Data & Analytics

From science fiction to business reality: Making AI work for your fashion supplier to retail business


This time last year, the fashion industry was grappling with a severe supply chain crisis, dealing with deadstock taking up warehouse space, coming to terms with shifting consumer behavior and fluctuating demand. A year later, little has changed. While the fashion industry faces different iterations of the same challenges every year, technology is rapidly evolving. The advent of ChatGPT, new breed of how Large Language Models (LLMs), and generative AI technologies have awakened the world to the transformative potential of artificial intelligence. While AI itself is not particularly new, business leaders are starting to recognize the technology’s fashion industry potential.

The third edition of the Fashion, Technology & Wine event series sought to explore exactly how fashion brands can reimagine their value chain with AI and ML technologies. Held last month in New York, the exclusive forum organized by Infor, in collaboration with Fortude, opened up with an engaging session featuring Infor’s Sandeep Anand who called for fashion brands to ‘Think Big, Work Smart’ with AI and ML. Diving deeper into this area of innovation, this blog explores some of the most prominent use cases for AI technologies within the fashion value chain and how business leaders can adopt them and attempt to stay ahead of the competition.

Artificial Intelligence and Machine Learning: Making sense of it all

As technologies like ChatGPT take over the news cycle, you see the terms artificial intelligence (AI) and machine learning (ML) thrown around and often used interchangeably. As fashion brands look to tap into the benefits these technologies bring, we must take a step back and deepen our understanding of the distinctive differences between each of these terms. Artificial intelligence (AI) is an umbrella term that refers to the type of technology that mimics human intelligence and human cognitive functions like problem-solving and learning. While early AI tech was rules-based, limited in its scope of capabilities and often relied on significant human input, modern AI learns from historical data to optimize and solve complex tasks that were previously only performed by humans, such as facial and speech recognition, decision making and translation.

Machine learning (ML), on the other hand, is a branch of AI that enables your system to optimize itself by tapping into historical and current data (remember what happened yesterday is now part of history) to make more accurate predictions and minimize errors. Business leaders can leverage this in two ways. First, ML will be used to augment analytics, solutions, and draw out connections between data points and detect recurring patterns, anomalies, or cause-effect relations, which is particularly useful in trendspotting and demand forecasting. Second, businesses can leverage the pattern recognition capabilities of ML to replicate humans’ cognitive skills. In practice this could be used to “learn” the linguistic patterns of human communication to understand and mimic them or detect visual patterns and relate them to pre-defined objects.

The fashion industry’s reliance on cycles and trends makes it the perfect candidate for innovation with AI. The industry’s recurring struggle with unsold inventory can also be addressed by AI—from production to shipping, AI can help fashion leaders optimize every aspect of the product lifecycle.

New product development with a (smarter) twist

Trend forecasting is a key part of the product development process. This laborious process often involves scouring through trend reports, market analysis, social media data and consumer sentiments to inform designs for the next season’s collection. Fashion brands can tap into AI to collect and analyze all these data sources to support the product ideation process. AI can work alongside human designers to draw inspiration from past product lines, trending styles and social media sentiments and convert these into new product designs and variations. This data also allows brands to decide how much of each design needs to be produced to keep in line with the predicted demand.

It doesn’t end there. AI can help combat fashion retail’s infamous returns problem by helping brands sell it right the first time. Given that poor fit is the leading cause of returns, AI can be used to customize products for individual consumers at scale and meet their demand for personalization. Fashion businesses can also use AI-powered virtual sizing tools to customize products for individual consumers at scale.

Redesigning the fashion supply chain with AI and ML

  • Optimizing inventory

The supply chain disruptions of the last two years have put the fashion industry through the wringer – from brands first struggling to get their hands on enough inventory and then overbuying to compensate, and now struggling to clear out excess stock amid an economic downturn. Now brands stuck with inventory overhang are slashing prices to get rid of the excess stock before the products go out of style. For fashion manufacturers looking to optimize their inventory with maximum selling potential while reducing the risk of overproduction, AI can be truly revolutionary. Manufacturers can tap into AI to better monitor their inventory levels and stockouts via real-time inventory tracking and predict demand more accurately by analyzing sales data and trends so they can adjust their inventory levels and production schedules accordingly. This data can then be fed back and run through ML algorithms to identify popular products and forecast demand patterns. Based on this information, the company can optimize production processes, avoid overstocking and understocking, and reduce waste to ensure it is meeting its sustainability commitments.

  • Getting it right on the factory floor and warehouse

Venturing beyond demand planning/forecasting and inventory management, manufacturers can also leverage AI and ML to optimize the supply chain. AI tech can help analyze the mountains of supplier-related data sitting within your business to ensure the right materials are delivered at the right time. In the warehouse operations space, AI and ML can be used to ensure that warehouses carry the right amount of stock at all times and automatically replenish inventory when it runs out. When AI/ML models are deployed in conjunction with automation solutions, it enables the production line to benefit from predictive maintenance. Fashion manufacturers can ensure that any potential issues are detected early on before they start impacting business continuity. This kind of preventive maintenance powered by AI/ML can boost operational efficiency and reduce downtime for fashion business leaders.

  • Smarter logistics and better quality

What if you could automate your quality control and defect detection processes? Manufacturers can implement deep-learning algorithms in the production line by integrating these capabilities with their manufacturing execution systems and using smart cameras. This can be used to spot defects in fabric and ensure that the colors of the finished product match the original design. AI-powered automated quality inspection minimizes the risk of costly defects, product recalls and raw material wastage by checking each product for surface defects efficiently, resulting in quicker time-to-shelf and high production quality. In the logistics space, AI and ML-augmented tech can also help fashion brands optimize delivery routes, reduce transportation costs, and improve delivery times by analyzing shipping and transportation data. This in turn can inform anticipatory shipping and smart route planning activities.

AI-powered fashion retail for smarter shopping, online and offline

  • A better product discovery experience

Consumers today are more accustomed than ever to personalized content—this expectation stands true for their online shopping experiences. Shoppers expect brands to offer them an individualized experience on their site, including in the form of personalized product recommendations and virtual assistants. When they visit your e-commerce site, they want to find their desired product at the right price, in the right size, color, style and material, all in the shortest time. This is where automated product tagging comes in. Retailers can adopt AI-powered product tags to enrich each product in their catalogue by automatically adding attribute labels related to color, fit, fabric, prints, sleeve length, necklines and more.

Ever tried to search for a product on a website only to end up with a bunch of unrelated results? With AI-automated product tags you can find the exact ‘knee-length, three-quarter sleeve, white dress’ without having to sift through pages of random products. Not only do these accurate, relevant product search results bring the customer closer to a buying decision and prevent them from leaving the site without making purchase, but it also helps retailers understand which products are performing well at an attribute level. This real-time feedback loop and the insights derived from it can inform decision-making in every aspect of the fashion value chain.

  • Powering personalization

AI and ML can also support another aspect of product discovery—personalized product recommendations. When customers click on a product they are interested, AI can be used to show visually similar and relevant products, suggestions on how the selected piece can be styled with other products based on their purchase history or products they have previously expressed interest in. Not only does this prevent the customer from dropping off if they don’t like the item they clicked on or if it’s out of stock, but also helps increase their overall basket size. For customers that may struggle to put into words what they want, AI-enabled visual search can also be an option on the table.

Another area of opportunity for fashion brands to leverage AI is price optimization. AI can be used to analyze pricing data, competitor prices, and consumer behavior to inform your dynamic pricing strategy. On an ecommerce site, AI can offer personalized pricing – by analyzing competitor pricing and data related to products that shoppers click on/add to their cart/purchase, brands can entice first-time buyers and price-sensitive customers with better deals.

  • Bringing AI to the physical realm

AI and ML doesn’t have to be relegated to the digital realm. In physical stores, AI-powered smart mirrors can recommend products similar to the style of clothing that the customer is wearing at the moment or offer styling suggestions with other products in the store. On the e-commerce front, this could take the form of virtual product fitting and demos. For brick-and-mortar store operations, AI and ML can also be used to optimize store layouts by generating and testing layout plans under different factors such as customer footfall, store size and local consumer base.

Making AI and ML a business reality: Where to start

Fashion business leaders should identify where AI and ML can offer the greatest value to their business—whether its design, manufacturing, or retail. To start off they can look within their business to understand their biggest pain points, what external pressures lie on the horizon, and how their competition is taking advantage of advancements in AI. Leaders can then prioritize the AI use cases based on the level of impact the identified opportunities will have on their business. As Infor’s Sandeep Anand notes, even a single high-impact use case can transform your operations, and some of them can bring significant results for your business in as little as 90 days.

Once you identify the top priority use cases that fit your broader business strategy and goals, you would have to take a look at the feasibility of implementation. This includes ensuring that you not only have the right technology architecture in place, but also the right talent and culture. A short- and long-term AI adoption roadmap can help address any barriers to adoption such as technical skills gaps or the lack of a robust data analytics foundation. Keep in mind that the accuracy and usefulness of AI is only as good as the data it has to work with so it is crucial to build a solid data analytics foundation to capture the full value of AI. Once you overcome these barriers, you would also need to build in appropriate governance and control structures to mitigate any risks associated with deploying AI in your business.

Want to be notified for the next edition of the Fashion, Technology & Wine event series? Register your interest here:

Robert McKee
Chief Strategy Officer, Fortude