In a follow-up to our AI Report 2024, The Interline sat down with Isaac Korn, Senior Director of Innovation for Perry Ellis International, to talk about how the introduction of AI differs from other digital transformation strategies – and how technology and culture need to be sensitively weighted if artificial intelligence initiatives are going to achieve their aims.
The Interline: In your role as a lead innovator for a brand, you’ve spent a lot of time working in 3D / digital product creation (DPC), which itself has evolved a fair amount over the last few years. Now that we’re firmly into the AI era, tell us about what your day-to-day looks like Perry Ellis International and how that’s changed over time?
Isaac Korn: Perry Ellis International has been at the forefront of technology for the past 20+ years – using predictive analytics, business intelligence, and now AI/machine learning to drive innovation in fashion. Day to day we continue to be focused on DPC. We have made great progess as we implemented our Design 3D strategy early on which allows us to conduct business in a sustainable fashion with cost efficiencies. Alongside that, my main role is looking for new technologies, understanding the trends in the market, and from there considering how they might service the needs of the company and streamline our processes. I do this by running proof of concepts (POCs) and different tests to see how those technologies could improve the way that we work. Additionally, AI tools have helped with streamlining our processes tremendously.
The Interline: Unlike DPC (where the initial push was spearheaded by the biggest brands, with the biggest budgets) this time around, the core AI technologies are open and available to everyone. Perry Ellis International is an interesting proposition, too, because as an agile brand you have a history of going ahead and pushing practical applications in several cutting-edge technology areas over the last few years. How, in your opinion, is the progress you’re making now with AI different from the notable work you did in NFTs and the metaverse? And do you think brands should be thinking differently about this compared to those hype cycles?
Isaac Korn: I believe that brands should start thinking about this AI era totally differently because, this time around, the technologies have already measurably changed the way we think and work. It’s also growing at a speed that is markedly faster than before. This time it isn’t a case of waiting and seeing what others are doing and then deciding whether to do it or not. We saw a lot of hype around NFTs and the metaverse, but I think a lot of that was centred around circumstance. At the time that they became popular we were firmly in the COVID-19 pandemic and fashion businesses needed to generate something of value. If generative AI had been around at that time,I believe it would have had a very positive outcome. The number of people already using generative AI now is major evidence of the fact that it is here to stay, and will have a significant impact. Unlike the metaverse and NFTs, the barrier to entry is much lower, and it’s not the case that only people who understand Web3 and cryptocurrencies can get involved – there’s very little gatekeeping in AI.
We also have a different mentality at Perry Ellis International that we’re very proud of and pride ourselves in being innovative. We have always been early to start testing new technologies and start creating different use cases internally. And in my position at the company, I have been able to drive that value-added testing to understand what technologies are worth investing in and will bring tangible benefits to how we work… and that’s how I see generative AI. The same tools are now accessible to brands and people of all different shapes and sizes. It’s not like the traditional enterprise technology cycle where bigger brands with bigger budgets start the usage, and then slowly the capabilities filter down to smaller organisations. Now, everyone has access to the same tech at the same time. I think that’s what makes it so interesting and engaging.
The Interline: The next logical question is what stage you think the fashion industry is at on its AI journey. Are we in the stage where ready-to-use, off-the-shelf fashion specific AI tools are actively being used? Or are we still at the stage where the big, frontier cloud hosted industry-agnostic models are primarily being used, and people are still working out the shape of the scope of what fashion-focused applications should look like? Where do you think we are on that journey and how should brands be thinking about that?
Isaac Korn: The technology available today can be used in so many shapes and forms. In some instances it can be applied in many use cases for fashion, but at the same time it is not quite where it needs to be yet, industry-wide.
When it comes to data analytics and using that to make decisions, there are already great AI tools available. AI is very good at consuming a lot of data, processing it, and allowing us to glean insights from it. The fashion industry utilises tremendous amounts of data in the form of sales and cue counts etc. companies just need to ensure that they have good quality data. It could be that it isn’t all there or it is inaccurate. But if you have good quality data as a brand or retailer – there are many benefits to leveraging AI tools.
The limitation comes in when it comes to using AI tools for generating content that is brand specific and of a high quality. It takes a lot of time to train a model to give an output that isn’t generic. So that’s why, as of today, generative AI is not going to replace designers or creative people. It’s not yet able to design an on-brand collection for spring/summer 2025, and provide us with the different specs and so on. It’s currently a very good assistant or co-pilot to those in creative positions. This goes for assisting in productivity and automation of mundane tasks, and also for ideas and inspiration. We’re doing a lot of internal experimentation with generative AI and fine tuning it. We’re working towards that brand-specific output, but it’s still a work in progress.
But soon there is going to be a company that is going to apply all of AI’s lessons and make a general purpose tool, specific to fashion, that is going to assist with workflow, understands a brand’s specs and guidelines, and wider market trends. We aren’t the only ones thinking about this kind of AI tool for fashion that helps to create a very accurate line plan. This plan will detail what the next season’s collection should look like, and the benefits will be better sales conversions and less waste. We have always had the data, but now we have new capabilities and new functionality. It’s a case of bringing all of that together synthesising those sorts of insights.
The Interline: Where do you see generative AI adding the most value in the short and long term? Or is it impossible to predict?
Isaac Korn: Short term, there are small wins that everyone can take when it comes to automating mundane tasks, and also when one needs to get the juices flowing for creativity. It can help with small tasks like generating emails or work in Excel, for example. This doesn’t mean that it’s taking responsibilities away – it is actually just saving you time on tasks that you, as a creative person, execute, but that are necessary as part of your role. Fashion can be glamorous and creative, but the reality is that there are so many repetitive, administrative tasks that need to be done every day, on repeat. Longer term, it’s impossible to say because things are changing so fast.
The Interline: Turning now to the non-generative side of AI: is there anything you’ve been working on not thinking about that fits into this category?
Isaac Korn: We’ve long used AI-driven tools in e-commerce, such as computer vision for generating tags and enabling smart search. Many of these tools, used for years, are AI-based, though not everyone realises it. Teams rely on AI for planning, aiding them in understanding specifics and forecasting needs like quantities to purchase. The extensive analysis and data synthesis behind the scenes are managed by AI tools, which present the necessary information in a way that facilitates informed decision-making.
The Interline: Casting your mind forward, what are other brands not preparing for that they should be?
Isaac Korn: The area of copyright infringement and generative AI. I think all brands should be aware of the risks and be very careful with how they proceed with how they train their models if the information used is open source. I think that the fashion industry is still acting relatively conservatively when it comes to AI because the legal constraints when it comes to generative AI are still a bit murky. On our end, we are focusing on internal testing for now and are being very mindful of the legal implications of what we do because sooner than later the law is going to be black and white on what can be used and what cannot. I think all brands should take this seriously and not use AI just for the sake of it.
The Interline: It’s highly likely that we are all going to be surprised by the capabilities of AI again in the near future. What do you think is next? And how are you positioning Perry Ellis International to deal with it?
Isaac Korn: I think that we’re going to be able to streamline the process of generating a collection that is in line with brand guidelines, and what is being created is actually going to be needed in the market. That way, waste will be limited and efficiency will be at its peak. With the use of data and AI tools we are going to be able to take away a lot of the guesswork that comes from creating collections. Creativity will still be front and centre, but the entire process will be better automated, more sustainable, and needed.
