by
“Here’s 200 jackets, pack shots. Can you please put them all in models? … Can you put these earrings on, can you take this handbag, and can you give it a 360-degree spin?”
In a recent webinar hosted by the global fashion wholesale platform JOOR, Simon P. Lock, Founder and CEO of the fashion technology company ORDRE Group, gave a glimpse of what the group’s clients are requesting today and how the company’s way of work has already changed using artificial intelligence tools.
While much of the retail industry’s discussion about artificial intelligence is still framed around the future and what AI will eventually be able to do, at ORDRE Group AI tools have already found their way into daily business. With ORDRE’s ORB AI platform, complete photo shoots are created without photographers, studios or physical sets. Fashion models who do not exist wear real collections. Videos can be generated from simple product images. Campaigns can be adapted almost instantly for different channels, markets and audiences. And the economics are changing just as dramatically as the technology.
The French ORDRE Group is by no means the only company that has fully embraced AI in the content production process. The Hamburg-based SCAYLE Studios are another example revealing how far the technology has progressed into everyday operations. The company offers its clients an AI-powered production environment capable of turning basic product images into complete e-commerce image sets, including AI-generated models and videos. Fashion brands such as Betty Barclay, s.Oliver, Guido Maria Kretschmer, EDITED and ABOUT YOU are alongside more than 100 additional brands that have embraced SCAYLE’s AI platform as a highly economical tool.
By July 2026, ABOUT YOU had already generated more than 120,000 images using the technology. According to SCAYLE Studios, the company has achieved production cost savings of more than 90 percent and annual savings of more than €8 million, alongside a 9.2 percent uplift in gross merchandise value and a 5.1 percent increase in add-to-basket rates.
Although mentioned brands may still be among the first movers, the operational figures clearly indicate that artificial intelligence technologies are no longer confined to experiments in innovation labs. And they point towards something considerably bigger than cheaper fashion photography.
From One Product Image to Unlimited Content
The traditional economics of fashion imagery are relatively straightforward. Products must be transported. Studios need to be booked. Models, photographers, videographers, stylists, make-up artists and production teams need to be available. Every additional image, setting, market or video creates additional cost.
AI is now changing that equation. Ordre Group CEO Lock explains: “I can’t really afford to do an extended video shoot. I can’t really afford to shoot everything individually. So can I use AI to really enhance my suite of digital images?”
The simple answer is: Yes. AI can do more than create a single image. It can multiply an existing asset. A conventional packshot can become a model shot. The model can change. The setting can change. A still image can become a 360-degree presentation or video. Products can be combined into complete looks. Backgrounds can be adapted to seasons, markets and campaigns. A collection photographed in a day can subsequently become the raw material for an array of additional assets. Even CAD drawings can be transformed into realistic product imagery, even before physical samples are available. At dramatically lower cost than ever before.
The result is a transition from content production to content generation.
Content used to be scarce because every new asset required production resources. Generative AI makes the marginal cost of another variation dramatically smaller. Once the product information, brand rules and creative direction are available, potentially hundreds or thousands of derivatives can be created.
Lock summarizes the old economics rather nicely. Turning 50 packshots into 50 conventional videos means another shooting day, another model, another photographer, another videographer, more lighting, “and they’ve all got to have lunch.”
AI does not eliminate all those professions. But it fundamentally changes the production equation.
When Content Becomes Abundant
The consequences extend beyond savings. AI opens the door to an almost unlimited number of variations.
When producing another image or video becomes inexpensive, brands and retailers no longer need to ask only: Which content can we afford to produce? They can ask: Which content would work best in this situation?
The same dress could be shown on different body types, age groups and models. The environment may change according to season or geography. The imagery presented to a customer in Vienna need not be identical to that shown to someone in Madrid, Dubai or Stockholm.
Matthieu Blanc, Machine Learning and AI Product Specialist at Google, explains that current systems are already being deployed for applications including virtual try-on, 360-degree product visualization and the contextualization of marketing assets according to season, location or environment.
The technology is also becoming conversational. Instead of generating an image and starting again when something is wrong or needs to be edited, users can increasingly tell the model what should change and what must remain untouched. That sounds like a technical improvement. Commercially, it is much more significant. It transforms AI from a generator of interesting images into a production tool that can gradually be integrated into professional workflows.
The Rise of the Virtual Model
The implications become particularly visible when people enter the picture.
ORDRE is already developing a set of virtual models for one client that can be used seasonally, with different models matched to different products and territories. “The creative process in creating those models is just as extensive as casting for a real model,” Lock says. “But once it’s done, it’s done.”
A virtual model can become a reusable brand asset. Once the desired appearance and visual identity have been established, that model can potentially wear hundreds or thousands of products. It does not need to travel. It does not have scheduling conflicts. Its appearance remains consistent. And different virtual models can be created for different markets or customer groups.
This does not necessarily mean the end of human models. It does, however, begin to separate two things that fashion has traditionally bundled together: the human image and the human being behind it.
For basic product presentation, a retailer may simply need a body and face that present a garment effectively. Increasingly, AI can provide that. A human model, celebrity or influencer can provide something else: personality, reputation, community, credibility, cultural relevance and a relationship with an audience.
That distinction could become increasingly important for the influencer economy as well. If a brand merely requires attractive people to wear products across hundreds of social media assets, paying humans for their image becomes harder to justify when synthetic alternatives can be generated almost instantly. Human influencers will therefore have to offer what artificial ones cannot easily reproduce: genuine authority, expertise, entertainment, community and trust. AI may not eliminate the influencer. But it makes being merely photogenic much less valuable.
Lock himself draws a similar distinction. Physical models who have a profile, produce social content and act as brand ambassadors perform a fundamentally different role from virtual models used to present products. At the same time, he says the fashion industry needs “a big conversation about how it’s going to manage virtual models.”
There is another paradox. As production becomes increasingly automated, creative judgment may become more important rather than less.
More AI Could Mean More Human Creativity
“Creative input is maybe paradoxically extremely important in this new AI era,” Blanc says. Generating an impressive image is becoming easy. Ensuring that thousands of generated assets remain faithful to the product, brand identity and creative intention is not. Fashion brands will not accept a garment that is almost correct or a logo that is approximately right.
Blanc therefore expects creative professionals to use their visual knowledge to control the technology, adding that they will be able to produce “way more assets” than before the AI era.
That suggests a broader organizational change. The scarce resource in marketing may increasingly cease to be production capacity. It becomes judgment. When content becomes unlimited, editorial and creative selection become more valuable. The critical questions become: What should the brand look like? Which stories should it tell? Which images are meaningful? Which variations deserve to reach consumers? Which experiences strengthen the brand, and which merely add more digital noise?
From Media Production to Product Creation
But focusing only on advertising underestimates what is happening. “We see the fashion industry using also this AI model in the fabrication process and being able to imagine products that were not imaginable beforehand,” Blanc says. He points to AI being used not merely in the presentation of fashion, but increasingly as part of the creative and product development process itself.
This is where the implications become much larger. Blanc compares AI to earlier creative breakthroughs such as photography, the digital camera and computer-generated imagery (CGI) – a technology that became famous by being used in blockbuster movies such as Avatar.
Imagine combining generative design with customer data, body scanning, computer vision and increasingly flexible manufacturing. A customer might no longer browse only through products that already exist. She could begin with an existing dress and ask for a different color, altered neckline, longer sleeves or another material. AI may visualize the result immediately.
Take the idea one step further. A body scan or a handful of smartphone images may provide measurements. The garment may then be adapted not merely visually but physically to the individual customer’s proportions.
The traditional fashion model is based on producing a limited number of designs in standardized sizes and attempting to predict how many units consumers will buy.
Generative design and increasingly automated manufacturing lead to another possibility: AI does not merely generate the image. It helps generate the product itself. The product becomes variable. The retailer might no longer sell only 50 dresses. It could sell 50 starting points from which an unlimited number of variations are generated.
“You will be able to send anything as an input, text, image, video, audio, and you will get whatever you want as an output”, Blanc explains why this isn’t simply an image-generation revolution. The boundaries between image generation, video generation, product visualization and other forms of content are disappearing. The input and output increasingly become interchangeable.
What Happens to the Store?
That scenario has profound implications for physical retail. Today’s fashion store is largely an inventory system that customers can walk into. Considerable space is required because products need to be presented in different colors, styles and sizes. The store carries inventory because customers need to see products, touch them and try them on.
But what if seeing, touching and trying no longer require dozens of physical variants? Imagine a store that carries only one carefully presented sample of a particular dress.
The customer can touch the fabric, examine the stitching and understand its quality. She stands in front of an intelligent mirror or another spatial interface. The system recognizes her dimensions and shows how the dress would look on her body.
She asks for another color. It changes. Longer sleeves. Done. A different pattern. Done. Perhaps the system suggests an adjustment based on her proportions or shows combinations with products she already owns.
When she is satisfied, one click places the order. The garment is produced or taken from the most suitable inventory location and delivered to her home, hotel or the store for collection.
At that point the shop is no longer primarily a place for storing merchandise. It becomes a physical interface between people, products, brands and digital intelligence.
Less Inventory, More Experience
The potential consequences for retail real estate are significant. If retailers eventually need fewer units of each SKU on site, the relationship between sales and physical space changes. Stockrooms could shrink. Some stores might become smaller. Others might use the space differently.
The value of physical retail would increasingly lie in precisely those things AI cannot digitize: touching materials, testing products, receiving human advice, discovering something unexpected, social interaction, hospitality, service and the emotional experience of being somewhere.
Paradoxically, more digital retail may make the quality of physical retail even more important. A mediocre store whose main function is providing access to merchandise has little advantage when virtually unlimited merchandise can be accessed digitally. A compelling physical environment is different.
For shopping centers and mixed-use destinations, that could push the industry further away from thinking primarily in terms of space for inventory and towards spaces for interaction, discovery and experience. Stores may become showrooms, studios, consultation spaces, fitting environments, community hubs and fulfilment points simultaneously.
And the border between e-commerce and physical commerce becomes increasingly meaningless. The customer may discover a product online, modify it in a store, discuss it with a human advisor, visualize it through AI and have the finished product delivered somewhere else.
Which part of that transaction was “online”? Which part was “offline”? It is simply retail.
Reality Still Matters
None of this means every image, model, product or store will become artificial. There are important limits.
Blanc points to beauty as one example. Brands may use AI extensively for contextual product imagery while rejecting synthetic representations of real skin where the authenticity of a product’s effect is essential. Different fashion brands will draw the boundary in different places. He compares today’s questions around AI imagery with earlier debates surrounding Photoshop and argues that disclosure will remain important.
The central question therefore is increasingly not whether AI can generate something. In many cases, it can, even when the result would have been difficult to imagine only recently. The more consequential question is whether doing so generates value.
Consumers may accept an artificial model demonstrating a handbag while demanding authenticity from an ambassador recommending it. They may happily visualize themselves wearing a digitally generated version of a jacket while still wanting to touch the actual fabric before spending €800 on it.
Human authenticity and artificial creation are unlikely to be mutually exclusive. Retail will have to learn where each is most valuable.
The Real Transformation Has Only Begun
Perhaps the biggest misconception surrounding generative AI in retail is that we are still waiting for it to become good enough.
Examples such as SCAYLE Studios, ABOUT YOU and the production workflows described by ORDRE and Google suggest otherwise. The quality threshold has been crossed for a growing number of commercial applications. The economics are changing with it.
Tomorrow, the same logic now transforming media production could move deeper into the value chain. From unlimited images to unlimited product visualization. From visualization to generative design. From standardized products to individualized products. From stores filled with inventory to stores built around discovery, interaction and service. And from retail real estate as the place where products are stocked and sold to places where physical experience connects with almost unlimited digital choice.
The most interesting question for the retail industry may therefore no longer be what AI will eventually be able to do. We can already see enough to ask a more consequential one: If products, models, content and customer experiences can increasingly be generated on demand, what should the physical world of retail become?
That is where the AI discussion becomes a placemaking discussion. And where AI becomes much more than a technology story.
