Pinterest unveiled a standalone experimental app called "Ask Pinterest" on Wednesday, June 17, betting that a chatbot-style interface can extend its visual discovery roots into more conversational shopping experiences. The company, which rarely breaks from its flagship pinboard layout, is using the new app to test how natural-language queries—like planning a dinner party or staging a room over time—can surface personalized product recommendations drawn from its proprietary "Taste Graph," the internal dataset that maps users to their aesthetic preferences and interests. The app will roll out in limited access to start, giving Pinterest room to iterate without disrupting the core product that more than half a billion people use each month.
The launch comes just days before the ad industry's annual Cannes Lions festival, where AI-powered marketing tools are expected to dominate the conversation. To capitalize on that moment, Pinterest is also rolling out Pinterest Model Context Protocol (MCP), a new framework aimed at advertisers running campaigns on the platform, alongside a broader suite of AI-driven ad products. The company hasn't shared pricing or advertiser-side metrics yet, but the timing signals Pinterest's intent to position itself as a serious player in the AI-advertising race alongside Google, Meta, and Shopify, all of which have been pushing their own agentic shopping features.
Pinterest's strategy differs from rivals in a notable way. Rather than licensing its data to feed other AI products, the company has chosen to train and power its AI features internally, leaning on the Taste Graph as the backbone for personalization. That decision keeps Pinterest's recommendation engine under its own roof, but it also means competing with the likes of ChatGPT, which has been experimenting with agentic shopping, and Google, whose AI tools already help shoppers track prices and check out. Ask Pinterest can also draw on a user's own saved Pins and Boards, allowing the assistant to retain context across sessions and tailor suggestions to what individuals have already curated.
By spinning Ask Pinterest off as its own app, Pinterest buys itself flexibility to experiment with multi-step queries and longer-form planning tasks that don't fit the format of a traditional Pinterest search. Features that gain traction could eventually migrate into the main app, while ideas that don't land can be retired without affecting the broader user base. For now, the company is treating the launch as a research vehicle as much as a consumer product, exploring how AI can reshape shopping behavior without abandoning the visual-first identity that has defined Pinterest for more than a decade.