The Edit is our weekly show, where Social Editor Grace Robinson quizzes editor-in-chief Ben Hanson on some of the most significant fashion and technology stories from the past seven days.
This week on The Edit, Grace Robinson and Ben Hanson discuss how Stitch Fix uses customer feedback to build its private brands, why SHEIN is expanding its logistics hub in Poland, what AI search volumes and shopping surveys tell us about agentic commerce, what fashion technology loses with Coperni’s closure, and how the political backlash against AI infrastructure could affect fashion brands’ AI plans.
Watch the video above or listen to the audio below. You can also find the show on Apple Podcasts and Spotify.
For deeper interviews with brand and technology leaders, our prestige interview show releases on Thursdays everywhere you find podcasts – and will be coming to YouTube soon.
Subscribers to our mailing list will receive notifications for both shows.
Grace Robinson: Welcome to The Edit from The Interline, the show where we run a quickfire analysis on our pick of the most important fashion and beauty stories from the last seven days.
I’m Grace, the social editor, and I’m joined by Ben, the editor-in-chief. And together, we have around twenty-five minutes to give you our analysis on the stories that we think really matter.
Ben Hanson: Hello, Grace. How are you doing?
Grace Robinson: I’m doing good today. How are you?
Ben Hanson: I am fine. It’s getting cold now. I’m also trying a beard. I don’t know if this works. I don’t know if I can carry this off. If you’re listening to this via audio, this is no different. This might be the one time viewers on YouTube go, ‘That guy can’t do that.’ But I don’t know. Let’s see if it persists.
Grace Robinson: I think it works with the jacket you’re wearing.
Ben Hanson: Maybe. Maybe that’s what it is. A bit of a lumberjack-y sort of feel going on, which is the first time anybody’s ever said that about me, I think. We’ve got a ton to get through, though, so let’s roll. Hit me with what we found.
How Stitch Fix turns customer feedback into products
Grace Robinson: Okay. Firstly, there was a great piece this week on Stitch Fix in Modern Retail. It specifically highlighted how Stitch Fix uses the constant stream of consumer data it gets to address product gaps in people’s wardrobes. It then designs its private brands around these needs to give consumers exactly what they want.
The piece highlights how AI, of course, supports this workflow, as well as how this technology is helping the brand to summarise trends, kick-start the design process and even collect client ratings. I know that The Interline interviewed Stitch Fix last year. So tell me your thoughts on this story and what you make of this approach.
Ben Hanson: Yeah. It’s a really interesting one. The simplest thing I can take from it is it underlines the importance of data to AI strategy. No matter what you’re trying to accomplish with AI, the prime determiner of whether it’s going to be successful or not is how much data you have available to you.
And the interesting part about Stitch Fix is, I think it was Tony Bacos, I forget his role and I don’t know if he’s still in it, but he wrote a piece for us last year for the AI Report 2025, and that was billed as lessons that you can learn from Stitch Fix as an early mover in AI. And they were an early mover in the sense that they’ve existed for a long time. I forget when they were founded exactly, but I suspect it’s a decade or more. And it’s always been an algorithmic business.
The goal is to provide human-signed-off, but to some extent algorithmically derived, style recommendations to people. And to do that, you need personal stylists. You need experienced human experts. You also need a lot of information about what people wear, when they wear it, why they wear it, what they wear with other pieces, when they want it, how it aligns with particular seasons, how it aligns with events and occasions and things like that.
And on top of that data, they’ve been able to architect a huge amount of new visibility into what they do with private-label brands. So I think they have ten in-house brands, and the information that they now have access to as a result of all of that foundational work that they did for a decade plus with machine learning is information that they’re now feeding into their own brand development, refinement and so on.
What makes this especially interesting is that this is something that your typical brand, like an individual brand, does not have access to. Even your individual retailer does not have access to this kind of data. They know what they sell. They know what they don’t sell. They know at the colour level, at the style level, at the size level, what goes out the door, what comes back. They have limited visibility into what people do with it afterwards.
That’s a lot of the thrust behind loyalty programmes. They’re not just there to get people to spend more money and to upsell. They’re there to try and build more of an ongoing channel of communication between the brand and the customer, because the more you know about how a product is actually used, valued, not valued, paired with something, not paired with something, the more you can bring that intelligence back into your next cycle of design development, collection building, assortment planning and so on.
For brands and retailers, the only way aside from loyalty programmes to try and build some of this seems to be to maybe partner with some of the wardrobe quantification, wardrobe intelligence platforms that you and I have talked about before. We talked about Whering, W-H-E-R-I-N-G, when they did a big raise relatively recently. We talked about Google’s efforts to sort of do virtual try-on and wardrobe building by scanning your Google Photos. I’ve just recently, in the last week or so, interviewed the CEO of Whering. That episode will be coming out sometime soon.
But to give you a bit of a preview of it, the brand value of those kinds of platforms is not, ‘Here’s a ready-made community that you can advertise at.’ There’s a little bit of that. It’s more, ‘Wouldn’t you love to know what people are doing with your clothes? Wouldn’t you love to know if you make tops and bottoms, but you don’t make bags, and people keep pairing your tops with this particular style of bag? Wouldn’t you love to know if that’s a gap that you can fill in your assortment?’
So as a regular kind of brand or retailer, you have to take a patchwork approach to this. You have to get the data where you can find it through loyalty programmes, ongoing engagement and those kinds of partnerships. If you’re a business like Stitch Fix, which deals in obviously third-party brands but also their own internal labels, you have all of this information baked into your business model because you’re shipping fixes, as they call them, boxes out to people, and you see what they do with it, and you see what they pair with it. There’s just a huge amount of information in there.
This sounds like a digression, but it’s not. I had a conversation with a CEO that I know who started a scanning company, which was based on the idea that they would do 3D captures of footwear so that brands could then put those 3D captures on their websites. They very quickly realised recently that it’s super lucrative to just scan as many things as you can across consumer products, home, textiles, everything else, because the data that that generates is incredibly valuable for physical AI and for home robotics and things like that.
There’s just a huge amount of value in data right now. And if you’re the kind of business that’s been doing this for a decade or more and you have that data, you’ve got a real edge on people who haven’t been doing it. So it’s a really good AI story in the sense that it’s not just about generative campaigns or a bit of advertising. It’s about a fundamental business model that has been accruing the essential matter that you need to do a good AI strategy over a long time, and suddenly they’re able to demonstrate what you can do with it. You’ve seen bits of it from Amazon and their private labels before, but this is a really good case study, and I encourage people to go read it.
Why SHEIN is expanding its logistics hub in Poland
Grace Robinson: So as we spoke about over summer, in July the EU introduced a €3 charge on parcels worth less than €150, a measure that, of course, has put a lot of strain on businesses like SHEIN, Temu and AliExpress. In response, it was announced this week that SHEIN is investing €500 million to expand its European logistics hub in Poland, specifically to counter the rising EU import charges and also to speed up deliveries.
Now this isn’t an uncommon approach by brands. Decathlon, for example, also restructured its network to lower transportation costs and to cut transit distances. So again, I know The Interline covered this in a previous interview with them. So give me your take on this news and this story.
Ben Hanson: Yeah. It’s an interesting one. So the conversation that you’re referencing was a podcast interview I did with a lady called Oliwia Guziel, I think. I may be mispronouncing her last name. I knew it at the time I did the recording. So she’s from Decathlon in, I believe it was Poland or the Czech Republic, I forget off the top of my head. But it’s a fascinating case study, that one. I encourage people to listen to it for two reasons.
One, it spotlights how different consumer attitudes are to delivery in different markets. You and I are both in the UK. I think we have a very UK-first mindset when it comes to what people want from e-commerce, where they want things delivered. One of the big takeaways for me from that Decathlon conversation was there’s a very different consumer delivery culture in those kind of Central European, Eastern European markets, where people prefer to have things delivered to work, for instance, or prefer to have things delivered to lockers, in a way that doesn’t quite manifest itself in the UK.
The other part of it was it’s a tremendously complex operation that Decathlon has set up there, which caters to two different countries from one central hub, and that aims to do, I believe it’s twenty-four- or forty-eight-hour delivery, as a test case for just how fast things can go and how quickly things can reach people. There’s a lot, as you would expect, of automation in that model.
And that seems to be the case with what SHEIN is doing here as well. I’m just looking back at the statistics. I think they talk about deploying something like six and a half thousand autonomous robots to support warehousing. It makes sense, right? You wouldn’t choose to build something in Europe if it was solely going to be staffed by people. But I think the other key thing to realise from this is they’re also hiring a lot of people. I think they’re talking about the 10,000 employees by the end of this year for this geography.
Now you and I talked last week, I think it was, about the crazy high shipping and logistics costs this coming holiday season. I think it was The Business of Fashion who did a case study of how that’s all playing out and why brands are turning to different shipping providers and different delivery providers and so on. There’s a strong argument for doing more of this. There’s a strong argument for having as much in-country distribution as you can.
Yes, okay, SHEIN are doing it for regulatory reasons. They’re doing it to try and get around parcel importation fees. That sounds dodgy, but in a way it’s no different to the loopholes that you see where something is made in America, for instance, but is only finally assembled there. We have to realise that we do deal and live in a global market.
I think we’ll see more of this. I think we’re going to see more of these so-called digital disruptors who European countries are trying to regulate away their business model. I think this is their natural response, to say, ‘Well, okay, if the issue that you have is that we ship a bunch of parcels from China and we try to get around loopholes, there are other ways that we can do this, and there are ways that we can manage distribution closer to the country.’
Distribution is an advantage, speed is an advantage, convenience is an advantage. There’s a big technology component to this story, as we talked about automation, but it’s also, at the end of the day, a first-principles economic, commercial story. And again, it’s an interesting one to read about, and I’d be keen to see how it plays out over time.
How AI search is changing the path to purchase
Grace Robinson: Next, we’re talking about agentic commerce and AI search. So McKinsey surveyed approximately 21,000 consumers across 11 markets, and some important stats to highlight from this survey are that 53% of global consumers have used AI for shopping in the past three months and 14% of global AI users have used AI to complete a purchase. On top of this, apparently, AI users are 3.2 times more likely to rank AI search as more influential than traditional search engines.
So what’s interesting about this is that it all fits in with the news that OpenAI now has around 20% of Google’s search volume. So talk to me about this and what you think it means for fashion.
Ben Hanson: Yeah. Do you use AI for search? Have you found yourself moving away from Google as a front end and using ChatGPT or Claude or what have you as a search engine?
Grace Robinson: It really depends what it’s for. I think I use it more and more to give me lists of different options, but I’m very aware that that is only a limited view. But if I really need to know some options for restaurants or something like that, it is a good way to just get an overview. So I’d say I’m fifty-fifty at the moment, but it is creeping in more and more.
Ben Hanson: Yeah. I think I’m about the same, and I think it’s partially affected by the fact that Google Search gives you AI Overviews as the default anyway, at least in the UK, the US and a few other markets. So you’ve always had some measure of, not quite curation, but you’ve always had some measure of preselection done in a Google search in terms of what’s on the first page, right? There’s so many studies out there to say people never get past the first ten blue links and get onto the second page. I found myself doing the same thing.
I’ve found myself doing something maybe weird, maybe not, which is I’d rather use ChatGPT for that search than I would use Google, to pick your restaurant example, just because I have ChatGPT plugged into an external search provider, Firecrawl, which I trust to do a decent job of search out there. I would still, if I was looking for a restaurant, so Manchester in my case rather than London in yours, I would go to a specialist publication now. There’s one called Manchester’s Finest, which tries to keep on top of the restaurant scene and bar scene and stuff in Manchester.
I’ve got three kids, so I don’t go out a lot. When I do, I want to make sure I go somewhere decent. I turn to human experts, and I turn to human experts in a way that I go direct to the website. I’ll use ChatGPT for that initial kind of canvassing sometimes and a bunch of other stuff. But the interesting part here, I think, is that behaviour that you and I have just described is fairly common. I think a lot of people use AI in place of traditional search.
When you say 20% of Google search volume, we don’t mean that it’s stolen that away from Google necessarily. What we mean is, of the people that exist and were surveyed for this, about a fifth of people use AI for search compared to the total volume of people who use Google for search.
The other interesting story that I saw this week that pairs with this one is the CEO of Uber was talking about how many people place ride requests through AI agents. And they did it on Halloween, which is apparently quite a busy time for Uber rides. I imagine a lot of people go to parties and things and need to move around. And it was like single-digit people across the whole of the US actually placed an Uber ride request with an AI agent, compared to the number of people who just open the app and do it.
And the interesting statistic for this one, so behind all of this research that McKinsey has done and the Big Technology newsletter have done, is yes, 14% sounds like a lot of people using AI to complete a purchase, but it’s 14% of people who already use AI. It’s not 14% of shoppers or 14% of the overall market. That’s not a great conversion rate, really. If you take your own user base and one of the functions that you’re selling and advertising is order placement, placing Uber Eats orders or ordering a taxi or completing a transaction or what have you, if only 14% of your user base are doing it, it’s not a significant behaviour just yet. Now this McKinsey study does seem to have been based on summertime data rather than something a little more up to date. So I think it may have changed a bit over time.
The final interesting bit that they did find, which I do think is valid and I do think is interesting, is that AI assistants specific to retailers’ own websites trail standalone AI tools by as much as two times. So if you go on the website, just pick a hypothetical brand, and they have their own chatbot designed to help you navigate through their selection and find something for what you’re looking for, more people would prefer to go through this process with the AI that they already use, ChatGPT, Claude, Gemini, whatever it is, than would prefer to interact with another chatbot on a retailer’s website and start to fragment the journey somewhere there.
I think all of this just speaks to the fact that AI is here in the sense that people are actively using it. It’s on most people’s phones. It’s taking a meaningful share of search volume. The agentic stuff that everybody is very keen to say is happening, I don’t know if it’s happening at the volume that they want it to be happening at. And I think we need to keep a close watch on this as we go into next year to see how it plays out.
I have also not placed an Uber request through an AI agent, and neither have I yet done a full end-to-end shopping transaction through an AI agent. Maybe I will, and when I do, we’ll talk about it.
What fashion tech loses with Coperni’s closure
Grace Robinson: So a huge fashion headline this week has been the news that Coperni is shutting down, and this is following a dispute with its distributor and majority stakeholder Tomorrow.
The news is relevant for The Interline because the brand is really known for its speculative fashion tech designs. Some of the most viral examples include the Bella Hadid spray-on dress, the Boston Dynamics robot dog on the runway, the Tamagotchi Swipe Bag, the Ray-Ban Meta smart glasses collaboration. The list literally goes on with examples like this.
So in that way, the closure is a shame because their brand was a way for more general fashion lovers to understand fashion technology concepts. So what do you make of this story and what’s your take on it?
Ben Hanson: It’s sad to see any brand shut down for what sounds like business and personal reasons. Disagreements behind the scenes are never a nice way for something to finish. And I lament the closure of any brand that is interesting at the product level and at the marketing level.
They were really good provocateurs, right? Aside from a real art collective like MSCHF, I think Coperni were the most successful of the ‘We’re a real fashion brand, but we’re also going to make a nuisance of ourselves and create a bunch of conversation through innovation and through technology.’
I have a conflicted relationship with that because on the one hand, it’s interesting to put a robot dog on the runway, I guess, as long as there’s something that you’re really saying about it. And I’ve never really figured out what they were trying to say with some of this provocative stuff. I’m also one of those people who, with some of the edgy or modern art and stuff, I tend to think, ‘Cool. But what’s behind it? What’s the message?’ Sometimes the message is just, ‘Hey, look at this. We’re being in your face. We’re starting a conversation.’
I’ve had so many conversations with people who don’t work in our industry or who maybe work in fashion but don’t interact with the technology component to it. When you say, ‘Oh, I work in technology for fashion,’ they’ll be like, ‘Oh, cool. You mean like the spray-on dress? Or you mean like the robot dog?’ And I’m like, ‘Yes and no.’
I’m glad that those conversations are happening. I’m glad those things are in front of people, but there’s also a part of me that feels like some of that provocative stuff clouds the real role that technology is supposed to play in helping brands to make better products, create better experiences, do more interesting advertising, build more data-driven collection development and stuff.
And I’m aware that as I’m saying this, I sound like a boring person. I sound like somebody who works on spreadsheets and back-end systems and stuff in fashion, and I have no place talking about runways. But there is that part of me that’s like, they did a great job of using technology to get people talking. Sometimes that talking ended up detracting from what technology was supposed to do for fashion rather than actually being productive in that kind of way.
But I do think I will probably look back on this in six months and be like, ‘I kind of miss some of that provocative role that technology plays on the runway,’ as long as it’s not done the way that we talked about with Vogue World a couple of months ago.
What the backlash against AI means for fashion
Grace Robinson: Now we’re ending this week with another AI story, but it does have a slightly different angle. So it’s no secret that AI has been getting a lot of backlash and this backlash is really snowballing in fashion from all angles. So much so that even a handful of luxury brands are now purposefully commissioning real artists instead of using AI to generate campaign imagery.
Now zooming out, the backlash against AI and big tech can also be seen in political news as well. For example, the 2026 US elections showed that voters are really pushing back against data centres, and this is amid the concerns around electricity bills, pollution, jobs, surveillance and also the influence of big tech executives. So the backlash will probably only increase. So what do you think fashion brands need to be thinking about when it comes to their AI strategies?
Ben Hanson: Yeah. So we’ve talked a few times about general surveys and stuff that provide a referendum on how people feel about AI. And depending on where they are in the world, you and I talked last week about the fact that sentiment is generally more rosy in emerging markets than it is in countries like the US and the UK and in Europe. Generally speaking, people don’t seem to like it.
And it is a very nebulous and exotic sort of feeling because on the one hand, you have people saying, ‘Well, I don’t trust AI. I don’t like it. I don’t see it in my advertising.’ Then you see the studies that we just talked about a minute ago, which is an increasing number of people using AI in place of search or starting to use it for agentic shopping or so on. That’s always been there in polls, it’s been there in sentiment. We found some of the same sort of stuff in our AI report survey from earlier in the summer, although the electricity use and the ethics and stuff kind of came way down on the priority ladder compared to people who just thought that the capabilities were maybe a bit inflated.
What we haven’t had is a direct referendum on everything that you just talked about, like the infrastructure build-out that needs to come with it. And you and I are recording this a few days after the US midterm elections, which have demonstrably turned against the Republican Party and towards the Democratic Party. And there’s a lot of outlets that are framing that as exactly that kind of referendum on AI, because the current US administration has been very permissive when it comes to allowing technology companies to build out the data centres or compute campuses or whatever OpenAI decides it wants to call them now in states across the US.
Texas was a particularly strong example. It seems like that was a big factor in some of the voting behaviour that occurred in the US. A lot of it is down to cost of living, I’m sure. A lot of it is down to other policies. A lot of it is down to trade and war and everything else. But it seems like this is evidence of the fact that people are willing to vote on AI as an electoral issue. They’re willing to vote on the fact that they do not want data centres constructed in their community because, as other people smarter than me have put it, they don’t get much from that bargain.
There’s no real great kind of AI product for them. There’s a lot of great AI products for companies, but not a lot of great AI products for consumers. You and I have just talked about it. In our everyday lives, we use it a bit. We don’t use it for some other things. It’s hard to justify that.
Now does this end up running counter to the story that we opened with here, which was the importance of data to AI-centric business models like the one that Stitch Fix has? You’re increasingly going to find new brands that are predicated on the existence of AI, or you’re going to find existing brands that have really pivoted hard into bringing AI more into the heart of their design, development, planning, market and competitive analysis processes, and so on. So what you have is an outer world and a government that is now increasingly hostile to AI, at least in America. And then on the other side, you have brands who are increasingly using it. Adoption is going up. Use cases are springing up and so on.
Now I don’t think there’s any world where the US starts to really walk things back in a massive way. I think the moratorium on data centre construction seems potentially there. We’ve seen a big sell-off in US tech stocks in the last twenty-four hours, on Thursday, the day before we record this on Friday, which all seems to have been catalysed by people’s assumption that AI is overinvested and that the regulatory environment is going to turn against it.
Now does that affect the ability for fashion brands to get AI tools? No, as things stand. What it might affect is how quickly the frontier of capability moves, because a lot of brands will have invested in agentic systems and other AI initiatives on the assumption that they will continue to get better over time.
A lot of brands have done layoffs, not just in fashion, across industries. A lot of companies have laid people off with the assumption that AI is going to continue to get better. For it to continue to get better, you need people building better products from it, better things, better solutions for particular industries like fashion to use. But you also need companies to continue to invest in training new models, pushing forward the capability frontier and so on. That might get harder to justify as the political pendulum swings in the other direction.
This is one that we can’t guess at, but I think it’s very true that you and I are recording this at the end of a week where, at the start of the week, I don’t think anybody was seriously debating whether data centre construction was going to stop, whether AI build-out was going to stop. Now in light of the political results of the US, there’s potential that it might, and that could have ripple effects for how fashion brands think about their AI strategies.
This one’s going to take a while to play out, but we will continue to watch it and certainly do so as we get into our AI report next springtime as well. But I think those effects might start to show themselves.
