This article was originally published in The Interline’s AI Report 2024. To read other opinion pieces, exclusive editorials, and detailed profiles and interviews with key vendors, download the full AI Report 2024 completely free of charge and ungated.
Key Takeaways:
- The fashion industry, like other sectors, is undergoing significant changes due to generative AI, which is revolutionising higher-order tasks and basic societal pillars.
- Brands face numerous challenges at various layers, including financial risk, skills scarcity, and time constraints, which generative AI can potentially address to streamline processes.
- The proposed AI-native, community-led brand framework suggests a collaborative ecosystem where generative AI removes creativity bottlenecks, enables market validation before production, and matches designs with manufacturers for efficient product creation.
Technological change comes in waves. This is something we all understand intuitively, because we are accustomed to having it happen to us and around us. In telecommunications, we went from 1G to 5G. In transportation, we moved from mechanical, to fossil, and now electric energy. Our monetary systems transformed from basic barter to forms like trading cards, gold, paper money, and now, digital currencies. There are thousands of examples like these.
Even if the difference between how things are and how they were in the past can feel enormous in retrospect, transformation often happens incrementally, going broadly unnoticed until we zoom out and realize the progress we’ve made.
Yet, every so often, a breakthrough technology emerges that is so revolutionary it catalyzes an entirely new paradigm and leaps us forwards in a way that’s not just perceptible but extremely prominent. These types of new frameworks do not simply build incrementally on what came before. Rather, they introduce a transformative way to rethink old problems, opening doors to solutions previously unimaginable, and they give rise to new opportunities that were impossible to envision – at least with any degree of practicality – previously.
These types of innovations are black swan events that fit the adage: “there are decades when years happen, and years when decades happen”. I don’t think it’s an exaggeration to say that 2022 and 2023 were examples of the latter.
The rise of generative AI, I think (and many agree) , represents such a pivotal shift. Vaulting to the front of people’s minds in late 2022 and continuing to up-end industries, our understanding of “work” and much more over the 18 months that have followed, generative AI models as experienced through user-facing applications like ChatGPT and Midjourney are fundamentally changing how we think about not just higher-order tasks but some of the basic pillars of society.
And the world of fashion is not exempt.
As entrepreneurs, designers, and trendsetters, the disruptive potential of this technology forces us to question the current state of affairs and wonder what the world could look like if we restarted anew, leveraging the cutting-edge technology available today. How different would our processes, industry dynamics, and economics be? Would our products be better or worse? More human, or less?
What would fashion look like if we built it with AI in mind?
Would we, after some initial experimentation, wind up settling for some fairly prosaic use cases and getting incremental innovation, or would the re-imagined industry be a transformed one, pushing us to rethink the things we considered to be hard truths about the fashion beforehand?
Brands’ Main Challenges
First, let’s ask the question: what are the biggest challenges brands face today?
The brand value chain is both complex and simple at the same time. Complex in that it involves a large number of different parties – each with their own set of incentives. Simple in that this framework has been, by most measures, unchanged for the last 500 years. To an outsider, the way fashion products make the journey from idea to market looks, frankly, messy. To an insider, that complexity and disorder is just the way things work and have always worked.
Below is a highly simplified view of a brand’s value chain as it exists in 2024. I want to walk you through it, step-by-step, starting with the context of the industry itself.
Layer 0: Industry Dynamic
The first challenge of a traditional brand value chain lies in its inherent competitive dynamic. Brands all compete for mindspace and share of wallet, bidding against each other to the point where unit economics have become broadly unsustainable for most players.
At first glance, this doesn’t seem like a challenge AI can fix. Every industry that has a number of players growing faster than its market size is exposed to economic cannibalization and a race-to-the bottom in terms of profitability.
From a purely commercial perspective: with low (and lowering) barriers to entry, the number of brands in the fashion industry should continue to grow until a market equilibrium is reached where no one is truly making money, and where weaker actors are competed out of the market.
Layer 1: Inspiration
The inspiration layer for most of us is the fun one. It’s where passion lives. At this layer, designers are thinking: What should I create? What message do I want to share with the world? While this layer is obviously informed by the previous one – market realities increasingly dictate what gets made – this is also where the ineffable “creative” spark sits. This layer also flows through many stages, namely: personal experience, idea, and sketch.
Here the challenges are obvious: there is a limited supply of great ideas, getting unbiased market validation on the merits of an idea is hard, and transforming great ideas into great sketches requires technical skills that not everyone has. The technical skills required to transform an idea into a sketch reduce by several orders of magnitude the number of inspirations we get to ultimately consider to be developed into new products.
Layer 2: Product creation
The product creation layer is my personal favorite. It’s technical, messy, and requires a lot of coordination with outside actors. This layer is defined by the following stages: tech packing / 3D design, manufacturer selection, golden sample production (often with many rounds of iteration), inventory production.
Challenges are basically all there is at this part of the process! But if one wants to be specific, here are the key pain points most brands face:
- Transparency: it is hard to find good manufacturers and it is even harder to know if you are paying a fair price to work with them.
- Parallelization: each brand has to do the same work as the others. Functionally the same tech pack will be designed thousands of times by thousands of different brands, with every one reinventing the wheel multiple times over.
- Skills: Tech packing and 3D design, visualization and simulation require expertise, as well as dedicated hardware and software for the latter.
- Quality assurance: Finding the good manufacturing partners is costly as it involves good through multiple rounds of sampling, knowing what to look for, etc.
- Financial risk: Finally, the biggest challenge of this phase is the negative working capital involved in building inventory before knowing if you can sell it. This is the main reason why so many brands die or never see daylight: guessing what the market wants, committing to producing it in volume, and then finding that it doesn’t sell is an extremely common occurrence.
Layer 3: Commercialization
Finally, the third layer: getting sales.
This one has already been transformed by AI in a meaningful way through product recommendation algorithms, better ad targeting, dynamic pricing, etc.
Commercialization can be broken down into two phases: (i) marketing – i.e. creating a purchase intent for your product, and (ii) commerce – i.e. handling the mechanical aspects of completing a transaction (payment processing, ERP update, fulfillment, shipping, duties optimization, after-sale support, etc).
The commerce side is usually what people identify as the “boring” part of the business. The name of the game here is efficiency more than anything else. There are best practices, and for most brands, the key is to follow them diligently, not innovate. Other industries, too, have already set fairly cast-iron templates for having the most seamless routes to market.
The marketing side gives more space for creativity and experimentation.
Which channel to use? How to frame your messaging? How to design visuals that stand out and stick? Here the biggest challenge is optimizing ROAS (return on ad spend) by creating better promotional material, faster, and distributing it to the right people, at the right time, and in the right way.
Wrapping Up On Challenges
As we’ve seen, there are a bunch of challenges that brands need to overcome at the inspiration, product creation, and commercialization layers. Although there are hundreds of them individually (and any one can easily become a bottleneck that holds back the speed, quality, creativity or other outcomes of a final product), we can bundle them into three core categories:
- Financial risk: (i) product development risk, (ii) inventory risk, (iii) high customer acquisition costs
- Skills scarcity: shortage of (i) inspiration, (ii) sketches, technical, and 3D design skills (iii) quality insurance expertise, (iv) marketing efficiency
- Time constraints: the inability to move through each layer faster, or even in parallel.
Leveraging GenAI to Propose a New Framework
Now that we better understand the inherent challenges coming with launching a fashion brand or a new collection, let’s explore if and how generative AI might help us.
From that perspective, the first question to ask is: has it done so already? Did the world truly change over the last year or so in a way that’s delivered meaningful AI applications into the hands of brands and retailers? A few observations:
- We have already gone from a world where only a handful of trained designers can conceive breathtaking concept mock-ups to one where everyone with taste has that capability. We hear a lot about generative AI’s ability to “democratize” creativity (and there are certainly strong opinions on both sides as to whether that is a desirable thing) and this is the neatest encapsulation of it. If you want to bring an idea you have for a product – no matter whether it fits into the apparel, footwear, or accessories categories – to life, you can now do so without needing to learn to draw or model in 3D.
- We have already moved from a world where creating captivating marketing materials and validating consumer demand could only be done post production with professional assistance, to one where AI can do the job at much earlier stages of the product lifecycle, in a way that requires far fewer professional skills.
- We have already moved from a world where manufacturers had little visibility and exposure to end-buyers to one where data can both inform and connect them with customers. And while that data itself is not generated by AI, an increasing number of supply chain connectivity and visibility platforms are making extensive use of AI to provide clarity, accountability, and a new channel for engagement and exposure.
- We have already moved from a world where creatives might be able to collaborate instead of compete, to one where universal accessibility of digital creative tools, digital production methods, and community sourcing platforms has made it easier than ever to work together and to make use of both innovative manufacturing approaches and collective buying power across traditional ones
Taking stock like this, I think it’s important to recognise how quickly some of the fundamentals of fashion have already changed. But, building on these facts, let’s consider how AI could empower us to propose a completely rethought value chain for emerging brands.
The framework proposed below obviously has its own flaws and does not claim to be the answer to all of the industry’s challenges. However it is a new structure only made available by the rise of generative AI we’ve witnessed over the last year – and by linking those new possibilities into some of those pre-existing innovations I listed in design, marketing, and manufacturing.
I believe this approach is interesting not just because of its innovative approach, but because it also offers a realistic way for emerging entrepreneurs to offload risk. For the moment, let’s call it the Community-Led Brand Framework, and let’s consider how it would differ from where fashion finds itself today.
Layer 0: Dynamics
The traditional competitive landscape, where brands vie for market share and visibility, is transformed into a collaborative ecosystem. In this community-led dynamic, every new player – creator or consumer – adds value. Network effects, often non-existent in the classical approach are now amplified as each participant not only consumes but also potentially contributes ideas, enriching the pool of designs.
By crowdsourcing ideas and having all involved actors incentivized in growing the network, the community-led brand transforms the net zero competitive nature of a multi-brand ecosystem to positive-sum multi-creator ecosystem.
This may sound unrealistic in such a cutthroat market, but we only need to glance at the communities that have grown up around online creators and hobbies, or the communities built around AI tools like Midjourney, to see that this model is not just feasible but that it also potentially better fits the reality of content creation and consumption in 2024.
Layer 1: Inspiration
In a generative AI-infused world, the bottleneck of creativity can be removed. Anyone with taste and patience can generate an amazing product mock-up, irrespective of technical skill. Market validation, previously a stumbling block, is now facilitated by a captive ecosystem of actors voting on the concepts they would like to see move forward.
This measurably meritocratic community idea validation process derisks the chances of producing something nobody wants, turns passive buyers into active participants, and more importantly, allows creators to finetune their approach and designs, significantly accelerating the feedback cycle necessary to their development as designers.
There are, of course, social and copyright considerations to take into account here, but broadly speaking the closer fashion gets to culture, and the more the two are allowed to grow in symbiosis, the better they will feed into one another.
Layer 2: Commercialization (not production creation)
A significant shift in the traditional value chain occurs here, as commercialization leapfrogs product creation, at least up to a certain degree.
With generative AI, the market validation phase can inform and precede manufacturing. This is what on-demand producers like Shein are doing, but this can also be accessible to smaller brands if they join an ecosystem where speed of delivery is removed from the equation for consumers.
This is a profound change. Moving to a pre-order and on-demand production model, effectively mitigates the financial risks associated with inventory and unsold stock. Commercialization becomes an intelligent, data-driven process that actively shapes the products to be created. And the sustainability benefits of effectively removing waste could be profound.
Obviously, customers will not wait forever to receive their products, though. While shoppers are increasingly understanding that fashion cannot be cheap and fast without the planet or another person paying the ultimate cost, fashion is still driven by whims and desires just as much as by practical, long-term planning. For every time someone commissions a piece of on-demand occasionwear for a pre-planned wedding, there will be a scenario where someone needs something much sooner – just because.
So is there anything meaningful that can be done with the now new layer 3, product creation, to accelerate that cycle? Could we reach a point where delivery delay doesn’t exclude such a big market segment that the community-led brand model becomes effectively unviable.
Layer 3: Product creation
Product creation in the community-led framework must be agile, proximate, and proactive.
In this new model, the generative AI concept generation model of the community-led brand leverages a visual database of manufacturers’ design capabilities to arrive at something producible. When creators input their concepts, the AI analyzes this visual history and steers the design towards the nuances and capabilities of the most fitting manufacturing partner – covertly matching inspirations with components, materials, and machinery.
This AI-native system wouldn’t be about pondering variables such as cost or timing, which are where the traditional fashion value chain spends the most time circling back on itself; instead, it’s a direct conduit between the creative concept and the manufacturer’s proven expertise, ensuring a smooth transition from idea to physical product. As the AI is exposed to more historical data, patterns, operations, components and so on,its ability to match designs with manufacturers becomes even more precise, accelerating the production process without compromising on the manufacturer’s strengths or the creator’s vision – and without needing anyone to completely re-engineer a basic t-shirt.
Just like ChatGPT, Perplexity, and others are opening the era of “answer engines” in opposition to “search engines”, the community-led brand’s ever evolving network of manufacturers can create the foundation of a new “answer-machine” when it comes to handling how to transform a given design into a fully fledged product.
To put it simply: AI-native fashion could get us to a stage where we don’t need to ask so many repetitive, rhetorical questions along the road from idea to finished product. And if you don’t need to ask them, how many designers, creators, and brands would choose to?
Conclusion
Through the lens of generative AI, each layer of a brand’s value chain could be imbued with new capabilities. Inspiration is no longer confined to a select few; commercialization shifts upstream, influencing what is produced; and product creation becomes a dynamic and demand-driven process.
A lot needs to be achieved to get there, but I believe the evidence is out there, in both our personal and professional lives, to suggest that this is not a wild idea. I see it as feasible and desirable to reconfigure the entire value chain around these possibilities, heralding a new era where community, collaboration, and creativity are at the heart of fashion innovation. More importantly, this model could open a new path for aspiring brand owners to tip their toes in the world of entrepreneurship while reducing their risk and need to focus on the “boring” parts of having a brand.
If this model sparked your curiosity, you might want to consider downloading Off/Script from the app store. My team and I have been working for a long time on turning this vision into a reality, and we’d love to hear what The Interline’s audience thinks.