OpenAI Pushes Into Personalized Shopping
OpenAI is expanding its AI search experience with highly personalized product recommendations. The new features integrate images, reviews, and direct purchase links, signaling the company's ambition to build context-rich recommendation pipelines that blur the line between information discovery and commerce.
OpenAI's new shopping features are part of a broader strategy to monetize its massive user base beyond subscriptions by embedding e-commerce directly into ChatGPT. The company plans to take a commission, reportedly around 2%, on transactions that originate from its product discovery and research tools. This move diversifies OpenAI's revenue streams as it reportedly faced a $5 billion loss in 2024 and aims for significant revenue growth. The "Shopping Research" tool allows users to describe their needs in natural language to receive personalized buyer's guides. ChatGPT asks clarifying questions about budget and features, then scans online retailers for real-time pricing, reviews, and availability to inform its recommendations. The feature is powered by a purpose-trained version of the GPT-5 mini model and is designed to disregard low-quality websites. For direct purchasing, the "Instant Checkout" feature allows users to buy products without leaving the chat interface. This is facilitated by the "Agentic Commerce Protocol" (ACP), an open standard co-developed with Stripe that connects ChatGPT to payment systems and allows merchants to link their stores. Major retailers like Walmart and platforms such as Shopify and Etsy have already partnered with OpenAI to make their products available through this feature. This push into e-commerce positions OpenAI in the emerging field of "agentic commerce," where AI assistants act on behalf of users to complete tasks like shopping. The company is entering a competitive space, with giants like Google and Amazon also developing their own AI-powered shopping agents and conversational commerce capabilities. For instance, Amazon has features like "Buy for Me," and Google is integrating similar agentic capabilities into its search and ad ecosystems. While the new tools aim to streamline the path from discovery to purchase, they also raise questions about data privacy and algorithmic bias. OpenAI has stated that user chats are not shared with retailers and that product recommendations are organic and unsponsored. However, the system's reliance on training data means it could mirror existing internet hierarchies, potentially making it harder for smaller brands to gain visibility.
