Amazon launches AI-powered ‘Shazam for clothes’ fashion search

Amazon launches AI-powered ‘Shazam for clothes’ fashion search

- in Fashion
105
Comments Off on Amazon launches AI-powered ‘Shazam for clothes’ fashion search

Amazon is a behemoth of online retailing, but there’s one sector it has yet to crack: fashion. To help dig into this lucrative market, the company is introducing a new AI-powered tool named StyleSnap which will help customers find clothes to buy.

StyleSnap is basically Shazam for clothes, built into the Amazon mobile app. Users can take a picture or upload an image and StyleSnap will use machine learning to “match the look in the photo” and find similar items for sale on Amazon.com.

Amazon’s consumer worldwide CEO Jeff Wilke unveiled the feature at the company’s Re:MARS conference in Las Vegas today, saying: “The simplicity of the customer experience belies the complexity of the technology behind it.”

But lots of startups offer similar services, including online fashion giant Asos. And it remains to be seen if StyleSnap can live up to Amazon’s claims. It might be easy for AI to identify basic items of clothing, like pencil skirts or flannel shirts, but that doesn’t really help consumers who can usually identify these items themselves. Whether or not the feature can match niche items and styles will be the real challenge.

Amazon launched the Echo Look in 2017, which uses AI to analyze users’ outfits.
 Photo by Vjeran Pavic / The Verge

Amazon hasn’t exactly proved it’s a fashion guru, either. Back in 2017 it introduced the Echo Look, an AI-powered camera that gave users fashion recommendations. In our review, we found the software regularly made mistakes, failing to identify different outfits in images and churning out the same tips over and over (it particularly liked our reviewers’ jacket).

But the launch of StyleSnap shows Amazon isn’t going to give up on the fashion business that easily, and the company believes its expertise in artificial intelligence will give it an edge over competitors. It just needs to find the right look.

[“source=theverge”]