AI in Aisle 4: How Artificial Intelligence Can Transform Apparel Retailing

Artificial intelligence (AI) holds the potential to revolutionize retail, and retailers know it. Nearly 30 percent of the top 250 global retailers are already integrating AI into their operations.

That’s according to a new Capgemini study, “Building the Retail Superstar: How Unleashing AI Across Functions Offers a Multibillion-Dollar Opportunity.” The study surveyed 400 retailers, representing 23 percent of global retail revenues, in the United States, Western Europe and India.

Where are these retailers investing? Most are focused on sales and marketing, with efforts similar to the partnership between Guess? Inc. and Alibaba to bring the ecommerce giant’s FashionAI concept to its stores. FashionAI combines smart racks and smart mirrors to help customers mix and match clothing items.

But the AI payoff should extend across the value chain. In fact, the Capgemini study identified $340 billion in opportunities for AI use cases in back-office operations.

With only 1 percent of AI initiatives at full-scale deployment, now is the time for retailers to carefully consider the promise of AI — and what their next moves should be.

Apparel gets smart

Retail was set to spend $4 billion on AI in 2018, IDC predicted, tying banking as the top spender on the emerging technology. While only 4 percent of retailers in the Capgemini study had deployed AI in 2016, 28 percent were rolling out AI two years later.

Which retail segment is leading the way? That would be apparel and footwear, with 33 percent pursuing AI initiatives.


In particular, apparel companies are using AI to improve the customer experience. For example:

·        Luxury brand Luis Vuitton integrated chatbots in Facebook messenger to drive personalized and conversational shopping. The chatbots ask questions and leverage natural language generation to showcase the retailer’s full product line.

·        Stitch Fix, the online personal-shopper subscription service, uses AI to identify trends, understand preferences and develop new styles.

AI in the back office

AI involves a range of technologies. But in retail, two-thirds of AI applications involve machine learning. That makes sense, because machine learning allows retailers to leverage their extensive data sets.

Pet-store chain PetSmart, for instance, uses machine learning to detect fraudulent orders. An algorithm aggregates millions of transactions and their outcomes, including approvals, chargebacks and refunds, and weighs the risk of fraud against the value of each customer. In just 250 milliseconds, it can separate a fraudulent transaction from a legitimate one. The technology saved PetSmart as much as $12 million in 2017.


Likewise, AI has many potential business applications. But in retail, three quarters of AI deployments are customer-facing. That’s also no surprise, because sales and marketing are the most visible retail functions.

The most common implementations involve personalized search, targeted recommendations and chatbots for answering customer queries. For example:

·        Macy’s, John Lewis and Zalando have deployed “find similar” tools that use image recognition to display products in a similar color, pattern, style or shape.

·        Clothing retailer Ted Baker rolled out chatbots to help customers get questions answered, complete direct purchases and track orders.

·        French retailer Auchan delivers AI-driven personalized experiences. Customers participate in a gamified process that leverages their responses to build unique profiles and targeted promotions.

Some of these retailers report measurable positive outcomes from their customer-facing AI. But focusing exclusively on sales and marketing would be a mistake. Why? Because the Capgemini study found that by 2022, retailers could avoid $340 billion in costs by deploying AI across a range of operational activities, from procurement to logistics to returns.


Intelligent use cases

Considering this operational AI promise, Capgemini analyzed 40 use cases across parameters such as feasibility of implementation and benefits expected. It then compiled a shortlist of 19 recommended use cases. Among them are:

·        Assortment rationalization — AI can rationalize SKUs. H&M applied machine learning to search results, purchases, receipts and returns to customize assortments to each store. As a result, it was able to slash SKUs by 40 percent and decrease unsold stock.

·        Fraud reduction — Image recognition can identify counterfeit products. Online sneaker marketplace Goat has resellers submit photos of their merchandise. The company’s deep-learning applications uses heuristics and data points such as color and texture to authenticate the items.

·        Sales support — Machine learning and chatbots can personalize recommendations. Activewear company North Face uses an AI-powered online shopping assistant to ask consumers questions and present merchandise that will best meet their needs.

·        Self-checkout — Image recognition can track when an item is removed from the shelf and the store. Tesco is experimenting with checkout-free stores that automatically debit consumer accounts.

·        Stock replenishment — AI-driven insights from varied data sets can automate stock replenishment. U.K. retailer Morrisons is deploying AI that will analyze internal data such as sales and external data such as weather to predict demand at the store level. During testing, the company reduced shelf gap by 30 percent.

Taking next steps now

Where should apparel companies begin with AI? Follow these five steps:

1. Treat AI as a strategic imperative. Retailers that have successfully scaled AI give it strategic focus. For 100 percent of them, AI is a top-three CEO issue. For 96 percent, implementation budgets are in place.

2. Pursue quick wins. A surprising 89 percent of retailers are currently focused on high-complexity use cases. While less-complex solutions may deliver fewer benefits, they have a greater chance of success, and they’ll give you valuable AI experience. Consider layering AI onto existing assets, such as your e-commerce website or your fulfillment route planning.

3. Invest at the enterprise level. At the same time, AI can require significant strategic investments. One-half of retail leaders invest 5 percent to 10 percent of their IT budgets in AI. Only 2 percent of their middle-of-the-road counterparts invest at a similar level.

4. Focus on data. Successful AI solutions require not just AI expertise, but also mature data management. Develop the maturity of your data practices, such as augmenting internal information with external data streams.

5. Look through the eyes of the consumer. Another recent Capgemini study, “The Secret to Winning Customers’ Hearts With Artificial Intelligence,” explores why some customer-facing AI projects have failed. The primary culprit? Most companies prioritize implementation costs and ROI. Successful customer-facing AI initiatives focus on solving customer problems and delivering superior customer experiences.

AI is sending shock waves throughout retail. Apparel companies can be toppled by those tremors. Or they can be the earth shakers, leveraging AI to disrupt their markets for competitive advantage. The key is to explore how AI can transform not only sales and marketing, but every aspect of your retail operations.

John Hoeller, Jr. is the Lead for Apparel and Fashion in North America and a client partner at Capgemini, a global leader in consulting, technology services and digital transformation.