How Pinterest’s VLM-Powered AI Assistant ‘Gets’ Your Style

The world of Pinterest aesthetics is getting an AI makeover with Pinterest Assistant, a visual-first AI shopping collaborator.

Built on a visual language model, the Pinterest AI Assistant can deliver content that is catered to the user's personal saves, boards, collages and data from users with similar tastes, all through a simple voice prompt. 

The company says that "unlike traditional search or chatbots, which are helpful once you already know what you’re looking for, Pinterest Assistant is designed for the moments when you don’t".

Pinterest's CEO Bill Ready adds: “People, especially Gen Z, say that the magic of Pinterest is that it ‘just gets me’, whether that’s finding the perfect outfit or knowing your distinct style,” says

“With Pinterest Assistant, we’re supercharging that magic by leveraging AI to help our users discover and shop like they would with that person who knows them best.”

What are Visual Language Models (VLMs)?

VLMs are trained to map the relationship between text and visual data, which allows the model to process natural language prompts against the backdrop of visual information.

They generally have two main parts: a vision encoder and a language encoder.

A language encoder links the semantic meaning of words and phrases, along with their contextual meaning, to vectors that allows the model to process natural language information. 

A vision encoder, on the other hand extracts visual information such as colour, shapes and textures from the input images or videos and transforms them into vectors, which gives the model a metaphorical pair of eyes. 

Pinterest AI assistant, a shopping game-changer

Pinterest’s latest multimodal AI, the company claims, outperforms traditional models by more than 30%, in the relevance of shopping recommendations produced. 

This significant difference is due to the fact that the Pinterest AI, with access to each user’s specific tastes, can leverage that information using their proprietary Taste-graph to create outputs that are solely cater to each user’s personal aesthetic preferences.

“Unlike traditional search, which relies on keywords and scrolling through results, Pinterest Assistant is designed for open-ended exploration,” says Pinterest CTO Matt Madrigal. 

“You simply start a conversation – ask about a vibe, style, colour or even show an image. It draws from your boards, saves, and Pinterest’s vast catalogue to deliver recommendations that fit your unique taste in real time.

“What sets Pinterest Assistant apart is our multi-modal AI and proprietary taste-graph, built for smarter, more intuitive AI. 

“It can understand complex, conversational queries that blend text, visuals and even voice, making it possible to capture intent in natural, human ways."

Pinterest Assistant adapts as your style evolves, responding to what catches your eye and what you save.

In doing so, it turns shopping and discovery into something more intuitive, more personal and closer to how inspiration naturally unfolds.

Matt says that this new assistant is "just the beginning of our journey toward a future of AI-powered, visual reasoning".