Artificial Intelligence's Impact On the Future of Fashion

Competitive advantage increasingly relies both on anticipating customer preferences and taking costs out of the supply chain in order to deliver better selling products faster.

AI offers that opportunity. JL

Hilary Milnes reports in Glossy:

Using AI to track customer behavior, high- and low-performing products, popular silhouettes and color patterns to predict what new categories and pieces will sell.“The smarter we get with AI, the longer our customer stays with us (and) the better we get at improving product.” Dig(ging) through customer data to find insight in product development faster (creates)more options for high-performing products. Its inventory strategy reduces waste.

Like many one-note fashion brands before it, luxury lingerie brand Cosabella wants to become a lifestyle brand. But as Cosabella moves into categories like swimwear, sleepwear, ready-to-wear apparel and athletic apparel, it’s looking past mood boards and runways for inspiration.
Cosabella is using artificial intelligence and machine learning to track customer behavior, high- and low-performing products, and popular silhouettes and color patterns to predict what new categories and pieces will sell.
“The smarter we get with AI, the longer our customer stays with us. The longer a customer stays with us, the better we get at improving product, fit, fabric and silhouette,” said Cosabella CEO Guido Campello.

Cosabella, which sells its items globally through its own channels as well as through wholesale partners like Nordstrom and Bloomingdale’s, operates a 100-person team. With already tight resources, Campello found it was impossible to dig through customer data in order to find relevant insight in product development to figure out what their customers were missing.
So last summer, the company hired automated marketing platform Emarsys and handed over its bulk of data to artificial intelligence.
“Everything is AI”
Cosabella has since gotten faster at rolling out more options for high-performing products and tweaking ones that aren’t selling. Its inventory strategy reduces waste and limits out-of-stock disappointments. What’s more, it has introduced a capsule collection based on what customers are wearing and buying in real time, which is designed and put on sale at a faster pace than its regular collections.
Cosabella also uses AI to tailor its email offers, place ads, improve product recommendations and change the way the e-commerce site is laid out depending on how someone arrived at the page.
While AI isn’t a magic fix-all for fashion brands, it’s taking a long time for it to be adopted. The reason: Not many people understand it.
“The term ‘AI’ has been co-opted,” said Allen Nance, Emarsys’s CMO. “Everything is AI, but no one knows what it actually can do. And the best part is that, even though data conversations scare people, expectations are sky-high.”
Artificial intelligence in fashion is stuck in the same catch-22 as technology like virtual and augmented reality. They’re all consistently on the cusp of disrupting the industry, but as a majority of cautious retailers wait in the wings to watch as others test it out, no real progress is being made. The edge AI has over VR is that it can actually make an impact on how companies operate internally.
“There is always hesitation around new technology. Brands don’t want to invest without seeing how it plays out in the market,” said Mark Jarecke, founder and creative director of agency Four32C and former creative director of Condé Nast Digital. “With AI, there is a concern around the lack of experience and expertise, but it feels like we’re at the point where the locomotive is just gaining speed, and soon it will be everywhere.”
AI as marketing
In 2015, IBM Watson and The North Face teamed up to build an online “personal shopper” using artificial intelligence. Users shopping for a new coat were prompted to enter information — like their climate, what the coat would be worn for and their location — and the AI bot would spit back a selection of “smart” suggestions, the same coats that a real person in a store might suggest.

The North Face and Watson bot.

This use of artificial intelligence has since become more relevant with the rise of chatbots on Facebook Messenger, which ping responses back to customers who interact with it and ask questions. But customer-facing AI tools, like The North Face’s partnership with Watson and Tommy Hilfiger’s TmyGrl chatbot, are launched mostly for the PR buzz they generate, rather than for the meaningful data a brand can gain from the technology. The reason: Not only do chatbots get more people talking than an internal data management system does, but overwhelmed marketers facing massive amounts of data don’t have the means to move the needle.
“These companies talk about data as the future still because I’m not sure their infrastructure is ready,” said Nadina Guglielmetti, managing director at Huge. “They’re getting a ton of data and don’t know how to filter it to get the right insights. Plus, CMOs are expected to do everything — PR, marketing, data activations — and there’s so much to know that they don’t prioritize the data side, because it’s not where fashion has traditionally played.”
The result is a surface-level understanding of databases.
“Look at email marketing. It’s typically not the CMO building these reach-outs, it’s the junior strategist,” said Emarsys’s Nance. “On the other side of a bad digital message is a piece of software and a 22-year-old dragging and dropping, basically guessing which customer set gets what.”
AI on the horizon
A slew of AI startups have popped up with different solutions for retailers. SupplyAI offers AI and machine learning tools that reduce return rates by making predictions based on prior consumer behavior and product performance. Propulse Analytics, launched by former Saks Fifth Avenue executive Eric Brassard, improves product recommendations online by taking customers’ personal tastes into account. Canadian retailer Frank & Oak reported seeing two times the conversion rate online after using the technology.
“Retailers are still scared to give away access to their data,” said Brassard. “It’s like showing your underwear. But they’re beginning to realize they need better personalization and recommendations, and that they’re not going to become specialists in big data.”
Jarecke expects AI’s power to lie in its subtlety — like more appealing product recommendations — as it comes out of the gates. Consumers won’t be hit in the face with creepy personalization or wooed by the novelty of a Watson bot.
“AI can be used in a subtle way to optimize purchase flow, aid with efficiency and insights and offer more personalized experiences,” said Jarecke. “People often assume AI will just help the consumer. But on the administrative side, they’ll free up staff hours to relocate elsewhere.”
Courtney Connell, Cosabella’s marketing director, said she had a hard time letting go of old mindsets when first working with AI.
“As humans, we want to hold onto as much control as we can,” she said. “But this takes out the hands-on work and Excel spreadsheets that people hate doing. Once the fashion world understands that, we’ll get to a new phase.”