Meta launched an experimental AI shopping feature called Shopping research inside Meta AI on the web, rolling it out to select US desktop users and featuring a query button labeled "Shopping research." The assistant responds to product queries by presenting a carousel of product images, prices, brand details and links to retailer sites, with a brief rationale for each recommendation.
The tool can tailor suggestions when Meta AI has access to basic profile signals, such as gender and location, and it showed region-appropriate results for testers in New York. Users cannot complete purchases inside Meta AI; the feature directs shoppers to external e-commerce pages to buy items. This rollout follows wider industry adoption of AI shopping assistants and signals Meta’s push to embed commerce-oriented agents into its AI layer, helping consumers discover options faster while preserving retailer checkout flows.
Personalized AI Shopping Features
Meta Launches Meta AI Shopping Research Experimental Tool
Trend Themes
-
AI-powered Visual Product Carousels — A move toward image-first recommendation displays that combine visuals, pricing and context to streamline discovery and influence purchase intent.
-
Profile-driven Personalized Recommendations — Greater reliance on basic profile signals like location and gender to surface region-appropriate and demographic-relevant product suggestions.
-
Commerce-embedded Conversational Agents — The embedding of commerce-aware assistants into conversational layers that present curated options while preserving external checkout flows.
Industry Implications
-
E-commerce Platforms — A redefinition of product discovery as platforms incorporate AI-curated carousels that can shift traffic patterns and merchant visibility.
-
Social Media Advertising — An evolution in ad formats where conversational AI recommendations blur lines between organic discovery and sponsored promotion.
-
Retail Technology — Enhanced back-end tooling and feed integration needs driven by AI agents that require real-time pricing, inventory and attribution data.