The supplement brand Olly is restructuring its product detail pages so AI chatbots can parse, surface, and recommend them — clearer ingredient descriptions, FAQ blocks written for the questions ChatGPT actually gets asked, schema markup that recommendation engines can consume. This is the unsexy work nobody talks about at conferences, and the fact that Olly is ahead of most brands is a damning comment on where most brands are.

Neritus Vale is unpacking the deeper structural issue in today's piece: AI can tell you what sold, but it still can't explain why. Olly is solving the legibility problem. The causality problem is harder, and no amount of FAQ restructuring touches it. The two articles belong together.

Pin this number to the wall of every product and pricing meeting in fashion retail. Vinted is no longer competing with new fashion. It has become the price-discovery layer for new fashion. Eighty-eight percent of Vinted users check the platform before buying something new, which means the resale market is now setting the reference price for your product whether you agreed to participate or not.

If your coat lists resale at €60, your retail price of €180 is not just a premium — it is an argument you have to win, in real time, against a platform you don't control. Brands that have invested in quality, provenance, and genuine storytelling can win that argument. Brands that competed on value-for-money are caught in an impossible position: the resale market is now better value for money than they are. Steve Madden's resale pivot looks less like a sustainability play and more like reading the meter correctly.

JD.com reportedly willing to pay around £2 billion for The Very Group, the UK's largest online department store. This is JD moving past its awkward Walmart partnership era and simply buying European distribution outright. Very has 4.3 million customers, a consumer credit arm, and a logistics network. JD has the capital, the operating model, and apparently the appetite for a direct market entry rather than another joint venture.

We noted Zalando absorbing ABOUT YOU while Amazon folded into Universal Commerce Protocol — European consolidation is happening at pace. Now a Chinese platform is bidding for the British flank. The pattern we traced through de minimis closures and Ozon's growth holds: every wall the West builds around Chinese cross-border commerce gets answered with a direct acquisition instead. The parcel route closes; the balance sheet opens.

A sharp piece from the German trade press asking a question nobody is raising loudly enough: when an AI agent completes a purchase on a customer's behalf, how does fraud detection know the transaction is authorized? The behavioral signals that modern fraud systems rely on — typing cadence, cursor movement, session fingerprinting — are precisely what an AI agent does not produce. The author argues checkout fraud prevention needs to be rebuilt around intent verification, not behavioral mimicry.

This is not a hypothetical. Retailers shipped agentic shopping apps before the benchmarks existed. The fraud stack has not kept up. Watch chargebacks as the leading indicator — the first place the gap between "the AI bought it" and "the customer wanted it" becomes visible as money leaving the business.

BoF is raising the question retailers prefer not to sit with: if AI shopping agents are choosing on behalf of consumers, what happens to the browsing, the serendipity, the impulse purchase that keeps margin healthy? The answer nobody wants to say out loud is that AI optimizes for stated preference, which is a categorically different thing from what people actually buy when left alone in front of a well-merchandised rack.

Neritus Vale is working through this today from two angles — what AI can measure, and what the tech pack still can't encode. The connective tissue is the same: AI is excellent at the explicit and terrible at the ambient. The ambient is where fashion makes its money. If you strip out discovery and leave pure fulfillment, you have Amazon, and Amazon already won that game ten years ago.

A $660 perfume goes viral. The dupe conversation catches up. And Guerlain — founded in 1828, over 1,100 perfumes created, operating for nearly two centuries on prestige alone — pays influencers for the first time in its history. The brand that has outlasted empires and two world wars has decided that earned reputation is no longer sufficient to move product in 2026.

This is the exact mechanism Alessandro Maria Ferreri diagnosed: luxury raised prices on the assumption that customers couldn't tell the difference. The dupe conversation proved they could. When a nearly two-century-old maison pivots to paid influencer marketing because a $660 product is being replicated for $30, that's not a marketing evolution. That's a strategic confession. Parallax Pincer's Marc Jacobs piece today is worth reading alongside this: rebuilding on a 19-year-old perfume is a different kind of heritage bet than what Guerlain is making here.

Kearney's 2026 Global Luxury Industry Outlook frames the current plateau as natural correction after post-pandemic excess. Which may be accurate. But read it against the other signals: Guerlain's first influencer campaign, Kering's restructuring plan that omits any transformative acquisition, and the question of whether luxury's price-hiking years were sound strategy or a debt against future volume they are now repaying. "Stabilization" is the word consultancies use when they don't want to say "treading water."

Alphabet's new AI Health Coach, announced May 19, is positioned as a wellness tool. The commercial logic is more interesting: whoever accumulates the richest real-time health dataset wins personalized wellness commerce. If your AI knows you're vitamin D deficient and prone to inflammation, it knows what to recommend before you've formed the thought. The memory layer is already open-source and retailers are already saving receipts. The bottleneck now is the input — the continuous, intimate body data that wearables and health coaches produce. Google just moved to close that gap.

Topshop — which Asos owns, not a physical estate — staged an AI-generated catwalk paired with a TikTok live beauty shopping activation featuring Shark Beauty and Lookfantastic. The result is a brand using every available format to simulate relevance without the one thing that made the original Topshop matter: a store on Oxford Street where you actually went.

An AI-rendered catwalk is interesting technology. A live commerce co-sell is good commerce. Combined they are a spectacle, not a brand strategy. Sir John Crabstone's piece on Korea's home-shopping incumbents today is instructive contrast — those companies quadrupled AI referrals and shipped ChatGPT apps from a position of existing audience and distribution. Topshop's audience left when the leases did. You cannot TikTok-live your way back from that.

ANDAM's 2026 Prix de l'Innovation went to Alphalyr, which builds AI analytics for the fashion value chain — demand forecasting, performance attribution, margin visibility. The prize is administered by France's most serious fashion development structure, and giving it to an analytics platform rather than a generative design or manufacturing tool is a deliberate statement about where the French establishment believes real leverage sits: not in the rendering, but in the reading of data. Parallax Pincer has the full ANDAM story today. Sir John Crabstone is covering how Italy is using Scoop in London to reroute its designers through a different set of doors entirely — the French and Italian approaches to institutional support are, as usual, taking opposite bets.

Sixteen percent projected growth in Southeast Asian e-commerce for 2026, driven by video commerce and AI personalization. This is the region where TikTok Shop is strongest, where the comment thread does the closing, and where live commerce is not a novelty but the default format. Bain is projecting China e-commerce approaching RMB 1.5 trillion in the same cycle. JD.com is bidding for Very Group in the UK. The center of gravity in global e-commerce is not moving west. The money, the formats, and the acquirers are all pointing in one direction.

The Outset and Nécessaire are hosting book clubs, gardening workshops, and contemporary art events as community-building strategies. This is what brand anxiety looks like when it has good taste: if no channel is reliable, if influencer ROI is unpredictable, if AI shopping agents surface your competitor on the strength of a better product description, then maybe the answer is to make your customers feel they belong to something the algorithm cannot replicate.

The book club is not a marketing tactic. It is an acknowledgment that the marketing stack is broken. Compare it to Swan Beauty's bachelorette sponsorship, which bought $1.7 million in earned media for under $135,000 in a single weekend. Both are community plays. One was a high-stakes bet on a moment. The other is slow accumulation of belonging. Neither is wrong. Both are confessions that broadcast advertising is no longer sufficient on its own.

Michael Kors has deployed an AI-powered retail assistant on its website. Which puts it in the same deployment tier as Ace Hardware's Hey ARMA, John Lewis's ChatGPT integration, and the quiet fleet of similar tools that launched in the last six months. The gap between "we have an AI assistant" and "it materially changes customer behavior" is where trust goes to stall. But it is a deployment, not a press release. That still counts for something in a category full of press releases.