AI for retailers is progressing
AI in retail adoption has been relatively slow, but it's starting to pick up as retailers see the benefits of AI technologies and the realities of e-commerce competition.
As e-commerce becomes the default mode of shopping in the smartphone age, traditional retailers are increasingly turning toward new technologies and forms of advertising to attract and retain customers.
The retail industry is readily using analytics, AI and IoT software to create better online shopping experiences, more personalized smartphone apps and marketing campaigns, and, for some retailers, in-store robot workers that perform basic tasks and help augment existing workforces.
The progress of AI for retailers, in particular, has been slow, with the retail industry typically lagging behind other industries, including healthcare and finance.
Traditional retail industry has fallen behind
For one, the brick-and-mortar retail business has trouble finding talented engineers and data scientists, said Nikki Baird, vice president of retail innovation at Aptos, a vendor of a data management platform for retailers.
That puts retailers in a difficult position, as the tech giants and other industries such as financial services suck up most of the AI talent, she said.
In addition to a limited talented pool, retailers historically have been wary of black box AI, Baird said. The models, complex to the point that they are difficult or impossible to understand how exactly they work, essentially ask retailers to rely on their data and predictions without fully explaining them.
At one time, retail was predictable, Baird said, as companies tended to enjoy bigger profit margins and more consistent growth rates. That's generally no longer so, and "retailers are waking up to the idea that it's not business as usual," she said.
For retailers, AI represents an inherent, and previously unwanted, risk, as they put faith in an emerging technology that is not always entirely understood.
It's a gamble that can pay off, however, especially for personalization.
"They collect everything in the hopes they collect the important thing, but they don't really know what the important thing is yet," Baird said. AI can help cut through all the data and pluck out what's important.
With AI for retailers, "now we can make recommendations that actually make an influence on that customer behavior," she said.
AI can help
Nikki BairdVice president of retail innovation, Aptos
It's still a learning process, though, and many brick-and-mortar retailers experiment with new ways to market their wares.
In-store, retailers have been experimenting with augmented reality (AR), encouraging customers to take out their smartphone and point their camera at different products in the store. An AR overlay can then give more information about a product, compare its price in the store to other places or even offer special coupons.
Online, retailers create smartphone apps, websites and advertising campaigns on social media that cycle user data through algorithms to create personalized advertising.
Numerous tech startups are selling into the retail market, including Fyusion Inc., a six-year-old vendor that sells a platform that can create life-like 3D models of real-life objects using nearly any type of camera.
Using machine learning and advanced capturing techniques, the San Francisco-based vendor sells a platform that generates 3D models of real-world objects through 2D images captures by users.
Radu B. Rusu, CEO of Fyusion, explained that the model relies on geometric and position data, which gives the otherwise static images depth. An algorithm then compiles the images and data to create a 3D image.
It's a relatively quick and easy process as users can use their smartphones to take the photos that will be translated into the 3D model, Rusu said. Over time, the model will automatically recognize and categorize different types of objects, which can speed up the process even more.
Fyusion has clients in the retail, automotive and fashion industries. Retailers, Rusu said, use the platform to better advertise their products, giving potential customers a life-like view of a product without having to go into a store and see it personally.
"It helps to better visualize something online that you're going to be purchasing," Rusu said.
Still, despite new technologies designed with retailers in mind, the take up of AI for retailers will be relatively slow, especially on the forecasting side of business, Baird said.
"A wrong forecast of how much product a store should order would be a problem," Baird said.