Beauty is the fifth-most-searched category on Pinterest. However, the company regularly hears from its 200 million monthly users that searching for beauty images and tutorials that match their skin tone is next to impossible.

In response, the social network is adding an “inclusive discovery feature” that allows users to narrow search results by selecting a skin-tone range. The goal is to improve discoverability by featuring more relevant images that are better tailored to users browsing for product inspiration and beauty tips. The function, which is currently in an experimental stage, is powered using ModiFace’s artificial intelligence software that uses machine learning to scan images on the site and group them into respective skin hues. According to the Pinterest spokeswoman, the algorithm then works by analyzing and tracking the most liked and viewed images by individual users, and sharing results similar to those.

Users are currently able to select from four color palettes to receive more nuanced results, though a Pinterest spokeswoman said the platform plans to expand the tone selection as the technology advances.

“During the test period, ‘Pinners’ shared positive and helpful feedback, which we’ve incorporated. We’ll continue to listen to feedback, improve our technology and expand the feature to more skin-tone ranges,” the spokesperson said.

Skin tone ranges closer view_preview

Pinterest’s inclusive discovery tool in action.

The improved search functionality is, in some ways, a response to the continued push for diversity in product development and marketing within the beauty industry. The movement has continued to gain momentum in recent months, with the launch of buzzy collections like Fenty Beauty by Rihanna, intended to service a wider consumer demographic of women of color. While Pinterest does not share information about the racial demographics of its user base, the function will largely help minority users bombarded with beauty tips for white women find content more relevant to them.

For Pinterest, which has strategically ramped up e-commerce partnerships with fashion and beauty companies over the past few years, the enhanced search will ultimately help raise affiliate marketing profits by driving conversions through better connecting users to products they’d be likely to buy. According to data from the company, 70 percent of weekly active users go to the platform to discover new beauty or grooming products, and 40 percent create customized “boards” of products they intend to buy. Looking to the future, Pinterest aims to enhance the feature, not just for beauty but also to deliver personalized results in other areas, like size.

Though ModiFace already had an existing library and catalog of images, many of which were already on Pinterest, the search function took six months to build, in order to improve skin-tone detection, particularly for user-generated photos that can be poorly lit or out of focus.

Parham Aarabi, founder and CEO of ModiFace, said this is ModiFace’s first foray into working with a major social media platform to optimize search results. Following its recent acquisition by L’Oréal, ModiFace’s partnership with Pinterest points to the technology’s growing capabilities and versatility, after it already served as a hot commodity in the beauty industry for allowing consumers to virtually try on makeup from their mobile phones.

“I think personalization is important, and something that users will expect in a variety of situations including shopping and searching,” Aarabi said. “In the context of searching on Pinterest, we believe this will be helpful.”