Here's how this Bay Area startup is helping online shoppers find clothes that match their style

Shopping for clothes online can be a frustrating experience.

Fashion e-tailers often sort clothing into very broad categories — gender, size, color, price. But those categories can lump together very different items that appeal to people with totally different fashion sensibilities. Finding something you like can be like searching for a needle in a haystack.

Purva Gupta founded Lily AI to solve that problem. The Mountain View startup offers a service that's designed to help consumers more easily find clothes they like and to help online clothing stores better understand customers' individual styles.

The fundamental shortcoming of the online fashion shopping experience is a lack of data, Gupta told the Business Journal. While retailers know and list the basic attributes of particular clothes they sell, those likely aren't the reasons why customers will buy them. Price, size and color are important. But people buy clothes because of their particular cut, the way they're shaped, or the feel or the specific pattern of the fabric.

"There could be 40 other attributes about this blouse that could be the reason why I like this blouse," Gupta said. "Lily AI is built on this fundamental hypothesis that the e-commerce (ecosystem) is starved of this basic data."

Lily's system uses artificial intelligence to automatically detect the features of particular pieces of clothing. Its service allows clothing e-tailers to list 10 times as many attributes per item as they normally would, according to the company, including things like fit, style and the kinds of occasions for which the outfit is best suited. Clients can integrate its service into their existing online storefront systems.

The company's service makes it easier for customers to find the exact outfit they're searching for, because they can search for particular features. But it also allows online clothing stores to track which attributes are most important for specific customers. The company, legally known as OneLook Inc., aims to develop its service to the point where e-tailers will be able to accurately predict what individual consumers will buy, Gupta said.

Already, Lily's customers, which include Bloomingdale's Inc., ThredUp Inc. and Backcountry.com LLC, are benefiting from its service. Its software has boosted clients' revenue 10%, she said.

Lily is branching out

More customers could be signing up soon. The company announced last week it has begun to offer its service on a global basis.

Lily charges customers different rates based on the size of their product catalogs, the number of different applications with which they want to integrate its system and how frequently they connect to its services.

Gupta and co-founder Sowmiya Narayanan — the company's chief technology officer — launched Lily AI as a standalone consumer app in 2017. But they found that model wasn't very lucrative, so two years later, they shifted the business to focus on enterprise customers.

That model seems to be taking off. The company raised $12.5 million in a Series A round in January 2020. And since the pandemic, Lily's grown from 25 employees to 71, Gupta said. The company plans to hire 29 more by the end of the year, she said.

Unlike many startups, gender diversity has been a founding principle at Lily. Not only does it have two female founders, it has an all-female board and more than half of its employees are women, Gupta said. Given the company's fashion focus, that emphasis on hiring women of all levels has been a strength, she said.

"We've been really fortunate in how diverse the team is in the type of product that we're building."

Lily's next move will be to offer its service for use by other kinds of retailers, Gupta said. Top of the list: home and beauty product e-tailers, she said.

But its technology can be applied to a wide range of products, she said. Lily's service just needs to be able to examine digital photos to find the products' attributes and then it can track which of those features interest customers, she said.

"It's definitely not something that we want to limit to fashion, but take to other industries, especially industries where there's some emotional context," Gupta said.