Generative artificial intelligence (AI), or gen-AI, is a machine learning form that generates new content from user prompts, input, or triggers. Today, text and images represent the pinnacle applications of this technology. Across different domains and industries, there is a unified drive to broaden its application to fields such as animation, sound effects, music, integrating virtual characters, and leveraging AI in fashion for collection design.

Quick insights

• Most brands using gen-AI in fashion rely on tools developed by foundational AI companies like Midjourney. Currently, achieving high-quality output is time-consuming, with users depending on textual prompts to master the proper wording and combinations.
• Brands enhance the output through existing design tools, highlighting the need for solutions that grasp design principles and materials better.
• AI-generated designs are inherently digital, so proficiency in digital rendering for physical products is crucial. Solutions in this field must seamlessly integrate with existing tools and position themselves as more than just inspirational aids but facilitators for the subsequent manufacturing steps. Generating design ideas differs from creating a complete 3D model ready for production.
• Human intervention remains vital in guiding and optimizing the final output.
• Current tools face challenges in understanding clothing structure and delivering comprehensive designs.
• The key advantages lie in time efficiency and creative inspiration.

How can gen-AI be applied to the fashion industry?

From the 2023 AI Fashion Week to Casablanca’s AI-generated campaign, the fashion industry has been one of the most explorative with the technology. Projects are being done in collaboration with agencies like Maison Meta and Lalaland.ai and sometimes in combination with domain-specific software. According to McKinsey, generative AI could increase the apparel, fashion, and luxury industries’ profit by $275 billion in the following three to five years.

One of the most common use cases tested in the brand and retail industries involves marketing automation with companies like Pimento for marketing imagery inspiration, Writer for SEO or e-commerce content optimization, and Jasper AI for copywriting. For instance, Casablanca used the technology for its S/S 2023 marketing campaign. Another tested use case has been around online shopping assistance and customer service tailored to shopping. Companies like Kering, a luxury group, and Zalando, an online fashion retailer, launched a chatbot using ChatGPT to give product recommendations.

RELATED: Unleashing the Future: AI Innovation Reshaping Industries Today

Finally, product design is also seen as one of the use cases with the most potential, given the nature of the technology to support creativity, give answers, or provide newness. McKinsey highlights that design and product manufacturing are where technology will bring the most value. This is also supported by BCG’s study, stating that using AI for creative innovation performed about 40% better than a control group that did not use it. Conversely, participants who used AI for business problem-solving performed 23% worse than the control group. Product design is a critical step in manufacturing for any physical consumer product company. The current process involves a combination of methods and tools depending on the industry and the designer’s personal and preferred creative process. According to Business of Fashion, generative AI is a strategic priority for 73% of fashion executives, and 28% have tested it in the design and product development processes.

Looking at the current design process, almost every product is built, starting with an inspirational phase, then building prototypes, and then the development phase. Using AI can help designers be inspired by getting more ideas. The added value comes mainly from AI models being able to ingest a large amount of data like trends, past collections, and customer interests to provide ideas.

Many brands have explored the use of this technology to build collections and products, especially in 2023. The benefits of the technology can vary depending on whether a fast-fashion or luxury brand uses it. Fast-fashion or mid-market brands can use generative AI to create more designs based on the latest trends. In contrast, luxury companies might use it more at the start of the process purely for mood board visualization using different sources of inspiration.

One can say that the challenge of using the technology is that the final rendering does not reach current standards in terms of structure or design. However, these tools should be used in parallel with human creativity. Some designers mention that “hallucinations” or errors made by AI systems can also be a source of inspiration.

Who are the space's key players, and how do they tackle AI in fashion?

The use of gen-AI is still in its early stages, given that it has been less than two years since tools have been developed and commercialized. In product design, solutions are at an even earlier stage due to the diversity of products to adapt to - for instance, cars, jewelry, and furniture. Currently, the market is structured between existing design tools, like Cala, that integrate gen-AI as an added feature, and emerging companies purely focused on asset generation, like cre[ai]tion or Raspberry AI. Using gen-AI can also serve benefits for optimized customization. Indeed, by integrating consumers’ preferences from first-party data collection or their own inputs like images and text, brands can create a new type of consumer relationship by manufacturing personalized products. Raspberry AI is even more focused on the fashion industry, providing market search for trends detection to accompany designers right from the start of the process. Six Atomic enables optimized production, with one of its enablers being rapid prototyping thanks to AI.

Overall, the maturity of adoption from fashion brands is correlated with the emergence and maturity of startups in the space. Acne Studio launched a collection in 2020 with AI's help; since 2023, many brands like The Kooples have been using the technology. Currently, no brands are seen leading the space, as most have been in the testing phase or not sharing results. The key challenges in adopting these tools have been the actual interest of designers in using them and the ability of software to provide the proper interface for creative people.

AI Fashion Week 2023: Revolve x Maison Meta

For instance, Revolve partnered with Maison Meta to organize an AI Fashion Week in April 2023. Most designers directly used Midjourney and Stable Diffusion with Photoshop to make final edits. Revolve also launched a marketing campaign that involved models, illustrations, and clothes being generated with AI. The imagery includes large mushrooms and flowers, which would not only be complex to set up in real life but also take much more time and be more costly to render, even with traditional computer-generated design methods.

RELATED: 5 Startups Share Their Views on the Future of Commerce

Our discussions with designers, not only in fashion, usually show mixed reactions to the interest in using the tool. For instance, one emerging bag designer based in France focused on using unused fabric is not interested in moving faster when looking for new designs. In fact, he enjoys taking the time to follow his own personal creative world and identity. On the other hand, when speaking with one of the designers who participated in the AI Fashion Week 2023, she mentioned how much she enjoyed playing with the different AI tools to amplify her creativity. First, she had a concept in mind and then used different tools, from Stable Diffusion to Midjourney, to create ideas. Then, she selected the pieces she wanted to dive into further and created a technical drawing from there. Only then did she use software like Clo3D to create the 3D visualizations and patterns. Hence, she perceived the AI tools as seamlessly working alongside the traditional workflow. There are limitations, however, as one still has to use different tools, so an appropriate user interface is needed for designers. Also, these public tools are not granular enough, so she could not touch up the final details. Hence, existing editing tools are necessary for post-production. Moreover, gen-AI tools still cannot go directly from image to pattern.

Source: AI Fashion Week website

Collina Strada

Founder Hillary Taymour used Midjourney to support her ideation process for her S/S 2024 collection and sees the technology as a crucial tool for upcoming collections. The brand adopted a feedback loop method by feeding the tool with past archives and collections, making and starting the process again.

Source: Business of Fashion

Heliot Emil

The Copenhagen-based brand Heliot Emil partnered with an AI company to build a custom algorithm and feed it past collections to create new products.

Source:GLASS

Interview with Marco Limm, Founder and CEO at cre[ai]tion

Cre[ai]tion is an AI-powered tool that allows any type of fashion, automotive, footwear, and jewelry designer to create designs and products through an easy-to-use interface.

Could you describe the state of the market and developments in the past year?

The excitement of late 2022 and the entirety of 2023 is currently transitioning into practical implementation at the end of 2023. 2024 is poised to be the year of substantial implementation. In the realm of gen-AI within the design space, testing only began in 2023. At cre[ai]tion, our goal is to bridge the gap between designers and AI models to facilitate the creative process as the field evolves.

How will gen-AI impact the design and product development process?

I won't make a definitive prediction, but it will change everything. The design flow and product creation process will be completely rethought and different. I hope that we build better products thanks to AI. For the first time, you can let the customer create and start the creative journey. The next stage will be mass market personalization.

Currently, user interactions with models are limited. Methods like Stable Diffusion or other solutions are cumbersome. It should be visual, not textual, so people can focus on what they excel at — the design aspect. This allows users to concentrate on the creative flow and enjoy upgrading product features.

Cre[ai]tion is an AI-powered design tool that supports theinspirational phase of design. We aim to be like Instagram, providing customers with a feed. Moving into theproduct creation phase, cre[ai]tion allows users to see their entire archive and every step in the design process, presented as a flowchart. The third step is theproduction phase, where cre[ai]tion can significantly reduce time-to-market.

The fashion industry stands to benefit greatly from this type of solution. For example, in fast fashion, Western companies are under pressure from Asian competitors to produce faster. Overall, industries across the board are reaching out to us. Any company that creates physical products requiring inspiration and creativity can leverage AI in their processes. For instance, we are working with Blanco, a water fountain company.

Which part of the design process will benefit the most (inspiration, prototyping, production, etc)?

While cre[ai]tion can provide value throughout the design process, it shines brightest in the inspirational phase. It empowers designers to explore and brainstorm various designs easily, fostering creativity from the very beginning.

Your tool is quite impressive in the way it simplifies the brainstorming process for designers. Can you explain the rationale behind this user interface?

Our user interface is designed to make the creative process intuitive and accessible. We believe that by providing designers with a user-friendly platform, we can encourage more innovative ideas and streamline the initial design phase, setting the stage for a smoother overall process.

RELATED: An Essential Guide to Corporate Innovation

Moreover, our interface aims to help designers tap into a unique emotional state — an inherent understanding of the direction a design should take, even when it can't be easily articulated at the outset. We recognize the importance of designers entering a flow state, where creativity flows effortlessly and exploring the vast realm of possibilities becomes second nature. In this state, known as "Möglichkeitsraum," or the space of possibilities, designers can push boundaries and discover novel directions for their designs, leading to groundbreaking outcomes. Our interface is meticulously crafted to facilitate this journey, ensuring that designers can immerse themselves in the creative flow and unlock their full potential.

What is a key trend in the space?

I see mass personalization and customization as a consequence of gen-AI-enabled designs. For example, companies have trouble getting people into their shops, and they're struggling to have consumers connect with their brand. Imagine that you go into a shop in the future and create your own running shoe or whatever apparel product, and then maybe several hours later or the next day, you have the shoes printed. You can pick it up, and then you have a unique pair of shoes.

If everybody can create their own Adidas or Nike shoes, what does this mean for them? I don't know yet. But I think that's a really interesting question for them to ask. And that is also what we are researching at cre[ai]tion.

How does this align with other strategic topics, such as on-demand production?

Cre[ai]tion aligns perfectly with the concept of on-demand production. By streamlining the design phase and reducing time-to-market, cre[ai]tion supports the agile production of products based on real-time demand, making it an ideal fit for this strategic approach.

How do you see gen-AI-enabled design tools as part of the already existing ecosystem of tools (for instance, Clo3D in fashion)?

The tools will still be there, but the creation process is before the actual creation of patterns, 3D models, or even 3D simulations. Hence, a solution like ours would be used before. In the future, however, the vision is for cre[ai]tion to be one platform, so instead of having 2D designs on cre[ai]tion, we want to also create 3D models automatically.

How can gen-AI tools establish defensibility?

Ultimately, success hinges on offering a superb user experience in tandem with a high-performing model. How you interact with the model is just as crucial as the model's performance itself. Customized model training for each customer within their unique environment also provides a competitive edge.

At cre[ai]tion, our team consists of a diverse mix of designers, tech experts, and business professionals. The user interface we are developing is unquestionably one of cre[ai]tion's unique selling points.

How do you envision cre[ai]tion being used by luxury fashion brands versus fast-fashion brands?

Cre[ai]tion can cater to both luxury and fast-fashion brands. For luxury brands, it can enhance the creative process and provide a platform for designing exclusive, bespoke products. In the fast-fashion segment, cre[ai]tion can accelerate design and production processes, helping brands meet the demand for quick product turnover.

At Plug and Play’s Brand & Retail vertical, we have some of the first startups making innovations with AI in fashion. Join our platform today.