Original article was published on artificial intelligence
- The $2.5 trillion fashion industry seems vulnerable to a long drawn out recovery, due to the discretionary nature of consumer spending on clothes.
- Retailers will now experience first-hand, the need for a digital-first mentality to even stay afloat.
- Empowered by AI insights, designers, merchandisers and buyers could soon predict customer preferences, way before they become fashionable.
Even before the coronavirus pandemic crashed the global economy, disrupting supply chains and stalling consumer demand, the
industry was not looking forward to walking the 2020 ramp. Headwinds such as slowing demand, rising competition, shifting demographic-driven preferences and the rising carbon footprint of fast fashion, were already hanging over the sector’s head.
The fallout of COVID-19 pandemic is beginning to look like the biggest economic shock in over half a century, the 2008 financial crisis included. While many industries have already been hard hit,
seems vulnerable to a long drawn out recovery, due to the discretionary nature of consumer spending on clothes. Meanwhile, deep discounting stock clearance, supply chain interruptions and even bankruptcies, are the themes playing out in the short to medium term.
The reshaping of the prevailing environment and business landscape will also present an existential opportunity for the industry to reinvent itself. Innovation and digital acceleration will be at the middle of all future-proofing initiatives, tying to already accelerate consumer trends like personalisation and customer experience. A low contact economy is presently emerging from the ashes of the pandemic, but retailers will now experience first-hand, the need for a digital-first mentality to even stay afloat.
If digital technologies like
(AI) had its way, the much-dreaded practise of discounting in fashion and retail, could vanish overnight. Trained algorithms would pre-empt such situations by informing retailers on which products to procure, in what quantity, whom to sell them to, and when – at full price. Empowered by AI insights, designers, merchandisers and buyers could soon predict customer preferences, way before they become fashionable. Online shopping could become a seamless experience in engagement, where the shopper describes the dress of their dreams, and an AI-powered search engine responds by tracking it down.
In short supply – fashion intelligence
While demand side use cases are the glamorous poster children for AI in fashion, the behind the scenes reimagining of the supply side, is where the tech journey begins. Currently, fashion retailers work with limited data to predict what products to make, in what quantity and when to discount them for clearance. Data brought together from disparate sources can help brands to align supply with demand – trends recognition on social media, including colours and silhouettes, identification of keywords from text searches, voice assistants and chatbots, and the browsing and shopping history of customers. AI’s ability to make predictions on trends even before they happen, is the holy grail in a trend-driven industry like fashion.
Skip to production. Robotic automation of manufacturing, leveraging computer vision, industrial robotics and machine learning, boosts production speed several fold. It also eliminates human production related issues like error, fatigue, safety and hazard constraints, cycle time bottlenecks and labour costs.
The next stop in the supply side journey is pricing decisions. Big data insights and AI-enabled real-time pricing as well as predictive pricing, compare exact and similar product matches across the competition. Market-based demand assortment helps in stocking products which sell more, thus contributing more to margins. Finally, AI-based algorithms consider factors such as consumer behaviour, seasonality, likely demand, and expected discounts to formulate optimal prices.
Accurate demand forecasting can make the difference between excess stock that locks up capital, consumes warehouse space, and is eventually sold at markdowns, missed sales due to understocking, or hitting bulls eye. Machine learning solutions greatly reduce forecasting errors. AI-based management systems recommend optimal stock levels for thousands of SKUs at distribution centres, while automated sensors and robots issue repurchase orders when inventory drops, to restock in-demand products. A good example is Hitachi using AI to issue instructions to warehouse employees to continuously improve on the operations of supply chain to respond quickly to the market condition.
Consumer satisfaction, on demand
If the supply side is all about operational efficiencies and number crunching, the demand side unleashes the sexy in AI. AI-powered visual search engines empower image search, or even with patterns and shades, returning with the closest matches from catalogues. When integrated with mobile phone apps, search can be activated via phone cameras too. AI can analyse the browsing and shopping history of customers, to understand and make intelligent recommendations when they visit physical stores. Virtual personal stylists help customers choose the right styles and fits that best suit one’s body structure.
Virtual trial rooms are the latest addition to fashion retail from the AI armoury. They simulate the trial room experience on a large screen, using computer vision-enabled cameras that project the shopper’s shape and size as a virtual mannequin. The shopper can see how the clothes fit, with additional recommendations that include matching accessories to complement a look. As part of Walmart’s Innov8 VR competition, Obsess created a photorealistic CG virtual store for Rebecca Minkoff. The experience included the ability to shop racks or runway using a headset and complete the checkout process in VR.
Virtual assistants and chatbots behave as shopping assistants, making personalized recommendations via chat. Apart from responding directly to queries, suggesting products and assisting with right fits, chatbots route customers to sales representatives and fashion designers, when human intervention is required. Self-learning algorithms improve the relevance of recommendations. These assistants also suggest similar items to the product that is being examined, helping prevent exits from out of stock situations or when expectations aren’t met.
Retailers lose about
in preventable returns annually, while also losing customers in the process. AI can help retailers personalize shopping experience, leading shoppers to make informed purchase decisions based on previous purchases, sizes and returns. For instance, Zara has evolved its size chart and suggests sizes not only based on customer inputs but also by tracking their preferences (example: loose or tighter clothing) and the return of that garment by other users with same tastes and measurements.