AI Shopping Discovery Shifts From Keywords to Conversations
AI shopping discovery refers to the use of machine-learning systems, conversational interfaces, and recommendation models that help people find, compare, and buy products through natural language, contextual understanding, and personalised suggestions rather than relying only on traditional keyword-based search results. That shift is now visible across major platforms. Pinterest is testing Ask Pinterest, a conversational product search app. Snap is rolling out Smart Assistant and AI agents for campaign setup and chat-based ads. Microsoft Advertising has introduced Product Explorer to make product catalog management more searchable and action-focused. Together, these moves signal a new phase where discovery becomes more dialog-driven and less dependent on typed product keywords. For marketers, AI shopping discovery is no longer experimental; it is the new front end for both product search and smart campaign optimization.
Pinterest’s Ask Pinterest and the Rise of Conversational Product Search
Pinterest’s Ask Pinterest is a limited-access experimental app that turns the platform’s visual inspiration into conversational product discovery. Instead of typing “green sofa” and sifting through Pins, people can describe complex plans—like furnishing a living room over months—and receive tailored ideas grounded in their taste. The app uses natural language to power conversational product search and pulls on Pinterest’s Taste Graph, which maps people’s interests and aesthetics, plus saved Pins and Boards, to personalise responses. It also keeps context across sessions so ongoing projects do not have to restart from scratch. Pinterest is running the app as a standalone web experience, allowing quick iteration without reshaping its main interface. The company has framed this as proof that “the future of discovery won’t be driven by keywords alone,” and as a way to turn inspiration into shoppable, AI-guided journeys.
Behind Pinterest’s AI Stack: Business Assistant, MCP and Performance+
Alongside Ask Pinterest, the company is upgrading its AI advertising tools to close the gap between discovery and paid performance. Business Assistant, in closed beta inside Ads Manager and mobile, acts as a guide that surfaces trends, content opportunities and campaign health through visual summaries instead of long text replies. Pinterest Model Context Protocol (MCP) connects Pinterest campaign, analytics and keyword insights to external agents and copilots used by agencies, built with partners such as PMG, Pacvue, Dentsu, Havas, Innovid by Mediaocean and Jump450. Performance+ creative now includes an AI model that selects the best-performing ad variant for each impression; Pinterest reported that the new model increased click volume by 7.5% over its previous single-variant approach. Combined, these tools push toward smart campaign optimization where setup, analysis and creative testing are handled by AI layers that stay in sync with conversational shopping behaviour.
Snap’s Smart Assistant and Conversational Ads for Shopping
Snap is expanding AI across the entire Snapchat advertising workflow. Snap Smart Assistant lets advertisers describe goals in plain language and receive recommendations on campaign objectives, audience targeting and optimization settings, along with ongoing health checks and suggested next steps. The company is also opening an MCP server so third-party AI agents can plan and optimise campaigns that include Snapchat inventory. On the consumer side, Snap is turning chat into a shopping surface through AI Sponsored Snaps in Chat, where branded AI agents answer questions, recommend products and help with purchase decisions; Experian is among early testers. For shopping brands, updated Dynamic Product Ads add AI-driven recommendation models that combine behaviour, product preferences, funnel position and real-time intent to deliver more relevant items. These moves show how conversational product search and AI shopping discovery are spreading from feed ads into one-to-one chat environments.
Microsoft Product Explorer Brings Searchable Insight to Ad Catalogs
While Pinterest and Snap focus on front-end conversations, Microsoft Advertising is targeting the operational backbone of AI advertising tools. Its new Product Explorer feature gives advertisers a searchable, filterable view of their whole product catalog inside the ad platform, currently for accounts with fewer than 100,000 SKUs. Marketers can filter by traits such as title, product ID, brand, price, category, condition, availability and custom labels, then add performance metrics like impressions, clicks, conversions, spend, CTR and conversion rate. Title filters include options such as contains, does not contain, starts with and equals, making it easier to isolate product sets. Product Explorer also highlights which products are serving and performing, with a Recommended Actions tab suggesting fixes to bring items back into rotation, and supports exports for offline work. This kind of structured insight is a quieter but essential layer for smart campaign optimization.
