The eTail West conference in Palm Springs, California, celebrated the brand’s 25 years of hosting live events with perspectives on core trends in online retail. True to industry priorities in 2024, many of the discussions focused on emerging technologies. One pervasive thread was artificial intelligence (AI), and especially the topic of how retailers are using AI.

“One of my favorite things to say is ‘Technology for technology’s sake: Probably not the best idea,’” said Rose Hamilton, a founder and principal advisor at the firm Compass Rose Ventures, during one of eTail West’s Feb. 28 morning sessions.

Rose Hamilton, founder and principal advisor at Compass Rose Ventures, speaking at eTail West | Photo credit: eTail West

Hamilton acknowledged the new opportunities being opened up by using AI, machine learning and data in combination to improve capabilities. However, she encouraged retailers present to not lose focus.

“What I don’t believe is actually going to change in another 10 years is the importance of intimacy of knowing your customer,” she stated.

At Hamilton’s panel and others throughout the show, retail leaders swapped stories and best practices from their recent experiences. These insights covered early approaches and how to test AI tools before deciding to deploy them widely. However, they also included accounts of how AI is already being used and what they have learned. Those discussions eventually led to opinions on standards and guardrails that some retailers have set up to avoid missteps.

Editor’s note: This is part one in a two-part look at the AI perspectives presented at eTail West. Read the second part, where retailers share recommendations.

How to treat testing before adopting AI tools

One eTail panel on Wednesday focused explicitly on current AI experiences in retail. There, how to treat the pre-adoption phase as an organization became a major question.

“You bring an open mind and a testing mentality,” said Suruchi Shukla, vice president of marketing at the designed products marketplace Minted. Shukla called this mindset “key” to getting started.

She also acknowledged initial barriers.

“Our biggest factor that comes in is inertia,” she recalled. “How do we even get started?”

As a solution, her team looked for ways to get past indecision to lay down a basis for informed decisions later on.

“Six months ago, my team was struggling,” Shukla said. “So we found a boot camp and I attended.”

They spread the boot camps out over a few days. Members of her marketing group attended. They gained exposure “to many tools,” she said, which they then brought back to Minted. As for the process of integrating those tools into projects and workflows, Shukla highlighted this testing phase before clear results were expected as crucial to getting better outcomes later on.

“I think it’s also giving your organization the permission to go experiment without having that expectation of performance immediately,” she assessed. “So people started to play with it and organically that adoption started. So I think it’s more about using AI as a starting point and not immediately putting any expectations and rules around it.”

Minted is No. 407 in the Top 1000. The database is Digital Commerce 360’s ranking of North America’s online retailers by web sales. Digital Commerce 360 categorizes Minted as a Flowers & Gifts retailer in its Top 1000 database.

How retailers are using AI tools now 

As Hamilton mentioned earlier in the day, AI and many of its associated use cases for online retail are not necessarily new. For example, Rachel Frederick, a vice president and general manager, ecommerce, at Sur La Table, noted that her company has been using AI tools at least as far back as 2017.

Sur La Table is No. 408 in the Top 1000. Digital Commerce 360 categorizes Sur La Table as a Housewares & Home Furnishings retailer in its Top 1000 database.

Frederick mentioned Bloomreach for smart site search and Contentsquare for behavioral analytics as two pieces of Sur La Table’s technology stack. She said these tools have been important for informing “product recommendations” and “customer behavioral analytics, being able to identify friction and understanding what the impact is or the opportunity” to prioritize.

In the meantime, she said Sur La Table is “testing and learning a lot” with use cases in email and specifically generative AI for “tools that help deflect WisMO [“Where is my order?”] calls for customer service.”

“That’s something that we’ve done,” Frederick said. “Voice response systems in AI, as well as looking at things like creating enviro-lifestyle assets.”

The assets she referred to can involve real product photos layered with generative AI-executed changes to provide new contexts for different scenarios. She used cookware to illustrate an outcome:

Frederick’s team may start a request “by using product images and being able to give a prompt to say, yes, here’s a Dutch oven product image, but now I want you to create another image that has it on a stovetop with stew,” she described. “Now, give me another image that has that same Dutch oven — different color, sitting on a cutting board with a loaf of bread. It’s just showing the diversity of a product, and it may be very seasonally relevant.”

“Through that gen AI, we’re actually able to create more assets very quickly and efficiently,” Frederick explained.

Generative AI use for product images

At Minted, Shukla detailed another visual use case, where she said the company worked with the vendor Cloudinary to produce customized art images on Minted’s product detail page (PDP).

“They built a unique service for us where we can do the base image shooting, but they then help us generate images in situ,” Shukla said. “So for an art product or a painting, people actually want to see it, and the sizes are important.”

The generative AI can help display the artwork against different types of backdrops with new surroundings.

“Where it’s placed on the sofa and stuff is important, but we don’t have the bandwidth or the resources to shoot so many setups and so that service has been fantastic for us to use this gen AI used in a very different way.”

Shukla also cited Adobe’s Firefly software, which Minted used, started during the 2023 holiday season to reduce time spent by its customer service team on certain problems.

“At Minted, when you’re building your holiday cards, people upload their images,” Shukla said. “And in the past, they used to reach out to our customer service team, asking them their help for retouching — you know, they want to remove a car from the background, something like that.”

She credited Firefly for its role in quickly accomplishing tasks including background expansion, filling and retouching.

“That actually really helped our design services team,” she stated. “Their bandwidth really opened up.”

Guardrails for AI and associated risks

Retailers at eTail were often enthusiastic and complimentary of AI’s capabilities. Still, they also outlined where limits and risks for AI tools can necessitate guardrails. The risks, as Shukla shared, can be external, as well as internal for organizations.

“It’s the brand reputation risk that’s top of mind,” Shukla said. “We’ve all kind of heard about AI inaccuracies, AI biases, and if people remember when a few years ago Google had come out and their AI couldn’t recognize certain faces as humans. The brand reputation damage risk is so high that we have to be very careful about how much AI is being used.”

The spectrum of consequences she addressed ranged from brands coming across as robotic and out of touch to matters of credibility when unverified facts are being introduced to messaging with generative AI.

“Internally, though, there is the employee morale and engagement risk,” she continued. “That’s becoming very real.”

Shukla framed the morale challenge as being one of empowering and educating teams versus stoking and fueling fears.

“People are resisting adopting AI because of the risk to their jobs,” she explained. She stressed that “any of these AI tools can really be productivity-enhancing tools and don’t have to be about workforce reduction in any way.”

Avoiding these pitfalls is not just a communications and management issue, according to multiple speakers on stage. As the conversation continued, they suggested confronting the very real anxieties raised by using AI to boost productivity. At the same time, they urged companies to value and leverage the human teams who use AI tools. That may include teams who set or execute strategy. In other cases, it may include those who keep human interaction alive in customer experience.

Read more about the role of human expertise and the controls and limits for AI that retail leaders discussed in the second part of this article here on Digital Commerce 360.

Do you rank in our database?

Submit your data with this quick survey and we’ll see where you fit in our next ranking update.

Sign up

Stay on top of the latest developments in the ecommerce industry. Sign up for a complimentary subscription to Digital Commerce 360 Retail News. Follow us on LinkedIn, Twitter and Facebook. Be the first to know when Digital Commerce 360 publishes news content.