We provide the only AI-based platform for denim design
KolAi is a denim design company that uses artificial intelligence to create laser designs for denim garments. Its software, KolAi Designer, can create a laser design in seconds, which can then be customised and refined by a laser designer. This saves time and money for denim manufacturers and allows them to create more unique and high-quality garments. In an interview with Fibre2Fashion, CTO and Founder Jack Payne talks about the role of AI in the denim industry, its impact on workforce, potential for improving sustainability, and the future of denim design and manufacturing with AI and automation.
How does the integration of AI technology transform the manufacturing process in the denim industry?
I know there are some huge changes with using AI to reduce waste in fabric cutting, for example. But our focus at KolAi is more on the creative and design side of things.
What are the key applications of AI in the denim industry, such as in design, production, and quality control?
You hit the major applications in the question: design, production, and quality control. Of the three, Quality control with machine vision systems has been developed the most (e.g., detecting fabric imperfections), but we are most excited about AI-assisted design because it has only recently been made possible in a meaningful way.
For example, AI-Assisted Digital Design. Soon, denim designers will be able to input the kind of information they might include on a mood board (inspiration imagery, sketches, fabric swatches, trims etc.) and an AI will be able to consolidate this information and generate realistic imagery of what that the resulting garment would look like. Designers will then be able to interact with this imagery using prompts (e.g., circling the pocket and instructing the AI to change its style). This will enable a lot more digital design and less sample-making.
Then, we have AI-Assisted Physical Garment Creation. Though turning a digital design into a physical garment is the hard part, AI can help with this too. For example, determining wash recipes or dry process laser files for digital garments.
What will be the impact of AI-driven automation on the workforce in the denim industry, and what are the potential implications for job creation and displacement?
It really depends on how it gets implemented in the industry. There is an understandable concern that AI will be taking jobs. In most cases, there will still be a need for human input. Even the best current AI systems require prompts and some editing from a real person. As the industry changes, there will be new roles that emerge, and as we adapt to new tools (because essentially AI is a tool) people will find new ways to use and build upon those tools.
I think a good analogy is the adoption of laser technology for dry processing. Before lasers, it was someone’s job to create physical dry process templates and someone else’s job to sand garments by hand using those templates. With lasers, you no longer need staff to make physical templates or to do manual sanding. However, you now need staff to make digital templates (i.e., laser files) and operate the laser machines. The nature of the work has changed but you still need people. Admittedly, you may need less staff than before because of increased efficiency, but the staff you do have are also doing safer work (fewer chemicals, less ingestion of airborne indigo etc.)
How are AI-powered analytics being used to optimise supply chain and inventory management in the denim sector?
Better analytics can help find weak points in supply chains, find better routes of production, and help reduce waste. They work by taking in vast amounts of data and finding connections or anomalies within that data. For example, the system might notice that there’s a fairly consistent 2-day delay in a certain shipping lane, and therefore it would recommend a different lane or perhaps a different manufacturing endpoint to reduce that time delay. Because they can take in such massive amounts of data, a well-designed AI can find things that many people would otherwise miss. Of course, there are cases where the opposite is true as well.
To what extent can AI contribute to the sustainability and eco-friendliness of denim production, and how might these advancements further develop in the future?
AI will help bridge the communication gap between the way designers envision their product and the way it must be described for manufacturing purposes. Thus, there will be less reliance on physical sampling in the future which means less fabric, water and chemicals used.
The ideal scenario for sustainability is a digital-to-production pipeline where designers can do all or most of their prototyping digitally. AI can help ensure the digital visuals can be matched by physical garments (by calculating washing formulas etc.). This scenario also allows for wider adoption of customisation / customer-created products (why not? If what you see on the screen can truly be matched), which means less dead stock.
Traditional methods of denim production use a massive amount of water and other chemicals. The use of laser finishing reduces some of that water consumption. Our system allows more laundries to use laser effectively, by making it easier for anyone to create a good-looking design. It also reduces the need for as many cycles of prototyping, so instead of having to send 5 samples back and forth across the entire planet, laundries and brands can get a high-quality finish much more quickly, saving time and fuel.
What was the inspiration behind starting KolAi Denim, and how did you identify the need for AI in the denim industry?
KolAi Denim started as an idea.
While working for VF Corporation (Lee / Wrangler) our CEO Kyle became familiar with laser finishing of denim and was at the same time keeping up with the academic literature in machine learning and artificial intelligence. Kyle realised that new machine learning models could be used to help automate laser design for denim. He presented this idea to a colleague and denim laser expert Aydin (at that time he was doing laser and wash development for 7 For All Mankind which was a VF brand). He thought the idea was fantastic and never forgot about it.
Several years later, after the COVID-19 pandemic broke out, Aydin reached out to Kyle. “Remember that idea about automating laser design?” He said, “Let’s do it.” And that’s how KolAi was born.
Kyle brought me in to help with the technical/programming side of things because I had built up a skillset using neural networks to perform data analysis during my doctoral work and afterwards in the industry. Aydin provided the data, and together we built software that speeds up laser design for denim. We incorporated in 2021 and called our company KolAi from ‘Kolay’ the Turkish word for easy + AI for artificial intelligence. I’m a big fan of multi-lingual wordplay like that.
Can you share the journey of developing KolAi’s AI platform from the initial concept to the current version?
It’s been quite a journey indeed! We started with the core idea of creating an artificial intelligence to translate an image of a pair of jeans into a usable laser design. One of our founders, Aydin, is one of the most experienced laser designers in the world (25+ years as a laser designer!). I spent about a year developing the central AI system before we were able to generate results that Aydin considered good enough for the industry! He has very high standards for design, one of the many reasons he’s been at the top of the industry for so long. So that was a huge milestone; it was a culmination of countless hours of data work, programming, and troubleshooting.
Once we got the core product ready, we started talking to more people about our product. Our early customers gave us a lot of great feedback for what they wanted out of an ideal denim laser design system. We took that feedback to heart, and created a suite of top-quality photo editing tools so that our customers can do all the tasks of the laser design pipeline in one spot. That includes working with layers, complex denim-specific automations, applying filters, drawing and painting, applying text, all-over prints, and more.
Recently, we made another huge update based on customer feedback, where rather than giving the designer a single AI-generated layer, we built several networks that give multiple layers. The AI now instantly generates separate layers for whiskers, thigh abrasion, and texture. This allows for easier customisation and a more natural 3D-looking design.
How do you envision the future of denim design and manufacturing with the increasing integration of AI and automation?
I see the process getting smoother, faster, and more personal. We envision a future that democratises the process, allowing consumers to have more say in how their clothes look, feel, and even how they’re made. I would like to see a bigger focus on sustainability in the denim world. AI will also help bridge the communication gap between the way designers envision their product and the way it must be described for manufacturing purposes. Thus, there will be less reliance on physical sampling in the future which promotes sustainability.
What sets KolAi apart from other AI platforms or solutions in the denim industry?
We are focused specifically on design tools. As far as I know, there are no other AI-based platforms for denim design. So far, we have built a tool specifically for denim laser design and we provide the only AI-based platform in this space. We are always taking in new customer feedback to build more tools and systems that people want, and because we are a relatively small team, we can move quickly to make those changes.
How do you address concerns about AI potentially replacing human designers in the industry?
That’s an excellent question, and one that every AI company should have to consider when they make their products. Because our system allows designers to get further along in the process more quickly, they can spend time applying their own artistic flourish to the designs they create. Rather than spending hours getting a basic design ready, they can spend that time refining and customising a design. Every new software tool that’s been made over the years helps people work faster, and the ideal is that they empower people to be more creative!
How have clients responded to using your platform, and what kind of feedback have you received from designers and manufacturers?
People have been enthusiastic! There are many small laundries in the world. Many of them only have a part-time designer, or minimal laser experience, maybe they have a single laser and no one who knows how to get good results out of it. Our system is great for that, because even someone like me, who didn’t know much about denim laser design, can make a decent design with minimal experience. I just recently taught a student from the Fashion Institute of Technology in New York how to use it, and she got good results on the first try! That was really cool to see.
Advanced laser designers appreciate the system as well because it takes care of a lot of the tedious busy work they would have to do just to get a base for their design. Most laser designers are under tremendous pressure because they have to make a lot of designs in a short period of time, and they have to be good. With KolAi, they can now focus their energy on just refining and customising a design using their own expert eye. In the end, they end up with a better result in a fraction of the time and their sanity still intact.
Our customers are always looking for more ways to save time in the process too, so they often request feature upgrades; things like a way to take an image of an all-over print, (for example, leopard spots) and make that into a usable laser design. The multi-layer output was another thing that a lot of people wanted to see, so we’re very proud to say that we listen to what our customers want!
How do you ensure that your AI platform stays up-to-date with the latest design trends and industry best practices?
The AI platform is a tool for denim design, meaning the designer who is using the platform is ultimately in control of the latest design trends because they are providing the inputs. This is true with any design tool (Photoshop, for example).
However, we take feature requests from our customers regularly (our most recent update, multi-layer laser designs, was our most requested feature from real-world users!) and we keep open lines of communication with brands, denim designers, laser manufacturers, laundries and textile fashion publications. We also go to a number of trade shows to make sure that we’re collaborating, sharing and learning with as many people in the industry as we can. Connecting with people is the best way to remain up-to-date!
What are the challenges you have faced while developing and implementing the KolAi platform, and how did you overcome them?
When we first started, we knew an AI system that makes laser designs would be helpful, but we didn’t know what else was needed or the best way to sell it. We figured this out through trial and error and by listening to customer feedback.
The first version of the system was just the AI and nothing else. But users said it wasn’t enough – they needed the ability to refine and customise the laser design without leaving our platform. So, we added Photoshop-like image editing tools around the AI. Then users said that was great, but it would really be helpful if the design was separated into layers. So, we built that in with our latest release, but it won’t end there! Software development is an ongoing, iterative process – there are always things to improve on either the front or back end.
Can you discuss any collaborations or partnerships that have been instrumental in the growth of KolAi Denim?
Not by name, but we have been lucky to work with several large brands the last couple of years. The brands have always given us great feedback for things to iterate on. We’ve got some more partnerships on the way this year too!
Are there any plans to expand your platform beyond denim, such as into other segments of the fashion industry?
Our current system is denim specific, but we have thought wider too! In our patent application, we’ve outlined a number of other AI systems that we could build to help fashion designers create rapid prototypes. The idea is a digital AI assistant for fashion design. You tell it to change a patch pocket to a well pocket or make the leg panel wider and it updates your digital model accordingly.
20230103183907.png)
