The Riviter team (from left to right) Dana Robinson, Emily Ruff, Andi Hadisutjipto, Hannah Jurkowicz
The Riviter team (from left to right) Dana Robinson, Emily Ruff, Andi Hadisutjipto, Hannah Jurkowicz (Courtesy of Riviter)

What if your Pinterest board was your personal shopper?

That’s the idea behind Riviter, a startup out of University of Chicago-Booth. Riviter’s tech uses image recognition technology and computer vision to analyze users’ fashion pins and social media likes to offer shopping recommendations on e-commerce websites. Founder and CEO Andi Hadisutjipto anticipates large retailers could use the tech as a plug-in on their own site, offering customers a personal shopping experience based on their already-curated social media.

The idea was inspired by an online shopping pain point: people are more likely to shop online than ever before, but large retailers still rely on cookies and keywords to understand what a customer might want. “Shopping for clothing is obviously a very visual decision, and yet a lot of the tools that are out there are built around language, and are not friendly for what a woman needs,” said Hadisutjipto, a current MBA student at Booth.

“Today all the data that is being used to form recommendations is based on clicks. Clicks are indiscriminate,” she added. “If you click on something it doesn’t say anything about whether or not you liked it. So whether you liked it [or not], it can follow you around on the Internet.”

Instead of tracking user preferences through cookies (which leads to a clothing item that “haunts” you around the web) she said Riviter could offer retailers a way to get an accurate style profile of a potential customer using the pictures they would pin and like on social media anyway. 

“Today all the data that is being used to form recommendations is based on clicks. Clicks are indiscriminate.”

The tech isn’t entirely new. You’ve probably heard of image recognition technology used in Facebook’s photo tags, Amazon’s (now-defunct) Fire Phone, and Google Photo’s organization tools. But it’s largely untapped tech in the fashion industry, said Hadisutjipto. The startup is also building a machine learning layer to the tech that can learn about a shopper’s taste and preferences over time.

“Your shopping experience changes because when you’re going to your favorite retail site, you’re actually seeing things that are fit to your tastes and similar to what you like,” Hadistujipto said.

Currently Riviter is halfway through the I-Corps program run through UChicago’s Polsky Center, which has helped the startup identify their end customer. Originally, they anticipated that this would be a consumer-oriented app, but as they interviewed potential customers they realized they wanted to reach a shopper when they’re in the “shopping mode”–already on a retail website. With that in mind, they’re focusing on large retailers who could plug the tech into their site (similar to Nordstrom’s True Fit, which helps people figure out their correct size when shopping online).

Next up, Riviter will pitch at SXSW’s MBA competition in March.