Meta's AI Shopping Push: Another Carousel Nobody Asked For
A Meta AI chatbot interface floating mid-air displaying an oversized product carousel of generic white sneakers with price tags, suspended above an empty suburban living room with a single unused sofa, highlighting th...📷 AI illustration
- ★Carousel product images with prices
- ★Competing with ChatGPT and Gemini
- ★No clear timeline for rollout
Meta is testing an AI-powered shopping feature inside its Meta AI chatbot that serves up product suggestions as image carousels complete with brand names, prices, and website links. The test, first reported by The Decoder, positions Meta to compete directly with similar shopping tools already deployed by OpenAI's ChatGPT and Google's Gemini.
This is not Meta's first e-commerce experiment. The company has spent years building and largely abandoning shopping features across Facebook and Instagram, from Marketplace to live shopping to in-app checkout. Each iteration promised to transform social browsing into purchasing; most ended up as infrastructure for ads rather than genuine retail utility. The AI chatbot integration at least acknowledges a shifted landscape: users increasingly expect conversational interfaces for product discovery, not static feeds.
The carousel format itself reveals the tension. It mirrors the visual language of Instagram shopping and Amazon product listings rather than breaking new ground. Meta's AI generates suggestions, but the presentation suggests the company is grafting AI onto existing commerce plumbing rather than reimagining how people actually want to shop.
The gap between social commerce and actual utility keeps widening
A cluttered desktop with multiple open browser tabs showing abandoned Meta shopping experiments: Facebook Marketplace, Instagram Live Shopping, and a defunct 'Shop' tab, all faded and inactive, contrasting the current...📷 AI illustration
The competitive pressure is real, if oddly timed. ChatGPT and Gemini have both introduced shopping capabilities that let users compare products, read reviews, and follow purchase links through natural conversation. Meta's advantage, if it materializes, would be its vast repository of user behavior data and established merchant relationships through Facebook Shops. Whether that translates into better recommendations remains unproven.
What distinguishes this test from previous Meta commerce efforts is the explicit AI framing. Previous initiatives leaned on social proof and influencer marketing; this one leans on algorithmic authority. That shift matters because it changes the accountability structure. When an influencer recommends a dud product, blame attaches to a person. When an AI hallucinates a price or invents a product feature, the failure mode is harder to trace and potentially more damaging to trust.
Early signals suggest the feature remains limited to select users with no disclosed geographic or demographic scope. Meta has not clarified whether results draw from its own merchant network or scrape broader web inventory, a distinction with significant implications for both accuracy and revenue model. The absence of detail is itself a signal: this remains experimental enough that the company would rather not promise specifics it may need to retract.