AI in fashion: the September 2026 review

KEY TAKEAWAYS Summary by the editors

1. In September 2026 Meta launched its Muse shopping-capable AI agent on 8 September and Amazon began blocking it from Amazon.com on 20 September, while Shopify opened checkout to browser-based AI agents on 28 September.
2. Macy's reported on 18 September that it is moving an AI forecast overlay for inventory replenishment from pilot into broader use.
3. At Zalando's Partner Day, About You said generative AI cut its product photography spend by about 90 percent, saving roughly 6 million euros, and that chatbots send less than 0.5 percent of traffic to online fashion stores.
4. A YouGov survey for ACI Worldwide found that only 7 percent of UK and US fashion shoppers would let AI buy without their approval, while Global Payments research found 69 percent would let an agent spend up to $100 on clothing and footwear.
5. No new EU AI Act milestone fell in September 2026, but Article 50 transparency obligations have applied since August 2026, with a grace period for marking AI-generated content ending on 2 December 2026 for systems already on the market.

September 2026 was the month in which AI shopping agents stopped being a forecast and became a commercial dispute. Meta launched its Muse agent, Amazon blocked it within two weeks, and Shopify opened checkout to browser-based agents, while fashion retailers such as Macy's, H&M and Zalando described AI mainly as a tool for inventory, content and efficiency. Two consumer studies published in the same month showed why the agent question matters for fashion: shoppers are willing to let AI help, but most still want the final say.

September 2026 at a glance

Who is allowed to shop for the customer?

8 and 20 September, Meta and Amazon. Meta introduced Muse on 8 September as a general-purpose agent that runs on its own cloud computer and needs user approval before it buys anything, according to The Rundown AI. On 20 September, twelve days after launch, Amazon began blocking Muse, telling users that its access violated the store's terms. An Amazon spokesperson said the company had asked Meta to remove the bot and Meta had declined. The report links the move to Amazon's advertising business: an agent that chooses products and checks out can bypass sponsored listings.

Why it matters: for fashion brands that sell through marketplaces, the agent that reaches the customer first may now decide which retailer, listing and price the customer sees. Whether a platform admits or blocks a given agent is becoming a distribution question, not a technical detail.

28 September, Shopify. Shopify took the opposite position. TechCrunch reported that it extended its WebMCP support to checkout, adding tools that let browser-based agents read a checkout, change details such as delivery options and submit the order once the buyer approves. The feature rolls out automatically to eligible merchants, and Muse already has a direct partnership with Shopify. For direct-to-consumer fashion brands on Shopify, this means agent purchases can arrive without any project on their side, so product data, size information and return terms need to be clear enough for a machine to read.

15 to 17 September, Salesforce. At Dreamforce in San Francisco, Salesforce introduced Carter, a shopper agent that answers product questions, compares merchandise and supports checkout inside a conversation, Retail Insider reported. It also announced integrations with ChatGPT, Google Gemini and AI Mode in Google Search, and a "Shop the Store" function that unifies inventory across its B2C Commerce and point-of-sale software. Many fashion retailers run on this kind of commerce stack, so the practical question is less about the agent itself than about whether store stock and online stock are accurate enough to be promised in a chat.

Read also

AI agents explained for fashion executives

How are fashion retailers using AI in operations?

18 September, Macy's. Retail Dive reported that Macy's is moving an AI forecast overlay for replenishment from pilot into broader use, as part of its "Bold New Chapter" plan. The aim, in the company's words, is "improved in-stock levels" and inventory efficiency. The retailer did not name a vendor. This is one of the clearest examples this month of AI being applied where department stores feel the cost most: stock in the wrong place.

24 September, H&M Group. In its nine-month report, H&M said third-quarter net sales rose to SEK 57,189 million, up 1 percent in local currencies, with an operating margin of 10.6 percent. Chief executive Daniel Ervér described more decisions being made "closer to the customer" with "greater support from AI" to improve relevance and efficiency, alongside more in-season purchasing. The report gives no figures for the AI contribution, which is typical: large fashion groups are naming AI as part of how they plan and buy, without yet disclosing its measurable effect.

Late September, Zalando Partner Day. WWD, republished by Yahoo Finance on 26 September, reported on Zalando's annual Partner Day in Berlin, held the previous week with about 750 attendees from brands. Zalando's vice president for content solutions, Matthias Haase, said around 70 percent of the visuals shown on stage had been created with AI. About You managing director Tarek Mueller said generative AI had cut About You's product photography spend by about 90 percent, saving roughly 6 million euros of a previous 7 million euro budget. Haase said that when Zalando closed its dedicated product studio, 90 positions were cut and about 60 employees moved to other roles internally. Mueller also said large language models such as ChatGPT and Claude currently send less than 0.5 percent of traffic to online fashion stores.

Why it matters: the savings on imagery are large and concrete, and so are the job effects. The traffic figure is a useful counterweight to agent enthusiasm: for now, AI assistants are a small acquisition channel for fashion, even if they are growing.

22 September, Parloa and Zalando. The Next Web reported that Parloa's co-founder Malte Kosub announced on LinkedIn that Zalando will use Parloa's AI for customer conversations across Europe. Neither company published an official announcement, Zalando did not comment, and no scope, markets or timetable were given. We list it because it was widely reported, but it should be treated as unconfirmed until Zalando says more. The article also notes that it came seven weeks after the EU AI Act's transparency obligations under Article 50 began to apply.

Product development, content and discovery

17 September, Lectra. Lectra launched Apogy, a cloud platform that combines AI agents, product data and design workflows for the phase between design and production. According to the company's press release, the agents automate routine tasks, support decisions, generate design concepts from prompts or images, and work with 3D simulation and digital prototyping. Lectra says the platform took three years to develop, names O'Neills and Oniverse as early adopters and gives no general availability date. For product development teams, the relevant test will be whether agents work on clean, shared product data; without it, generated concepts add work rather than remove it.

11 September, Amazon. Amazon announced Shop the Scene for Prime Video, which uses AI to identify clothing worn by actors so viewers can buy the same or a similar item from the Amazon Shopping app, Fortune reported. It is available on more than 600 titles in the US. Fortune notes the match "may not always be exact", which is the core limitation of visual product matching in fashion: similar is not identical, and brands may find their look sold through a substitute.

14 September, Daydream. The fashion discovery app Daydream added two features based on Apple's iOS 27 tools, TechCrunch reported: it can analyse outfit photos saved on an iPhone, identify individual pieces and surface matching products, and it accepts product requests through Siri. The company cites about 3 million products from more than 325 retailers and 10,000 brands, and 1.5 million shoppers. For brands, this is another discovery layer that depends on product feeds with accurate attributes and images.

What do shoppers say about letting AI buy fashion?

Two studies published in September look contradictory at first but measure different things.

• 16 September, ACI Worldwide. A YouGov survey of 3,328 adults in the UK and US, fielded in June 2026, found that only 7 percent of fashion shoppers would let AI make purchases without their approval and 53 percent were uncomfortable with autonomous AI purchasing. The most wanted features were price drop alerts and cross-retailer price comparisons (35 percent each); 18 percent were interested in personalised recommendations.
• 23 September, Global Payments. Research with The Lantern among 8,027 consumers in seven markets in May 2026 found that 69 percent would let an AI agent spend up to $100 on clothing and footwear, and 64 percent would accept the same limit for luxury clothing and footwear. Consumers expect AI agents to make 15 percent of their purchases within five years, up from 9 percent in the previous wave. Concerns remain: 50 percent cite payment security, 46 percent privacy and 42 percent the risk of wrong purchases, and 33 percent want to approve each transaction.

Read together, the message is consistent: shoppers accept AI that finds, compares and proposes, and many accept small spending limits, but very few want to hand over the decision without a check. That matches the design of Muse and Shopify's checkout tools, both of which require buyer approval.

Was there news on the EU AI Act?

We found no new AI Act milestone for fashion in September. The relevant changes came just before: the Digital Omnibus on AI entered into force on 27 July 2026, moving obligations for high-risk systems listed in Annex III to 2 December 2027, while the Article 50 transparency obligations applied from August 2026, according to Usercentrics. For fashion, Article 50 is the part that matters most, because it covers chatbots that talk to customers and AI-generated content. Systems already on the market before 2 August 2026 have until 2 December 2026 to meet the marking requirement in Article 50(2). Retailers rolling out conversational agents, as reported for Zalando this month, need to make clear to customers that they are talking to an AI.

What does September mean for fashion professionals?

• E-commerce and digital teams: decide which AI agents you want to serve and test how your checkout behaves when an agent, not a person, completes it.
• Merchandising and planning: the most concrete retailer deployments this month were in replenishment and in-season buying, where AI adds value only if stock and sales data are accurate.
• Product data owners: agents, visual search and AI discovery apps all read the same attributes, sizes and images, so product data quality now affects visibility in channels you do not control.
• Content and brand teams: AI imagery is producing measurable cost savings at scale, with real effects on studio jobs, and may fall under EU transparency rules for AI-generated content.
• Leadership: treat claims about AI traffic with care; one large platform reported that chatbots still send less than 0.5 percent of traffic to online fashion stores.

Read also

Is fashion AI biased? Sizes, skin tones and fairness explained

What to watch in October 2026

• 29 October: adidas publishes its nine-month 2026 results, according to its financial calendar.
• 3 November (just after month end): Zalando publishes its third-quarter 2026 results, its first report since Partner Day and the reported Parloa announcement.
• 2 December (preparation in October): the grace period for AI Act Article 50(2) marking ends for generative AI systems already on the market before 2 August 2026.

We will also follow whether Amazon and Meta resolve the Muse dispute, whether Zalando confirms the Parloa deployment, and how quickly agent checkout on Shopify shows up in fashion brands' order data. None of these has a scheduled date, so we list them as open questions rather than events.

Frequently asked questions

What were the biggest AI in fashion news stories in September 2026?

The main stories were Meta's launch of the Muse AI agent and Amazon's decision to block it, Shopify opening checkout to browser-based agents, Macy's scaling AI replenishment and Zalando's Partner Day discussion of AI imagery. Lectra also launched Apogy, an AI-agent platform for product development.

Will shoppers let AI agents buy clothes for them?

Partly. A Global Payments study found 69 percent would let an agent spend up to $100 on clothing and footwear, but an ACI Worldwide survey found only 7 percent of fashion shoppers would let AI buy without approval. Most consumers want AI to search and compare, and to keep the final decision.

Why did Amazon block Meta's Muse agent?

Amazon said the agent's access violated its store terms and that it had asked Meta to remove it. Reporting links the move to Amazon's advertising business, because an agent that selects products and checks out can bypass sponsored listings.

Does the EU AI Act apply to fashion chatbots and AI images?

Yes, through the Article 50 transparency rules, which have applied since August 2026. Customers must be told when they interact with an AI system, and AI-generated content must be marked; systems already on the market before 2 August 2026 have until 2 December 2026 for the marking requirement.

GuideThe complete guide to AI strategy for fashion companiesRead the complete guide