How Marks & Spencer uses AI for style advice and product descriptions
M&S uses AI to personalise online fashion, power a body shape and style quiz and write most product descriptions, and in 2026 gave Copilot to 11,000 colleagues. What has been reported.
KEY TAKEAWAYS Summary by the editors
- Marks & Spencer uses AI to personalise its online fashion experience based on shoppers' body shape and style preferences and to prioritise the products most relevant to each customer.
- A body shape and style quiz combines suggestions from M&S's in-house stylists with shopper input; by September 2024 it had been used by 450,000 shoppers and could draw outfits from 40 million options, according to SupplyChainBrain citing the Guardian.
- The share of M&S product descriptions written with AI rose to 80% from almost nothing a year earlier, according to the same September 2024 report.
- In March 2026 M&S bought 11,000 Microsoft 365 Copilot licences for colleagues including every store manager and support centre colleague, alongside a training programme, Retail Gazette reported.
- M&S's stated aim is for online to account for 50% of its fashion sales by 2028, up from about one third in 2024, which explains its focus on AI in the online journey.
Marks & Spencer uses AI in two visible places in its online fashion business: personalised style advice, built around a body shape and style quiz, and the writing of product descriptions, most of which were AI-generated by 2024. Both serve a stated goal of raising online's share of M&S fashion sales to 50% by 2028. More recently the retailer has extended AI to colleagues, rolling out Microsoft 365 Copilot to 11,000 staff in 2026.
How does M&S use AI to give style advice?
In September 2024, SupplyChainBrain, citing a Guardian report, described how M&S uses AI to help shoppers choose outfits. The retailer personalises the online experience according to a shopper's body shape and style preferences, suggests items to buy and prioritises products most relevant to each customer. The report gave a simple example: men are less likely to be shown bra promotions.
The centre of the experience is a quiz that asks for size, body shape and style preferences and returns outfit ideas. According to the report, it combines suggestions from M&S's in-house stylists with the shopper's input to indicate which garments may go together. Stephen Langford, director of online and omnichannel, was cited as saying that 450,000 shoppers had used the quiz, which could draw outfits from 40 million options.
The design is notable for what it does not do. It does not replace stylists with a model; it encodes stylist knowledge about combinations and uses AI to match that knowledge to an individual's stated preferences at scale. That is a lower-risk pattern than open-ended generative chat, because recommendations stay within a curated set of outfit rules and the retailer's own range.
How does M&S use AI to write product descriptions?
The same report stated that the share of M&S product descriptions written with AI had risen to 80% from almost nothing a year earlier. M&S has not published details of the tool, the prompts, or the review process behind those descriptions.
For any retailer considering a similar move, the general requirements are well understood even where M&S's specifics are not:
- Structured attribute data (fabric, fit, length, care, sizing) as the input, so that generated copy is accurate rather than plausible.
- Brand tone guidelines turned into prompts or templates, so that thousands of descriptions sound consistent.
- Human review rules, at least for new categories and for claims about materials or sustainability, which carry legal and reputational risk.
- Measurement of search visibility, conversion and returns for AI-written pages compared with older copy.
| Use case | What has been reported | Date reported | Audience |
|---|---|---|---|
| Personalised online experience | Products prioritised by body shape, style preferences and relevance | September 2024 | Online shoppers |
| Body shape and style quiz | Stylist suggestions plus shopper input; 450,000 users; 40 million outfit options | September 2024 | Online shoppers |
| Product descriptions | 80% of descriptions AI-written, up from almost nothing a year earlier | September 2024 | Online shoppers |
| Stock forecasting and ordering | Named by the CEO as an existing AI use | March 2026 | Internal |
| Marketing materials | Named by the CEO as an existing AI use | March 2026 | Internal |
| Microsoft 365 Copilot | 11,000 licences for store managers and support centre colleagues, with training | March 2026 | Colleagues |
Why is M&S investing in AI for online fashion?
The commercial motive is explicit. SupplyChainBrain reported that M&S aims for online to account for 50% of its fashion sales by 2028, up from about one third at the time. Personalisation and better product content are among the most direct levers on online conversion: if the right items appear first and their descriptions answer shoppers' questions, fewer customers leave or return goods.
The investment also sits on a long relationship with technology partners. In June 2018, Retail TouchPoints reported that M&S planned to integrate Microsoft AI technologies into its stores and operations, with Microsoft engineers working alongside the M&S Retail Labs team as part of a transformation programme to make it a digital-first business.
How is M&S using AI with store and office colleagues?
In March 2026, Retail Gazette reported that M&S had bought 11,000 Microsoft 365 Copilot licences for colleagues including every store manager and support centre colleague, supported by a training programme. In stores, managers use it to compile sales insights, summarise meetings, create shift rotas and prepare handovers. In the support centre, it drafts meeting updates and trading summaries and turns complex reports into recommendations.
Chief executive Stuart Machin said M&S was already embedding AI across the business, citing stock forecasting and ordering, the generation of marketing materials and a colleague help hub powered by AI agents. A store manager quoted in the report said she uses Copilot daily to prepare her morning huddle and shift handover notes.
What are the limits and open questions?
M&S has published usage figures, not outcome figures. The reports above do not state conversion uplift, return rates or sales attributable to the quiz or to AI-written descriptions. Body shape and style data is also personal information, so the value of personalisation depends on clear consent and careful handling. And AI-written product copy at 80% scale needs robust checks, because a small error rate repeated across a catalogue becomes a large number of inaccurate pages.
What can other fashion retailers learn from M&S?
Three lessons stand out. Start personalisation with stated customer preferences and stylist rules, not only inferred behaviour. Treat product description generation as a data quality project as much as a writing project. And give colleagues, not just customers, practical AI tools with training, because many of the gains in a large retailer come from faster reporting and planning rather than from customer-facing features.
Frequently asked questions
Does Marks & Spencer use AI?
Yes. M&S uses AI to personalise its online fashion experience, run a body shape and style quiz, write most product descriptions, support stock forecasting and ordering, and generate marketing materials. In 2026 it rolled out Microsoft 365 Copilot to 11,000 colleagues.
What is the M&S style quiz?
It is an online quiz that asks for size, body shape and style preferences and returns outfit ideas, combining in-house stylist suggestions with shopper input. By September 2024 it had 450,000 users, according to SupplyChainBrain citing the Guardian.
Are M&S product descriptions written by AI?
According to a September 2024 report, 80% of M&S product descriptions were written with AI, up from almost nothing a year earlier. M&S has not published details of its review process.
Which AI tools do M&S colleagues use?
M&S bought 11,000 Microsoft 365 Copilot licences in 2026 for store managers and support centre colleagues, with training. They use it for sales insights, meeting summaries, rotas and trading reports, Retail Gazette reported.
One edition every weekday morning. Read in five minutes. Free for industry professionals.
