Pinterest has introduced Ask Pinterest, an experimental AI shopping app designed to turn the platform’s visual discovery engine into a conversational shopping experience. Users can search with natural language, ask for product ideas, and receive personalized recommendations. For now, the company is keeping it separate from the main Pinterest app as it tests how AI shopping features perform in a standalone environment.
The company said the new app will serve as a proving ground rather than an immediate replacement for existing discovery flows. Features that work well are expected to be folded into Pinterest’s flagship product over time. At launch, Ask Pinterest is available as a web app for mobile and desktop, with access limited to a small group rather than a full public release.
Taste Graph powers context-aware shopping recommendations
The main feature is not simply a chat interface. Pinterest says Ask Pinterest is built around its internal Taste Graph, a data graph that links users with their interests and aesthetic preferences in real time. That system allows the app to respond with recommendations shaped by context instead of relying only on keyword matching.
Users can make multi-step requests such as planning a dinner party or furnishing a room gradually over three months. The app then draws on previously saved Pins and Boards to suggest products and ideas. It can also retain context across sessions, so earlier searches and saved inspirations can inform later recommendations without forcing users to restate the same details every time.
Pinterest takes a different route in the AI shopping race
The launch comes as large technology companies compete for attention in AI-assisted commerce. According to the source material, Google has been rolling out AI shopping features covering product discovery, price tracking, and agent-based checkout. ChatGPT launched an agentic shopping system in 2025, while Meta and Shopify are also expanding their AI commerce efforts.
Pinterest is positioning itself differently. Rather than licensing its recommendation data to outside AI services, the company is using its own data to train models and build products internally. Chief Business Officer Lee Brown said future product discovery will be shaped less by keywords and more by context, taste, and trusted recommendations, which is where Pinterest believes it has an edge.
New ad tools arrive alongside the consumer app
Ask Pinterest was announced together with several AI products for advertisers. In the US, Pinterest has been testing a beta AI assistant inside Ads Manager to help advertisers optimize campaigns. The company also rolled out Performance+ creative globally, a model that selects the ad creative predicted to perform best at each impression from multiple available assets.
Pinterest also introduced the Pinterest Model Context Protocol, or MCP, as an infrastructure layer for standardized campaign management and monitoring through third-party agent tools. The report says the concept is similar to Anthropic’s MCP standard, though Pinterest’s version is built specifically for advertising workflows. Taken together, the announcements show Pinterest pushing AI on both sides of the business: consumer shopping discovery and advertising execution.