How does ASOS use AI? Outfit recommendations, the AI Stylist and ChatGPT

KEY TAKEAWAYS Summary by the editors

1. In November 2025 ASOS said it was testing 'Styled for You', an AI feature that suggests outfits built around individual products and is trained on more than 100,000 professionally curated studio outfits.
2. On 20 May 2026 ASOS launched ASOS Stylist, a shopping app inside ChatGPT for UK and US customers, built with Bambuser, with purchases completed on ASOS.com.
3. ASOS says ASOS Stylist builds on an existing AI Stylist in its mobile app, which gave the company insight into how customers use conversational AI to shop.
4. In ASOS research with Attest among 1,952 UK consumers (July to August 2026), 33% said they had abandoned a purchase they could not picture wearing and 53% said complete outfit inspiration would help.
5. ASOS's AI styling work sits inside a turnaround: sales fell 12% in the year to 31 August 2025, while pre-tax losses narrowed to £281.6 million.

ASOS uses AI mainly to help customers decide what to wear with what, rather than only which single product to buy. Since late 2025 it has tested AI-generated outfit suggestions ('Styled for You'), run an AI Stylist in its app and, in May 2026, launched an ASOS Stylist app inside ChatGPT for UK and US shoppers. The common thread is outfit-led discovery built on ASOS's own curated styling data.

What AI features has ASOS launched?

ASOS's published AI initiatives from 2025 to 2026 fall into three layers. The first is outfit generation inside its own channels. When it reported results for the year to 31 August 2025, ASOS said it was testing a feature called 'Styled for You' that recommends outfits built around individual products. The company described the model as trained on its database of more than 100,000 expertly curated studio outfits, according to PA Media's report of the results.

The second layer is a conversational AI Stylist in the ASOS app. ASOS listed 'AI Stylist' in an August 2026 press release as a tool giving customers personalised outfit inspiration and styling advice based on their tastes, alongside Virtual Try-On, Buy The Look and a Follow Brands function.

The third layer moves outside ASOS's own properties. On 20 May 2026 ASOS launched ASOS Stylist, an app that runs inside ChatGPT, for customers in the UK and US. Shoppers can browse and complete outfits by category, occasion or trend, view products as video, and then click through to buy on ASOS.com, as reported by Retail Gazette.

Why is ASOS focusing on outfits rather than single products?

ASOS's own consumer research explains the bet. In a survey of 1,952 UK consumers conducted with Attest between July and August 2026, ASOS found that 68% said knowing what to buy is easier than knowing how to wear it, 45% had bought something without knowing what to wear it with, and 33% had abandoned a purchase they could not picture wearing. More than half (53%) said complete outfit inspiration would help.

For a pure-play online retailer with a very large assortment, these are conversion and confidence problems as much as inspiration problems. A shopper who cannot visualise an item in a wardrobe is less likely to buy it, and a shopper who buys without that confidence may be more likely to send it back. ASOS has not published figures linking its AI Stylist to return rates, so the returns effect should be treated as a hypothesis rather than a result.

The commercial context matters too. ASOS reported a 12% fall in sales for the year to 31 August 2025, with pre-tax losses narrowing to £281.6 million from £379.3 million. Chief executive José Antonio Ramos Calamonte said the priority for 2025-26 was 'to deepen our relationships with customers', and the company paired AI styling with a new loyalty programme, ASOS.World.

Read also

How Zalando uses AI: the assistant, personalisation and size advice

How does the ASOS Stylist app in ChatGPT work?

ASOS built the ChatGPT app with the video commerce company Bambuser. According to Retail Gazette, Bambuser's 'Intelligence Layer' converts ASOS's product catalogue and video library into structured, machine-readable data that large language models can return in real time as shoppable videos. A shopper can ask for a specific edit, for example 'pastel floral A-line dresses for spring', and receive a curated selection that reflects the conversation so far.

• Discovery happens in the AI assistant, but the transaction stays on ASOS.com, so ASOS keeps the customer relationship and checkout.
• Video is the differentiator: ASOS argues most AI shopping tools are limited to text and static images.
• The catalogue must be machine-readable: the partner's role is to turn product and video assets into structured data a model can retrieve.
• Learning came from the app first: ASOS says its existing in-app AI Stylist showed how customers use conversational AI to shop.

ASOS's head of product, Melissa Lim, said shoppers increasingly use AI for inspiration but find the experience 'fragmented and hard to visualise', and described the app as bringing 'fashion discovery, styling advice, and shoppable products together'. ASOS also said more shoppers are starting product research on agentic platforms, which is why it chose to appear inside ChatGPT rather than wait for customers to arrive at its own site.

What data does ASOS's AI styling depend on?

The most distinctive asset is not a model but a dataset: more than 100,000 studio outfits assembled by ASOS stylists. Generic recommendation engines learn from co-purchase and click data, which tends to surface items that sell together rather than items that look good together. A curated outfit library gives the model examples of styling intent. Retailers without such a library would need to build one, or accept recommendations that reflect purchase patterns rather than taste.

The second dependency is product content. Outfit generation, conversational search and video retrieval all require consistent attributes (category, colour, occasion, fit) and rich media. The ChatGPT app shows that external AI assistants can only return what they can parse, which makes catalogue structure a precondition for visibility.

What are the limits and open questions?

ASOS has published features and intentions, not performance data. It has not disclosed conversion uplift, basket size or return-rate effects for Styled for You, the AI Stylist or the ChatGPT app. The ChatGPT app is also limited to two markets, and its success depends on how much shopping behaviour actually shifts to AI assistants.

There are also familiar risks. Personalised styling can narrow what customers see, and outfit suggestions that ignore stock depth or size availability frustrate shoppers. Placing discovery inside a third-party assistant creates a dependency on that platform's rules and ranking.

Read also

Virtual try-on in fashion: how does it work and does it reduce returns?

What can other fashion retailers learn from ASOS?

1. Start from a specific customer problem. ASOS framed AI around 'how do I wear this?', backed by its own survey data.
2. Treat curated styling data as a strategic asset. Human-built outfits give a model taste that click data alone does not provide.
3. Make the catalogue machine-readable before chasing AI channels. Structured attributes and tagged video are what assistants retrieve.
4. Keep checkout and the customer relationship on owned channels where possible.
5. Measure returns and conversion separately before claiming that styling AI reduces returns.

Frequently asked questions

What is ASOS Styled for You?

Styled for You is an AI feature ASOS said it was testing in November 2025. It suggests outfits built around individual products and, according to ASOS, is trained on more than 100,000 professionally curated studio outfits.

Does ASOS have an app in ChatGPT?

Yes. On 20 May 2026 ASOS launched ASOS Stylist inside ChatGPT for customers in the UK and US. Shoppers can browse outfits by category, occasion or trend, watch product videos and complete purchases on ASOS.com.

Who built ASOS's ChatGPT stylist app?

ASOS built it with Bambuser, a video commerce platform. Bambuser's Intelligence Layer turns ASOS's product catalogue and video library into structured data that large language models can return as shoppable videos.

Does ASOS's AI stylist reduce returns?

ASOS has not published return-rate figures for its AI styling tools. Its research shows that a third of UK shoppers have abandoned purchases they could not picture wearing, which suggests outfit guidance could help, but no result has been disclosed.

GuideThe complete guide to AI in fashion e-commerce, marketing and retailRead the complete guide