? That's still a learning curve, but I do think people are already getting much better at it. As in one way, people learn to use tools like Trat GPT, they're realizing that oh, like tools like plush can just simply understand inherently very subjective, aesthetic, and contextual preferences in mind, and like translate that into shoppable results. And I do think that, you know, people have been talking about all sorts of trends like you know, like AI trials, like better opposite optimization. But I think personally being really excited about is number one, just like drastically better personalization, right? A shopping tool that truly gets you. For me, two keywords are curation and evidence, right? So curation is really important because essentially you don't want these tools to give you results, right? Some tool to help you narrow it down to maybe just like 10 to 20 results. How do we get to a state where no one you say you need a dress for this wedding, plush will give you only 10 or 20 results? And from that top 20 results, you can't find four to five dresses. You were just like, wow, like those are stunning. I'm so excited to buy and like try them on and like wear them, you know. And that's going to be a hard problem to achieve. And then the scene can we also want evidence, which is okay, tell me why I shouldn't like this dress. Right. Because a lot of times when you think about like the best job experiences, it's never just about like seeing the dress, it's also understanding like what makes it so cool. Like sometimes it's as my favorite influencer dress will wear this. It looks so good on her. Right. Sometimes it's understanding, oh, this kind of ruffle is really in. And like I want to wear a ruffle like this, similar to the Transessie on the runway. And sometimes it's understanding, okay, this brand is really, really chic. Like, like only the insiders are like wearing, I want to be like those like cool insiders wearing like the latest, like hardest brand, right? So if we can get to a stage where plush can come out with like five to ten dresses that are great, and we explain to you exactly what makes it so we so cool to wear, again, really excited. I think that's the end stage I want to get to. And for that to happen, we not only have to have really great personalization, but also have a, you know, a tool that thinks like a fashion editor, right? We're not just figuring out what you're trying to wear, but also we're kind of dictating a lot about what you should wear, right? And like we're trying to let like understand your preferences, but at the same time, like like a fashion magazine, tell you, you know, what are the trends. And I think another one, another trend I'm really, really excited about is basically, you know, how do you build truly a shopping experience that doesn't feel like what work? People have been talking about this idea of agenda shopping, right? Where agents just understand you so well that they shop on your behalf. But I think we need to define like what that means because you know what I think people some of this would get is shopping is an inherently joyful process, right? There's a lot of joy in the hunt itself, right? A lot of women don't like to be able to be handed like one or two stages, be totally like, okay, that's a thing you should wear. So how do you design this experience where it's people are only browsing beautiful things that they will like, but you know, something that doesn't take the joy of browsing away because there is a lot of hand joy and browsing and just looking at beautiful things. I do think, I do think that designing that kind of ultimate shopping experiences where it doesn't feel like infinite scrolling, but it is still delightful. It's it's a really it's a it's a model that like everyone industry needs to figure out. And I think, yeah, the end goal is definitely not this like complete agenic shopping where you know we don't get to participate in the process. I think fashion lovers will still want to look at beautiful things. Yes, definitely.

SPEAKER_01:

Since Plush lets users search for very specific things, I have to ask a fun one. What's the quirkiest or most memorable search where you've seen a user enter all that you've tried yourself, maybe something that made you laugh or think, wow, that is specific. And did Plush manage to find a good match for it?

SPEAKER_00:

Yeah, definitely. There's so many quirky ones. I was in one of them. There was a user that said, I got invited to a wedding where the dress code is upstage the bride. How do I find dresses like that without actually upstaging the bride? So I was like, oh wow, that's such a challenging question because yeah, you need one wear something super memorable for that location of weather. But at the same time, you know, you don't want it to be like overboard. You don't want to be like too revealing or something. We actually, I think I did a like a pretty good job on that one. And then I think I think the most memorable ones are always the ones where the user actually was willing to be vulnerable with us. I remember there was this woman that searched, I don't remember to this black tie gala in the city, and really want to dress the impress, but I'm also three months postpartum. Like, how do I find a dress that like doesn't really accenture my belly, but also, you know, black tie or even white tie appropriate? So I thought we did a really good job because we actually showed her dresses, you know, those formal gowns without the defined waist, so that you know, you you can find really structural, beautiful silhouettes that doesn't actually accenture your belly. So that was a really, really proud moment for me. Amazing. Amazing. So cool that plush could deliver even with these very, very specific, very personal requests. Yeah, I think most people don't realize is you know, we can accommodate so many nuance criteria. Things like we see people searching for like in-seam lid for jeans, and that way can accommodate too, right? Because some people might be really tall, really short, they just want a particular in-seam for jeans, which tradition traditional shopping just doesn't accommodate. Some people care a lot about natural materials, no synthetics. And then some people just like want only want dresses have a that have a fitted, a flexible waist. So all of those things we can we can deliver, and the more you use it, the better it gets.

SPEAKER_01:

Thank you so much, Chang, for that super inspiring conversation. That was Chang Liu, co-founder and CEO of Plush. I loved hearing how she combined her background in finance, data science, and product with her passion for fashion to build something truly innovative. The story is such a powerful reminder. The best companies often start with a personal frustration and the courage to solve it. If you enjoyed this episode, make sure to subscribe to She Builds with AI for more inspiring conversations with women who are reimagining industries with technology. And if you'd like to learn more about Chang and Plush, you can find all links in the show notes. Thank you for tuning in, and until next time, keep building with purpose and keep building with AI.

Head Shot

Julia Lach

Host