Pinterest is testing ‘auto-collages,’ a new AI-powered feature that enables advertisers to turn their existing product catalogs into thousands of shoppable visual collages.

This feature is part of Pinterest Ads Labs, a program launched in 2023 to test ad tools before a broader rollout. According to Pinterest, auto-collages use AI to cut out product images, match them based on visual and behavioral signals, and assemble them into new collages.

These collages are built around what users are engaging with, saving, or searching for. For example, if a retailer sells clothing, Pinterest can scan its catalog and generate new outfits using items from that catalog without the brand needing to manually design each look.

How brands can use auto-collages to create shoppable content

The platform says auto-collages create and group product images together based on themes like outfit ideas, product similarity, high user engagement, and items similar to those users have saved on their boards.

Brands can get outfit ideas from existing collages and use the tool to build their own. Auto-collages can also mimic the structure and style of high-performing collages. So, brands can use top-performing collages as templates to create new ones. For instance, if a collage of winter boots performs well, the AI can generate variations that follow the same visual pattern using different products.

It also groups similar products into a grid layout that users can shop from directly. Advertisers can create collages similar to products users have saved by using auto-collages.

Pinterest updates Pinterest Trends tool

Pinterest is also expanding what advertisers can do with its Pinterest Trends tool. This tool was already used to monitor what people are searching for. Now, it can also analyze what users are saving, curating, and shopping for.

Advertisers can now use the updated Trends tool to see which product categories their audience is saving on their boards and what they are clicking on across retail sites.

The tool also groups trends into high-level themes. Instead of browsing through countless individual search terms, marketers can pull up a trend and find what is popular within that category. Advertisers will also have access to visuals, trend summaries, and search volumes.

‍