Pinterest launched Ask Pinterest on Wednesday, a standalone experimental app that puts a conversational, chatbot-style interface on top of its product catalog and personal recommendation engine. The app is limited-access at launch and is the company's first serious attempt to package its Taste Graph — the internal dataset mapping users to interests and aesthetics — as something closer to an AI shopping assistant than a search box. Pinterest paired the release with a new Model Context Protocol (MCP) server for advertisers, an AI assistant in beta inside its U.S. Ads Manager, and a global rollout of Performance+ creative, a model that picks the best-performing ad variant per impression.
The timing is deliberate. Cannes Lions, the adtech industry's annual gathering, opens this week with AI for marketers as the dominant theme, and Pinterest is using the moment to position itself as something other than a passive image board competing with search engines for shopper attention. Ask Pinterest is the consumer-facing piece; the MCP server and Performance+ creative are the pitch to brands.
Ask Pinterest is built for queries the main app handles poorly — multi-step, context-heavy requests like planning a dinner party or furnishing a room over several months. The app retains context across sessions and can pull from a user's saved Pins and Boards to personalize what it returns. Pinterest describes it as an experiment that could eventually inform AI features in the flagship app, but breaking it out as a separate product lets the company iterate without disrupting the existing experience for hundreds of millions of users.
Key facts
- 01Pinterest launched Ask Pinterest, a standalone conversational AI shopping app, on June 17, 2026, with limited initial access.
- 02The app is powered by Pinterest's Taste Graph, the internal dataset mapping users to their interests and aesthetics.
- 03Pinterest also rolled out a Model Context Protocol (MCP) server letting advertisers manage campaigns through third-party agentic tools.
- 04A new Performance+ creative AI model went live globally to auto-select the best-performing ad creative per impression.
- 05The announcements landed days ahead of Cannes Lions, where adtech is concentrating on AI for marketers.
The strategic question Ask Pinterest answers is whether Pinterest becomes a destination for AI-driven shopping or a data source other AI services license. The company has chosen the first path. Rather than licensing the Taste Graph to ChatGPT, Google, or Meta, Pinterest is training its own models on its own data and shipping its own conversational surface. That's a meaningful divergence from the syndication strategy other content platforms have taken with frontier labs.
Lee Brown, Pinterest's Chief Business Officer, framed the bet around discovery rather than search.
“the future of discovery won't be driven by keywords alone. It will be shaped by context, taste, and trusted recommendations”— Lee Brown, Pinterest Chief Business Officer
Brown also called the area one where Pinterest has a "unique advantage" — a defensible claim given the Taste Graph has been collecting interest and aesthetic signals since the platform's earliest days, well before generative AI made that kind of training data valuable.
The advertiser-facing announcements are arguably the more immediate revenue story. The Pinterest MCP server lets brands manage and monitor campaigns through third-party agentic tools using a standardized protocol — the same Model Context Protocol that Anthropic open-sourced and that has since been adopted across the agent stack. For an ad platform, exposing campaign management through MCP means agencies and brands can wire Pinterest into whatever agent framework they already use, rather than learning Pinterest's bespoke APIs.
Performance+ creative, meanwhile, is a direct response to where ad spend is moving. Meta, Google, and TikTok have all shipped models that auto-generate or auto-select ad creative based on predicted performance, and brands increasingly expect that capability as table stakes. Pinterest shipping the feature globally — rather than U.S.-first — suggests the company sees creative selection as competitive ground it cannot cede.
The competitive picture is crowded. Google has wired AI into shopping search, price tracking, and checkout. ChatGPT has tested agentic shopping flows. Meta and Shopify both have agentic commerce projects in flight. Pinterest's wedge is that its users arrive already in discovery mode — saving, planning, considering — rather than searching with transactional intent. Whether that translates into a conversational app people return to, versus one they try once, is what the limited-access rollout is designed to test.
The risks are real. Standalone companion apps from incumbent platforms have a poor track record of retention, and Ask Pinterest will need to demonstrate it can answer multi-step queries better than a general-purpose chatbot that has also been shown the user's Pinterest activity through a connector. Pinterest has not disclosed engagement targets, geographic expansion plans, or how the limited-access waitlist will scale.
For Pinterest, Ask Pinterest is a hedge against becoming infrastructure for someone else's AI shopping experience. The company is betting its Taste Graph, accumulated over more than a decade, is more valuable as the engine inside a Pinterest-branded conversational product than as training data licensed to OpenAI or Google. If agentic commerce becomes the default way consumers shop online, that bet decides whether Pinterest captures the transaction or merely supplies the context. The MCP server is the tell — Pinterest wants advertisers wired in directly, not routed through an intermediary.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
