How Mango uses AI: generative design, virtual models and Mango Stylist

KEY TAKEAWAYS Summary by the editors

1. Mango has developed more than 15 machine learning platforms since 2018, applied to areas including pricing, design, personalisation and customer service.
2. In October 2023 Mango presented Lisa, an internal conversational generative AI platform for employees and partners, alongside Inspire, a tool that generates images mainly for its design teams.
3. Mango tested AI-generated models for its teen line in July 2024 and followed with a campaign created with AI models in autumn 2024, which drew public criticism.
4. In July 2025 Mango launched Mango Stylist, a generative AI fashion assistant for its Woman line on its online store and Instagram in nine markets, integrated with its after-sales assistant Iris.
5. Mango has not published quantified results such as cost savings, sales uplift or design lead-time reductions from these AI tools.

Mango uses AI across the value chain: an internal generative AI platform called Lisa, an image tool for designers called Inspire, AI-generated models in some campaigns and Mango Stylist, a customer-facing styling assistant launched in 2025. The Spanish retailer is one of the most openly experimental mid-market fashion companies on generative AI, but it has disclosed far more about what it built than about what it achieved.

What has Mango built?

Mango says it has developed more than 15 machine learning platforms since 2018. The publicly described tools fall into four groups:

Lisa, announced in October 2023, uses private models tailored to Mango as well as open-source models, behind a chat interface. Mango's director of technology, data, privacy and security, Jordi Álex, described generative AI at the time as a technology that "acts as a co-pilot".

For scale, Mango reported at the time revenue of 2.688 billion euros for 2022, operations in 115 markets and an online channel accounting for 36 percent of revenue. A business of that size produces a very large volume of product images, descriptions, campaign assets and customer conversations, which is the workload these tools are meant to support.

Why did Mango invest in AI?

Mango's stated reasons are speed and personalisation. Mango Stylist is part of the company's Strategic Plan 2024 to 2026, which emphasises technological development, data management and operational excellence. For campaign imagery, eMarketer reported in November 2024 that Mango wanted to accelerate content creation and strengthen its competitiveness against larger fast fashion rivals. A retailer that sells in more than 100 markets with a large online share needs to produce imagery, product content and customer service at high volume, which is where generative tools promise the most immediate savings.

Read also

How adidas uses AI: from design archives to generative content

How does it work (data, models, process)?

Mango has described its process in outline only:

• Design: Inspire generates images that designers use for inspiration on prints, fabrics and garments. Final designs are still developed by Mango's teams; the company has not said that AI produces finished garments.
• Campaigns: for the teen line, real garments were shown on AI-generated avatars instead of human models, and according to eMarketer the images on Mango's website carried disclaimers noting the use of AI.
• Customer assistant: Mango Stylist uses generative AI and conversational technology to recommend products based on stated preferences, suggest outfits and help with trends. It runs in chat on the online store and on Instagram, and hands over to Iris for pre- and post-purchase questions, creating one contact point.
• Organisation: Mango Stylist was developed jointly by IT, Data, Digital Product, Styling, Design, Visual Merchandising and Customer Service teams.

Analysis: the cross-functional team behind Mango Stylist is notable. A styling assistant needs product attributes and imagery from merchandising, a style point of view from styling teams and service processes from customer care. Assigning it to IT alone tends to produce a technically working but commercially weak tool.

What results has Mango reported?

Mango has not disclosed quantified results for its AI tools. There are no published figures on cost savings from AI-generated campaigns, on design lead times with Inspire, on adoption of Lisa among employees or on conversion and basket effects of Mango Stylist. The available evidence is about scope: the number of platforms, the nine markets for Mango Stylist (Spain, Portugal, the United Kingdom, France, Italy, Germany, Austria, Turkey and the United States) and the integration with Iris. Readers should treat claims that Mango cut production or photo costs by a specific percentage with caution unless they cite the company.

What are the limits and open questions?

The AI models drew the most public debate. eMarketer noted some scepticism among consumers. Analysis: generated models raise questions about jobs for models and photographers and about whether generated images give a fair impression of how clothes look on real bodies. Labelling the images addresses transparency, but not those underlying questions.

Other open questions are commercial. Mango Stylist launched for the Woman line only, so its usefulness for men's, kids' and teen ranges is untested in public. Generative design tools raise questions about the training data behind image models and about intellectual property in generated prints. And without published metrics it is hard to judge whether the platforms have changed margins, speed or customer loyalty.

Read also

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

What can other fashion companies learn?

Analysis: Mango's experience is useful for mid-sized brands because it shows a sequence: internal tools first, then marketing content, then customer-facing assistants. Each step raised the stakes, from employee productivity to brand reputation and customer trust, and each needed different data and governance.

1. Build internal capability first. Mango's internal platform, Lisa, preceded its consumer launches by almost two years.
2. Label generated content. Disclosing AI imagery is becoming a baseline expectation, and it reduces the risk of misleading customers.
3. Connect assistants to service. Linking a styling assistant to an existing after-sales bot avoids two disconnected chat experiences.
4. Expect scrutiny on people. Replacing models or photographers with AI affects reputation as well as cost, so the decision needs more than a budget case.
5. Define success up front. The absence of published results is a reminder to agree measurable goals before scaling.

Frequently asked questions

What is Mango Stylist?

Mango Stylist is a generative AI fashion assistant launched by Mango in July 2025. It recommends products and outfits for the Woman line through chat on Mango's online store and Instagram, and it is integrated with the after-sales assistant Iris. It launched in nine markets including the UK, Germany and the US.

Does Mango use AI-generated models?

Yes. Mango tested AI-generated models for its teen line in July 2024 and ran a teen campaign with AI models in autumn 2024. The clothes were real, the models were AI-generated avatars, and the images on its website carried AI disclaimers.

What is Mango's Lisa platform?

Lisa is Mango's internal conversational generative AI platform, presented in October 2023. It is used by employees and partners for content creation across the value chain and combines models tailored to Mango with open-source models.

Does Mango use AI to design clothes?

Mango's Inspire tool generates images that help design teams with prints, fabrics, garments and displays. Mango has not said that AI designs finished garments on its own, and it has not published data on how Inspire changes design speed or output.

GuideThe complete guide to AI in fashion design and product developmentRead the complete guide