Conversational Discovery via 'Taste Graph'

Pinterest has launched 'Ask Pinterest,' an experimental standalone app designed to shift product discovery from keyword-based search to a conversational, agentic interface. By leveraging its proprietary 'Taste Graph'—which maps user interests and aesthetics—the app aims to handle complex, multi-step queries that traditional search struggles with, such as planning a dinner party or furnishing a room over time. The app integrates directly with a user’s existing saved Pins and Boards to provide context-aware, personalized recommendations.

Strategic Decoupling and Experimentation

By releasing 'Ask Pinterest' as a separate application, the company is isolating its AI experimentation from the core Pinterest experience. This allows for rapid iteration and testing of conversational flows without disrupting the primary platform. Unlike competitors who may license their data to third-party AI services, Pinterest is doubling down on using its internal data to train its own models and power its proprietary AI products, positioning itself to compete directly with agentic shopping tools from Google, Meta, and Shopify.

AI-Driven Advertising Infrastructure

Alongside the consumer app, Pinterest introduced new AI tools for advertisers to modernize campaign management:

• Performance+ Creative: A global model that automatically selects the highest-performing ad creative for each impression.
• Ads Manager AI Assistant: A beta tool designed to streamline campaign management for U.S. advertisers.
• Model Context Protocol (MCP): A new infrastructure layer that enables advertisers to monitor and manage their Pinterest campaigns using third-party agentic tools in a standardized, interoperable way.

These moves reflect a broader strategic shift, as Chief Business Officer Lee Brown noted that the future of discovery is moving away from keywords toward context, taste, and trusted recommendations.