Pinterest Launches AI Business Assistant and MCP Protocol for Advertisers

Pinterest is rolling out a suite of AI-powered advertising tools designed to help brands navigate a shift from keyword-based search to conversational, recommendation-driven discovery.

The company announced Business Assistant, Pinterest Model Context Protocol (MCP), enhanced Performance+ creative capabilities, and Ask Pinterest, a standalone experimental app for testing AI shopping experiences. The moves come as the platform seeks to differentiate itself in an increasingly AI-mediated advertising landscape.

Visual AI for advertiser workflows

Business Assistant, currently in closed beta in the US, integrates AI directly into Pinterest's Ads Manager and mobile app. Unlike text-heavy chatbots, the tool presents insights visually—displaying trending search terms in graphs and surfacing high-performing Pins that align with emerging user interests.

If searches for "clean beauty routine" spike 42% in a given week, for example, Business Assistant can show advertisers how that interest grew over time and highlight relevant Pins to inform their creative strategy. The mobile version delivers proactive notifications about trends, performance changes, and optimization opportunities.

Standardized integration for marketing tools

Pinterest MCP addresses a different challenge: connecting Pinterest data to the AI copilots and workflow tools advertisers already use. Built on the open Model Context Protocol standard, it provides secure, structured access to campaign metrics, analytics, and keyword insights.

The infrastructure allows partner platforms to incorporate Pinterest's taste and intent signals into their own AI-driven guidance without requiring custom integrations for each tool. Alpha partners include PMG, Pacvue, Dentsu, Havas, Innovid by Mediaocean, and Omnicom's Jump450.

"What makes Pinterest especially compelling is the strength of its intent signals," said Chris Ivey, President of Jump 450. "People come to Pinterest to plan what they want to do next, creating valuable context that helps marketers make smarter decisions."

Dynamic creative optimization

Pinterest is also upgrading its Performance+ creative suite with a new AI model that evaluates creative variations at the asset level rather than the ad level. The system selects which variant to serve based on predicted performance for each individual impression.

In testing, the model increased click volume by 7.5% compared to Pinterest's previous approach, according to internal data from late November 2025. Advertisers will also gain access to new ad review tools and more granular creative performance reporting.

Experimental shopping app

Ask Pinterest, a separate app with limited access, serves as a testing ground for more conversational shopping experiences. The app is designed for multi-step decisions that don't fit neatly into single searches—planning a dinner party on a budget, finding a personalized gift, or furnishing a room over time.

The experimental format allows Pinterest to explore how AI can support complex shopping journeys while maintaining context across sessions. Insights from Ask Pinterest will inform future AI features in the main app.

Why it matters

As generative AI reshapes how consumers discover products and make purchasing decisions, platforms are competing to become the recommendation layer between brands and buyers. Pinterest's approach leverages its existing strength in intent-driven browsing—users actively planning future actions—rather than competing directly with general-purpose AI assistants. The MCP protocol in particular signals an effort to embed Pinterest data into the AI tools marketers are already adopting, rather than requiring them to add another standalone platform to their workflow.

Pinterest announced the updates ahead of Cannes Lions 2026, according to details first reported by the company's newsroom. Chief Business Officer Lee Brown noted that "the future of discovery won't be driven by keywords alone" but by context, taste, and trusted recommendations.