Pinterest’s newly launched “Ask Pinterest” AI shopping app: breaking through with an aesthetic edge

Image social platform Pinterest launches experimental AI shopping app "Ask Pinterest," allowing users to search for products and plan purchases through natural language conversations, with cross-session memory of preferences and styles. It charts its own course in the AI shopping battlefield surrounded by Google, Amazon, and Meta.
(Background recap: Amazon launches Alexa for Shopping: compare products, track prices, regular purchases, and support cross-platform shopping)
(Additional background: a16z warns "AI agents that don't see ads" will place orders directly: $291 billion in online ads will be destroyed)

Table of Contents

Toggle

• Conversational Shopping: Remember Your Style Across Sessions
• Google, Amazon, Meta Compete for AI Shopping Dominance
• MCP Protocol and Performance+ Launch Simultaneously

On Wednesday, Pinterest announced the launch of an experimental AI shopping app called "Ask Pinterest" on the eve of the Cannes Lions Festival, extending the platform’s well-known visual discovery experience into a conversational interface, allowing users to ask questions in natural language to receive highly personalized product recommendations and inspiration.

According to the explanation, this standalone app will not interfere with the main application experience in the short term but will serve as a testing ground for Pinterest’s AI shopping features. Pinterest stated that in the future, the results from Ask Pinterest will gradually be integrated into the main app, becoming the foundation of AI features for its flagship product.

Conversational Shopping: Remember Your Style Across Sessions

Unlike traditional Pinterest, which centers on image search, the biggest feature of Ask Pinterest is its use of the platform’s "Taste Graph," an internal data graph that instantly links users with their interests and aesthetic preferences to provide context-aware recommendations.

Users can make multi-step, complex queries such as "Help me plan a dinner party" or "I want to slowly decorate this room over three months." The app will suggest based on the user’s past pins and boards, even maintaining context across sessions.

This means users don’t need to describe their needs from scratch each time: the Nordic sofa searched last week or the table setting saved the day before yesterday will be remembered and naturally incorporated into subsequent recommendations. Pinterest believes that this continuous conversational experience, which retains context, is precisely the value that traditional keyword search cannot offer.

Currently, Ask Pinterest is available as a web application (supporting mobile and desktop) with limited experience slots and has not yet been fully opened to the public.

Google, Amazon, Meta Compete for AI Shopping Dominance

The emergence of Ask Pinterest comes at a critical moment when AI chatbots and traditional search engines are vying for consumer attention. Google has gradually launched AI shopping features, from product discovery and price tracking to AI-powered checkout; ChatGPT is also launching an agent-based shopping system in 2025; Meta and Shopify are also actively deploying in the AI commerce space.

However, Pinterest’s strategy is quite different from these giants. It does not, like some platforms, provide its product recommendation data to third-party AI services via licensing agreements. Instead, it focuses on training its own AI models with its data to power its own AI products. Pinterest CBO Lee Brown stated in an official announcement:

"Future product discovery will no longer be driven solely by keywords; it will be shaped by context, taste, and trusted recommendations." This is precisely where Pinterest believes it has a unique advantage.

MCP Protocol and Performance+ Launch Simultaneously

In addition to Ask Pinterest for consumers, Pinterest also announced a series of AI tools for advertisers ahead of the Cannes Lions Festival, demonstrating its dual-track strategy to simultaneously enhance consumer AI shopping experiences and improve enterprise AI advertising efficiency.

First, Pinterest is testing an AI assistant (Beta) within the US Ads Manager to help advertisers optimize campaigns. Second, the new AI model "Performance+ creative" has been launched globally, capable of automatically selecting the best-performing ad creatives from multiple assets for each impression.

Third, Pinterest released the Pinterest Model Context Protocol (MCP) infrastructure layer, enabling advertisers to manage and monitor ad campaigns via standardized methods through third-party proxy tools. This is similar to the open MCP standard proposed by Anthropic but tailored specifically for the advertising ecosystem.

AI shopping apps, AI ad models, open protocols—Pinterest aims to launch all three by mid-2026 to keep pace with the AI transformation wave.