How AI is Transforming Retail for the Better

At this point, most of us interact with artificial intelligence on a daily basis. Sure, we might not realize it consciously, but every time we search Google or see what our friends are up to on Facebook, we're interfacing with AI. Better computing power, ballooning data stores, and research momentum point to the fact that AI isn't just having a moment; it's here to stay.

And while we most often hear about the splashy successes of artificial intelligence – whether it's DeepMind besting a Go world champion or the continued refinement of self-driving cars – AI really is everywhere. And one area that's particularly exciting doesn't involve robots or the other trappings of science fiction. That area? Online retail.

When you get right down to it, the shopping experience online hasn't much changed. We still use the search bar and inexact language to look for the product we want. We're still presented with page after page of search results. We're still browsing those results or trying to search again or clicking facet boxes on the side to refine our query.

Which is to say: shopping online still feels like interacting with a database.

But advances in artificial intelligence are poised to change that. For shoppers, it means finding better products in fewer steps and with less hassle. For retailers, it means more intuitive sites with higher conversion rates. There are even implications for design and trend spotting.

So, again: think about what shopping online is today. We're used to it, but it really is a bit cumbersome, especially when you contrast it with the brick-and-mortar experience. Sure, in-store shopping means finding parking and waiting in line behind someone who still uses a checkbook, but it also means having the ability to talk to a salesperson who can help you find something you'll love, not just something you like. For a great salesperson, that means gauging your sense of style, listening to your questions, and having a deep understanding of the storeroom and the product he or she sells.

The promise of AI in retail is combining the best of online shopping with the deft touch of a great salesperson: convenience and hyper-personalization.

Instead of interacting with a check boxes and trying to use the right vocabulary to find the best product, what if you could just click a few items? What if AI could find the similarities between those products – both the big ones such as color and size and the subtle ones such as the scoop of a neckline – and intuit your sense of style? What if simply browsing a catalog trained an AI bot to understand each user's unique sense of style?

That's exactly what's possible and exactly what's happening already. Not only that, but the AI is doing this from the images themselves, not from product tags or manufacturer metadata. And the shopper? They don't know they're interacting with cutting-edge AI; they just know they're finding better products with less fuss. They're more satisfied and more likely to not only buy, but return, time and time again.

Simpler product discovery with AI means simpler sites too. Instead of facets and search bars and pagination, users can interact with images instead, wordlessly navigating to products that match their sense of style.

For retailers, that means more than just happier customers (though few would complain about that!). It means reduced friction too: instead of relying on unwieldy and inexact product tagging operations, retailers need only images to surface their entire catalog. Which brings us to another point: because AI doesn't have to rely on legacy techniques like cohort recommendations (i.e., "people like you also bought"), it can serve up product recommendations that are in fact more similar. That's because it isn't checking purchase histories or even browsing histories; rather, it's looking at the images and showing which products fit that user's preference, in the moment. Sometimes, that might be the most popular shirt on the site. Other times, it might be a product buried on page 45 of the search results. But the outcome is the same: better recommendations based on less interaction, that are more personal.

Let's go back to site experience for a moment. All those facets and pages and the rest of the bells and whistles that lead us to the products we want? Those are distractions. The goal shouldn't be giving a shopper a bunch of ways to filter a spreadsheet, it should be letting them explore and find products that delight them. Simpler interfaces with far fewer dependencies on consultants and system integrators means that brands and site designers can be more creative, putting together sites that more closely mirror a company's personality rather than each site being a reskin of the ubiquitous e-commerce storefront.

In other words, AI means less hassle for both retail sites and their shoppers. It means combining the best of the online world (massive catalogs, convenience, the ability to shop from our couch wearing our sweatpants) with the best of the brick-and-mortar one (the immersive experience, the personal touch, and the deep understanding of product nuance a tremendous salesperson has). It means more rewarding experiences for everyone.

And when that's the case, when retailers and their customers can enjoy tangible benefits of the same technology, you know you're onto something better.

Andy Narayanan is vice president of intelligent commerce, Sentient Technologies.