At Sentient, we believe that every retail problem is an AI problem. And while our Intelligent Commerce solutions solve some of the biggest issues facing retailers today, they don’t yet solve every problem. We’ve been listening to our customers to discover what their most pressing challenges are and two things keep coming up time and time again: mobile engagement and making editorial images shoppable. And we’ve now got the AI to solve them.
We’re debuting these solutions this week at ShopTalk, but we wanted to share them here on our blog for the folks who can’t attend.
Sentient Aware ‘Shop the Look’ Makes Every Editorial Image Truly Shoppable
The first, Sentient Aware – Shop The Look, allows customers to find fashions they love and personalize them with their unique sense of style. Here’s how it works:
Imagine a shopper browsing a retailer’s Facebook page. She sees a summer outfit she loves and clicks on it. That’s where Shop the Look kicks in.
First, it understands every article of clothing in the image. Then, it analyzes those items and surfaces the same or similar garments for that individual shopper. Instead of seeing only the identical outfit she clicked on, she now sees options: several hats or billowy skirts to choose from. Each item she clicks on trains the AI about what’s interesting to her and lets it pull new, relevant products from a retailer’s catalog to complete the look. In other words, Aware lets users browse effortlessly through vast catalogs and quickly find the styles and outfits that match their own unique taste.
After all, we’re all different. And while we might be interested in a particular outfit, we don’t always love the whole thing. Shop the Look lets users create their own version of a look or even use it as a jumping-off point to find items that are even more closely suited to their taste.
Sentient Aware – Shop The Look works by using deep learning object detection. Our AI scans the image and then derives X,Y coordinates of each product. It then isolates each object, scans them again against hundreds of dimensional features – think size, shape, color, and subtler features of a product’s visual style – and traverses our existing AI models to find the best apparel match. Once the AI finds that match, it merchandises that product or recommends similar products from the retailer’s catalog, so a shopper can complete their unique outfit or look.
Aware for Mobile Brings a Real-time Personalized Experience to Every Mobile Shopper
The second, Sentient Aware for Mobile, builds on the AI-powered personal shopper experience deployed by leading brands such as Sunglass Hut (‘My Frame Finder’). Powered by a new mobile-tuned algorithm, Aware for Mobile is able to change individual recommendations on-the-fly based on shoppers’ interactions. It does this by providing a new, Tinder-like experience that allows shoppers to quickly find products they are looking for in two or three swipes or taps on their mobile phone. To implement, retailers simply use the white-label UI and add their own branding after integrating Aware into their catalog.
Sentient Aware for Mobile™ works by evaluating every one of a shopper’s taps, swipes, and scrolls in real time. These actions are interpreted as signals of varying significance and preference for certain styles in the catalog, and are translated into certain vector values matching products in the catalog. The AI then matches these values against hundreds or thousands of vector values in a dimensional “feature space” that represent each product in a retailer’s catalog. The closer a shopper gets to what they like, the stronger those signals get. These signals then allow the AI to narrow down to precisely the product features and vectors that the shopper wants in just a handful of taps and swipes. The AI can also pick up on nuances and a user’s unspoken preferences in case shoppers want to explore a different path.
Other background materials on Sentient Aware: