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 edition – coming out in our Thursday slot this week, due to several vacation schedules crossing over – covers the court ruling that clears the way for agentic shopping; why AI investments are failing to pay off without solid data foundations; Pinterest’s new London office and its pitch to be fashion’s digital shop window; what GLP-1s mean for sizing, fitting rooms and in-store service; and New Balance’s new factory garden in Cumbria – and what it says about how the industry treats workers close to home.
This show is available in video format, embedded below, as well as audio, on Apple Podcasts and Spotify.
For deeper interviews with brand and technology leaders, our prestige interview show will be back next Thursday everywhere you find podcasts – and will be coming to YouTube soon.
Subscribers to our mailing list will receive notifications for both shows.
Grace Robinson: So 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 technology 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 less than twenty-five minutes to give you our analysis on the stories that we think really matter.
Ben Hanson: Hey, Grace. Good to be here again, and I’m back in the UK this time, so anyone who’s watching this on video will see my typical background. Weirdly, I’m going away again on Saturday for a few days, but I will be back in time for the next of these. So, yeah, good to be here. Good to be back on domestic soil. Let’s go.
Agentic commerce — a landmark ruling for a use case almost nobody uses yet
Grace Robinson: The first story this week concerns what might be a bit of a milestone legal case for agentic commerce. Amazon previously filed to prevent Perplexity from allowing its Comet agentic browser to search and buy things from Amazon on users’ behalf, making the claim that Perplexity had committed a federal hacking crime under the CFAA — the Computer Fraud and Abuse Act — accusing it of covertly accessing private Amazon customer accounts. However, it’s also worth noting that we can assume the claim stems from Amazon’s effort to prevent Comet from overtaking its own agentic AI, Rufus.
The new ruling says that it was the user, not the agent, accessing Amazon’s website, because the agent was running in the user’s browser under their command. So Perplexity’s servers were simply receiving what the user sends. This story does seem to show that we are at a significant inflection point with AI shopping. So what do you think of this, and what’s really going on here?
Ben Hanson: There are a couple of things going on. The biggest one is that I’ve never met anybody who actually uses an agentic browser for shopping. I’m sure they exist, otherwise these companies wouldn’t keep making them. But Comet, and OpenAI’s version — I think it’s called Atlas — did that get retired? Anyway, there was a whole push for a while towards the idea that AI would live where your tabs live, and that you would interact with an LLM as the primary layer in your browser. I think since then the prevailing sentiment in the state of the art has moved on a little, to where it’s more likely that people interact with Claude, or OpenAI’s Codex, or an agent harness like Hermes, and that that then drives either the browser they already have open, or a headless version that happens in the background.
So this is a weird case, in that it is a momentous decision in principle, but it applies to a use case that I think is really rare and really sparse, and that I don’t think is going to become the prevailing use case.
But the interesting part of it is that it was always a weird avenue to go after Perplexity on — hacking, essentially; computer fraud and abuse. What was really at issue, I think, was the question of automation: automated browsing, where the browser is taking action under its own steam, whether the user has requested it or not. To retailers, to websites and so on, that looks like bot traffic, and it looks like automated traffic. We see this at The Interline. There is a mess of automated traffic that breaks down into big, wide-format scraping and crawling, and then this kind of user-directed agentic side of things. The interesting ruling here is that they are basically saying this does not constitute that kind of bot traffic, because what is actually happening is that an AI agent is going out and taking action on behalf of a shopper who has directed it to do something, and that that is functionally and legally equivalent to the shopper taking the same actions.
Now, you correctly called out that Amazon have their own AI built into the Amazon website and the Amazon app, which is Rufus. Amazon also have a deep and abiding interest in not giving away a lot of information; they would prefer to capture things in the other direction. And I think the concern is that as more of this agentic activity takes place, more of the route between the consumer and the retailer gets disintermediated. You are not actively sitting there sifting through Amazon’s catalogue — it’s being done for you. And Amazon would like you to sift through their catalogue, because that’s a chance to upsell you and everything else.
This makes it part of the bigger conversation about whether the rise of ChatGPT, and the billion users or what have you, is going to reduce the amount of time that people spend interacting with retailers. It takes us back to the conversation you and I had, either last week or the week before, about Daydream and so on being offered up as a service to retailers instead of going through dedicated apps. It seems like there is something shifting here, and it’s shifting in a place that I think is maybe less consequential than the legal part of it would suggest, just because behaviourally, I don’t use agentic browsers. I never really have; it never really took root, and I’m going to guess you don’t either.
Grace Robinson: No, I don’t.
Ben Hanson: Well, there we go. So let’s see what happens on that one.
AI investment — why the returns depend on your data foundations
Grace Robinson: The second story is about AI again, but I still wanted to talk about it because we’ve just published The AI Report 2026, and really interestingly, some of the findings in our report have a lot in common with a piece we saw from Marc Bain at The Business of Fashion. Marc’s piece was called “The Reason Your AI Investments Aren’t Paying Off”, and it explains how fashion brands could get a better return on investment on AI if they first got a better grip on their internal data. The article heavily focuses on Zalando, who are claiming to get a lot of value out of their AI strategy, and who say, I quote, that they “have the best data on fashion and lifestyle in Europe”.
In parallel, the BoF piece also pulls from a new Bain & Company survey that looks at the barriers companies are facing in getting what they wanted, or expected, out of AI. Interestingly, that survey also shows that most respondents cited challenges with data access and integration as the reason for their lack of AI progress, with compliance concerns like data privacy, and competing business needs, following shortly after.
So do you think this lines up with what we found in our own AI survey? Does it contradict anything? And are brands really struggling to get value out of AI?
Ben Hanson: Yeah, I thought this was an interesting piece. Zalando is everybody’s go-to for AI success, for a number of reasons. I’ve interviewed somebody, Glazione Rocha from Zalando, who’s a creative director in their Lounge division, and I did that because they are the furthest ahead, I think, and the most open about their use of generative product images and videography. But as this article backs up, they’re also pretty far ahead in having built out the requisite data foundations that are now delivering value when they want to layer AI on top of them. So that’s all good and correct.
And I think the Bain & Company survey that’s cited in there does chime with what we found in The AI Report 2026. The trend that we described is that people at large use a lot of AI, but they don’t put a lot of trust in it when it comes to making business-critical decisions. Something like less than a quarter of professionals trust AI enough to make a vital choice based solely on its output; they would always go and take in something else.
Part of the reason for that lack of trust is the same finding as the Bain & Company one, which is that the majority of AI initiatives are either fully sandboxed — they live completely independently from the rest of the brand or retailer’s technology estate — or they’re largely standalone, with a few connectors and things like that. So the situation we have is that companies are spending a lot of money on AI, and people are using it, either through heavy incentives and mandates or of their own volition, but people don’t trust it because they don’t believe they’re going to get accurate answers from it.
They would get accurate answers from it, or at least they’re far more likely to, if they connected it to their wider technology estate. If you connected your AI chatbot of choice, through API, CLI or MCP, to your PLM, your ERP, your CRM, your supplier management, any of these systems, it would then have the context and grounding to give you answers that you could build trust in, and over time that trust behaviour would compound. But it’s a vicious circle: if you don’t trust AI, you’re not going to connect it to your sensitive systems; and if you don’t connect it to your sensitive systems, it’s not going to give you accurate answers, so you’re not going to trust it.
So that’s the key finding of our AI report. There’s a whole lot of other material in there, but the specific article you’re referencing, and the Zalando part of it, is evidence that there’s no real shortcut to that. The reason they seem to have been able to get so much value from AI is because they put all those foundations in place years ago. You cannot just go and do it now and then hope to shortcut that trust.
I think the key thing for listeners and watchers to take away from this is that you probably do want to watch what Zalando are doing with AI. It doesn’t have to be on the generative image side of things; at the pure data level, connecting AI to your important systems is going to deliver a greater return than treating it in isolation. But you can’t magically connect all your systems, and you can’t magically build those kinds of data foundations. It’s still a big governance and technical ask to get you on the way there. I’d suggest doing the behavioural work alongside it, but there’s just a lot of heavy lifting and architecting that needs to be done, I think, for people to get value from AI. And that’s the conclusion this article reached, and I think it’s the same one that our survey did as well.
Pinterest — the digital shop window opens in Oxford Circus
Grace Robinson: Next, I want to talk about Pinterest, which has just opened a new London office in Oxford Circus. What stood out to me about this news is that it feels like Pinterest is really doubling down on its mission to turn inspiration into actual purchases. We have already seen really successful retail partnerships from Pinterest: for example, they partnered with Primark, and Pinterest reportedly contributed to a 115% uplift in store purchases for Primark. They also had really similar results with H&M.
With regards to the new office, Pinterest’s VP of International, Greg Owens, said that this marks retail entering a new era where AI, visual discovery and physical stores blend seamlessly, and he also described Pinterest as a “digital shop window”, which was really interesting. Now, Pinterest isn’t the only platform moving in this direction. More and more social platforms are investing in features that actually help customers move seamlessly from discovery to purchase.
So my question to you is: do you think every social platform is becoming a shopping platform? And what do you think that shift means for brands trying to connect with customers and drive sales through digital content?
Ben Hanson: So the first thing I’ll say is that you’re more of a social media person than I am, and you also live in London, so this story is in your backyard in two respects. I’m going to give you my take on the latter part of it, which is that I’ve always been open about the fact that I don’t use social media personally. I do use it for work — LinkedIn, Instagram and so on.
It’s very clear that every social channel is becoming a transactional channel, first and foremost. It’s already the case that the entire Meta ecosystem — everything across Instagram, WhatsApp and so on — exists in service of putting ads in front of people. That’s been the stated and open business model for a decade plus. The transaction side of it emerged organically from that, I would say. So what you have is not TikTok and Pinterest and so on suddenly going, “oh, we should really try and become part of the path to purchase.” What it is is them recognising that they have become such a force in the path to purchase that they can now start to exert some power over the brands and retailers, and say: look at the contribution that we are making to your sales. Look at the contribution we’re making to your GMV. Look at the incidence of people coming from Pinterest and other places directly to you, and look at the value of those visitors versus the value of visitors coming from traditional search or from direct browsing.
Now, I don’t have those stats to hand, and they’re not part of this release, but you have to assume, based on the bets that Pinterest and other companies are making, that the value of the customer traffic they’re driving is higher than it is through other avenues. What that means is that “digital shop window” is both a statement of power and intent from Pinterest and others, because it recognises the position that they have in the product journey. It’s also a bit of a damning testimony for brands and retailers, as a recognition that they don’t control the digital channel in the way that they would like to. I am willing to bet that traffic to Pinterest is higher than traffic to the typical non-famous brand’s direct-to-consumer e-commerce portal.
So if you’re asking me, does social exert a big force over the way that people shop online? Yeah, the answer is very clearly yes. And I think Pinterest’s journey is especially interesting, because you’ve gone from a platform that was just intended to help people pin images — that was the initial thrust of it. I signed up for Pinterest in the early days; it was a pinboard, and that’s where it comes from. Over time, it’s leaned much more heavily into curation and influences, and into people assembling collections of assortments and ideas.
And then, logically, that’s become part of the handoff to transaction. This is a very interesting part of e-commerce, and a very interesting part of retail tech in general, and it’s further evidence of the force we talked about a couple of weeks ago around Substack and the native sponsorship programmes there. Traditional spending on web search and so on is probably on the way out, and spending on these kinds of channels is certainly on the way up, because they just have so much control over the path to purchase.
GLP-1s — the biggest change in size data fashion has faced in years
Grace Robinson: So the next story is about GLP-1s, and the fact that they’re causing more people to shop in person. This was covered a lot in the headlines this week, but I was looking at one piece from Reuters, and they were explaining that about 11% of Americans are on GLP-1 weight-loss drugs, and that with their new bodies, many are inclined to try on clothes in person to figure out what styles fit their new size. The piece goes on to say that with this change in shopping habits, many fashion retailers are going to need to re-evaluate things like the sizes they carry, their dressing room set-ups, and the styling services they offer.
So, from a fashion tech perspective, do you think something like virtual try-on could bridge this gap? Or is there any other kind of tech that could help brands better build their size ranges from the start?
Ben Hanson: Yeah. Any conversation about virtual try-on has this baked into it automatically now, because it has become such a prominent force. But notably, it’s not the first time that fashion has had to deal with changing bodies, and with bodies that don’t necessarily conform to either traditional expectations or traditional rules. And I mean rules from the point of view of sizing and grading rules, rather than social norms.
I remember speaking to somebody a couple of years ago — this was pre-Ozempic and Mounjaro — who was talking about having to change the way they approached sizing and points of measure because of the rise of strength training in women. So women’s bodies suddenly take on a slightly different and more muscular aspect than they have done traditionally, because strength training became much more universally appealing across genders. The other thing is cosmetic surgery, so that’s been in there as well.
Now, you have size grading rules that traditionally would have covered most of the allowance for chest size relative to waist size relative to hips, and all these kinds of things. When you start to deal with the rise of cosmetic surgery and implants, you start to separate the traditional sizing rules from the actual reality. But none of that was anywhere near as widespread as the GLP-1 uptake is. If it’s true that 11% of Americans are actually on Ozempic, Mounjaro and so on, I think it represents the biggest change in size data that fashion has had to deal with in a very long time.
It also changes, I think, the way that fashion communicates with its shoppers and buyers. Previously, outside of purposely size-inclusive brands — companies that would have been described as catering to “big and tall”, that kind of thing — there has always been some sensitivity around the way that brands talk to people with body types that fit outside the typical small, medium and large brackets. Now the conversation is much more open. You mentioned styling services, you mentioned fitting rooms, and you mentioned people coming into stores because they have experienced significant weight loss, and they want to talk about it, they want to spotlight it, and they want to be able to take advantage of fashion that they haven’t previously been able to.
Now, I think there’s an opportunity for fashion there, because right now a lot of brands design to outdated size surveys, size sets, size ratios and assumptions. It’s long been the case, at least for the last five or ten years, that fashion has wanted to build better demographic data about the people it actually sells to, and about the size of their bodies. There’s never been a great way to do it, and there’s never been a great incentive to say: you, customer, please scan your body, because I want to be able to better sell to you.
I think this might represent that opportunity, and it might represent that incentive — to have people who are now advocates for a different way of living, people who are now proud of their body, people who’ve been able to achieve a level of weight loss that they maybe haven’t in the past, and who are much more willing to have these conversations. And you can have those conversations in store, around fitting. You can have those conversations online, remotely, in store, in a number of different places, around capturing more granular, more current and more contemporaneous data about the size of your actual target demographic.
New Balance — what you’re buying when you buy domestic
Grace Robinson: So we’re ending this week’s episode on a bit of a positive note, which is nice. I came across this story in FashionUnited, and essentially New Balance has opened its first ever garden at its factory in Flimby, in Cumbria. It’s designed to offer a restorative outdoor space for the factory workers who are responsible for the craftsmanship of its Made in UK products. The garden has been conceived as a restorative environment where employees can step away from the production floor and spend time immersed in nature.
Now, this isn’t a typical story that you hear about fashion factories, but do you think we are going to start to see more of this, especially if more brands start to try to showcase their commitment to craft? Or is it just evidence of how differently companies treat their workers closer to home versus further away?
Ben Hanson: It is a nice story, so I’m going to try not to be my usual cynic. I will always commend New Balance for their Made in UK and Made in US strategy, because I went to a talk once, here in Manchester, from some of the folks behind the New Balance Made in UK operation. They are very candid about the fact that you can buy footwear made overseas, or you can buy the same footwear made domestically. It’s not always a 100% equivalent in colourways and so on, but the platforms, the uppers and the materials are the same, and the construction techniques are the same. They’re always very candid about the fact that you can buy a version of that made overseas, or you can buy a version of it made domestically, either in the UK or the US, and it’s more expensive. I don’t have an exact barometer, but it’s something like £100 on top of the price of the version that’s made offshore.
Now, I like the candour of that. I always have done, and I always like the idea of them saying: this is what it costs to do this thing close to home. And if you make that investment, you are making it because you believe in domestic production — either because you work in the supply chain and you think it’s interesting, because you are somebody with some measure of national pride, or, more likely, because you want to support a transition away from a more exploitative model of production that happens globally.
It’s been harder, though, to then be able to measure what your £100 did. If I was buying the Made in UK version versus the made in, I don’t know, Vietnam version — don’t quote me on that, maybe they don’t make them in Vietnam — the warm and fuzzy feeling that I would get is the return. Now, when you see stories like this, you start to go: okay, this is what that buys. If I spend on made in country, I am spending on measurably improving the quality of life of the people who made the products that I now wear. And that’s a good loop. That’s the kind of loop that leads to more domestic manufacturing, more domestic purchasing, and so on.
To put my cynic’s hat back on for a minute, the reason this is a story is because there are no restorative gardens in the factories in Bangladesh, as it were. If you go and work in a legal office, or you go and work somewhere else — if you go and work in a profession where this kind of worker safety and worker well-being is standard — this isn’t the press release, and this isn’t the story, because it’s assumed. The quality of life of the people who work in those kinds of knowledge sectors is considered to be good enough that this would not make the news. The quality of life and the well-being of your typical garment, footwear and accessories supply chain worker is bad enough that deviations from the norm of this type are newsworthy.
Now, I don’t want to play down the effort involved here. I think this is a very nice story, I think it’s a good effort from New Balance, and I think their strategy in general is great. I think it’s one of those things where we’re only talking about it because the plight of the typical garment worker is sufficiently poor that you need some good news in this area. So, yeah, sorry — I’m trying not to end us on a downer.
So, yeah: good news, and a nice restorative garden. Let’s think on that.
