Artificial intelligence has captured the interest of retailers, investors and consumers alike for a while now, thanks to the rise of generative AI making the technology more accessible to the public.
Some believe companies that drag their feet on plugging generative AI into their internal or external processes are already falling behind. As buzzwords fly around and companies try to grasp what emerging technologies mean for their businesses, some have started test-and-learn efforts while others work toward full-scale adoption.
Related Stories
Last week’s annual AI Summit at New York City’s Javits Center attracted more than 3,000 industry-agnostic professionals to the two-day Informa Tech-hosted event. And Coresight Research gathered several hundred retail leaders to discuss the technology in New York City’s Hudson Yards on Tuesday.
Key themes emerged from leaders’ discussions: generative AI remains of interest as internal and external adoption grows; what practitioners previously considered dangerous could be useful for creatives, and the fashion has much to look forward to—and look out for—in the new year.
Enabling user outcomes with AI
For some companies, creating a consumer-facing generative AI solution adds to their value proposition.
EBay has used generative AI to enable better outcomes for its sellers, executives said on an AI Summit panel.
The online marketplace company has leveraged generative AI to help sellers create stronger descriptions for their product listing, said Xiaodi Zhang, the e-commerce giant’s vice president of seller experience.
“It’s easy for sellers to write a title, [or to] put together the item aspects, but when they get to the description, we see a significant drop off and… a slowdown in the funnel,” she said.
Zhang said that while sellers told eBay they wouldn’t have an interest in using generative AI to help their creative process, the feature has become extremely popular among those trying to turn a profit on the site.
“We’ve been very surprised by how much sellers appreciated this feature,” she said. “It’s been the highest-rated—from a satisfaction standpoint—feature over the last few years.”
Zhang and her colleague Nitzan Mekel-Bobrov, eBay’s chief AI officer, said the company has worked to include a wide breadth of employees in AI-based initiatives, a critical part of the journey.
Zhang said Mekel-Bobrov organized an innovation event to drum up new ideas. During the hackathon portion of the week-long event, employees prototyped use cases for existing generative models inside a sandbox, a controlled testing environment that doesn’t affect actual operations.
That event alone brought in significant interest compared with previous iterations of the annual hackathon, Mekel-Bobrov said.
“There was a 10x increase in the number of teams that were building AI prototypes. That’s a massive increase—we didn’t hire 10x the number of scientists. Instead, we just made this available to non-scientists, to engineers and product managers,” he said. “But then, second to that, what I thought was really cool is a lot of the prototypes were actually developed by teams in one domain toward areas that have nothing to do with them.”
Cost-saving, consumer-facing generative AI
While some companies allow users to be behind the wheel, using generative AI for their direct benefit, other retailers have adopted the technology for consumer-facing advertisements and product photos.
At Coresight’s conference, Doug Weiss, senior vice president of digital and e-commerce at WHP Global, the brand manager behind Anne Klein and Bonobos, said the company has used generative AI to produce social and digital assets.
“The creation of visual assets is one of our largest cost centers, and the reality is, you do a photoshoot, you get assets three weeks later,” Weiss told Sourcing Journal.
WHP has saved money and time by looking to AI, he said. By doing smaller-scale photoshoots, the company tests different assets with varied global audiences, allowing it to adjust if a specific piece of a campaign doesn’t land right with consumers.
Weiss also noted that AI gives WHP the tools to create more representative and inclusive campaigns.
“Historically, as much as we would love to obviously have 100 different models to showcase the million different types of consumers we have, it just wasn’t feasible,” he said. “Now, we are able to have a broader set of representation that actually matches who our customer is.”
Weiss said WHP brands will continue to hire models, photographers and other talent for some of its campaigns. The company reinvests the savings from running fewer shoots into personalization for marketing campaigns.
“In a world where digital marketing has gotten more expensive, and ways of reaching consumers have become more difficult, the ability to be personalized, and to really meet the consumer where they are both in terms of touch point that we connect with them, but also the messaging and how we create content that connects with them, just allows us to stand out from the millions of different marketing messages that they see every year,” he said.
Weiss noted that WHP brands have been piloting AI with room to pivot if legal, regulatory or customer issues crop up.
“We always have a path for us to be able to course correct pretty quickly if things ended up potentially different than what we expected,” he said.
AI ‘hallucinations’: harmful or helpful?
While many companies have released solutions to help their customers, others look to generative AI for internal use cases.
Fossil, which makes watches and other accessories, has been using generative AI to inspire its designers. And in the process of using the tools, the company has uncovered an interesting trend: sometimes, AI hallucinations, when an AI model generates a factually inaccurate or impossible output, can actually be used for good.
“We were surprised to find out that hallucinations were good. If you have an e-commerce chatbot, you definitely don’t want hallucinations,” said Damian Fernandez-Lamela, Fossil’s global vice president of science and analytics. “But now our designers have the ability to utilize [this as an] inspiration tool… and the system can provide a breadth of different ideas.”
Shopify, which has a generative AI tool called Sidekick in place for its sellers, has also seen the effect of hallucinations, said Kriti Kohli, a senior manager of applied machine learning for the Canadian company.
“It also unlocks creativity; it comes up with things that humans didn’t come up with,” she said. “Hallucinated content from a gen AI system can actually be a really good idea or a suggestion for a future product.”
AI’s place in fashion in 2024
Though generative AI seems to be top of mind this year, fundamental AI systems that power capabilities like pricing, inventory management and personalization seem to be at the forefront for retailers.
And as AI adapts at “a double-exponential rate,” per Richa Agarwal, PVH Corp.’s former senior director of digital go to market, consumers have opinions on how companies could use AI to offer better experiences.
Data from Digital River shows that more than one-third of consumers indicated that enhanced search and filtering capabilities and personalized product recommendations are AI’s most impactful applications for enhancing the online shopping experience. Some 27 percent of consumers cited virtual try-on experiences.
Startups including Arvist, Syrup Tech, 3DLook and Zelig have recently raised millions or minted partnerships related to those exact use cases. One in every four dollars going to an American startup this year has some tie to AI, per Crunchbase.
And that funding flurry could spill over into the new year. Precedence Research data suggests that the global market size for generative AI could surpass $17 billion in 2024.
Per Agarwal, as the market expands, the industry should brace for legislative developments.
“We have a massive tsunami of legislation that is coming for fashion retail,” she said.
That in mind, retailers at the Coresight event voiced concerns about AI and data privacy.
However, Agarwal projects that legislation won’t dampen retail’s appetite for AI. When asked to rate on a scale from one to 10 the impact of generative AI will have on in-store experiences, she didn’t hold back.
“I think it’s going to just completely change the way that you walk into a store,” Agarwal said, assigning a score of “20” on the one-to-10 scale. “Generative AI, in certain store formats, is already having a very interesting impact on how people shop.”
Tools that offer sizing and styling recommendations have already started influencing the consumer store shopping experience. But it won’t stop there, she said.
“These are all just the tip of the iceberg,” Agarwal said. “There’s a huge opportunity to completely reimagine the store formats.”
