Lumesa CEO Vita Mallela on AI models, inclusivity and the end of the catalog shoot

Inclusivity has long been an issue in the fashion industry. For most e-commerce brands, showcasing their clothes on a variety of bodies comes with a high cost that few can afford. Generative AI promises to break that arithmetic, allowing a single garment to be shown on dozens of bodies without multiplying the budget. But showing more bodies doesn’t necessarily mean paying for them, raising the question whether these emerging tools are really solving the diversity issue or optimizing it away.

Over the past couple of years, fashion brands like Anne Klein and Norma Kamali have partnered with Lumesa, an AI-native visual commerce platform that sells brands the possibility of showing full size range, and multiple ways of styling and wearing the same piece. The platform was created by Malavika Reddy and Vita Mallela — Mallela, now the CEO of the company, said the idea was born out of her own personal shopping experience.

“As shoppers, we rarely found anyone on a product page who looked like us. As operators inside e-commerce, we knew exactly why, and that's the part most people don't see,” she told FashionUnited.

Lumesa’s pitch is straightforward: it wants to reduce the costs of photoshoots and streamline content calendars.

“Photoshoots are being asked to do a job they were never designed for,” Mallela said, adding that a photoshoot for a single garment to be shown in a single color on 20 bodies would roughly cost between 19,000 and 38,000 dollars, when adding the work of models, photographers, hair, makeup, styling, and studio rental.

By using Generative AI tools, brands can submit a shot list with ideas and approve the output. Everything in between is done by the machine. “The conversation changes from what we can afford to show to who we are actually for,” Mallela said.

From a budget point of view, the case is hard to contest. From a human side, there are few questions that brands need to answer. The main one, perhaps, is if customers will trust what they see.

A matter of trust

In 2025, H&M found itself in the center of an online debate after debuting a campaign featuring digital twin models posing against stylised backdrops of cityscapes. An analysis of 1,339 English-language Reddit comments related to the campaign, published by the International Conference on Human-Computer Interaction, revealed that, while a minority saw AI models as innovative and potentially supportive of inclusivity and technological progress, the prevailing sentiment was negative. Later, a survey of 250 readers of Vogue in 2025 revealed that only 24 percent thought AI-generated images and videos for fashion brand campaigns could be as valuable as human-made content.

Mallela said that customers are asking for the diversity generative AI tools can help make accessible. “What customers respond to isn't how the image was made. It's finally seeing the content they've been asking for: the product on someone built like them, in the colorway they're considering, styled the way they'd wear it,” she said. That is only possible through the use of digital twins, which Mallela sees as “the most interesting part” of the process. “Brands can now work with people the agency system would never have signed,” she said, adding, “Try booking a size 8 model who is 5'6".

Transparency

The EU’s AI Act transparency rules took effect in August, requiring certain AI-generated content to be labelled and machine-readable. New York’s Fashion Workers Act (2025) mandated model consent for digital replicas. California’s proposed AI disclosure and likeness laws remain under debate, with no comprehensive statewide regime yet in force.

Lumesa supports clear, consistent disclosure, Mallela said. “We build for it rather than around it. Practically that means metadata on every asset, so a brand can show what was generated and when, and usage tracking so they know exactly where each image ran and for how long,” she explained.

Industry changes

The other issue is the disruption of a network of professionals who rely on photoshoots to pay their bills, such as photographers, models and makeup artists.

In a 2025 poll promoted by the Model Alliance with more than 100 models and influencers, an overwhelming majority expected AI to negatively affect their careers. The UK Association of Photographers, which represents around 3,000 professional photographers, called for responsible, safe, ethical and economically fair regulation of generative AI through a letter in January 2025. Later that year, in October, the Association of Illustrators, Society of Authors, Equity and others described generative AI training on creators’ work without permission as “the greatest act of theft in modern history.”

“Some catalog shoot days will change, particularly high-volume product work,” Mallela said about the professionals who might be affected by the new tools.

“What changes is how content gets made, not whether people are needed to make it. The roles shift toward direction. Somebody still has to decide what's on brand, what the story is and what's good enough to publish, and that judgment is now the most valuable thing in the process. We're watching entirely new roles open up around creative operations: people who own the brand standards, run the pipeline, and approve what ships. Those weren't jobs three years ago.”

Lumesa reports that brand partners have seen conversion rate increases of up to 30 percent. The company also offers 3D garment renders, with no physical sample being produced, shipped, or shot for those SKUs. “What we never touch is the brand's taste,” Mallela said. “We optimize the creative, not the cast. Styling, pose, background, framing. Who a brand shows is a brand decision and the system doesn’t get a vote.”