OpenAI is restructuring how ChatGPT handles commerce, shifting focus from direct transactions to enhanced product discovery. The move follows significant challenges with its Instant Checkout feature, which launched in 2025 but stumbled over merchant onboarding, data accuracy, and complex cart management.

Instead of forcing a unified checkout layer, OpenAI now prioritizes search relevance and retrieval quality. Users can upload images or describe items with specific constraints like budget. Under the hood, engineering teams have optimized retrieval speed and product coverage to ensure results reflect current inventory. Major retailers including Target and Sephora are already feeding product data into the system.

For transactions, the strategy favors delegation through custom apps. Merchants can build dedicated integrations within ChatGPT, retaining control over loyalty programs and payment flows. Walmart recently deployed an in-app service handling links and payments directly. This architecture reduces the burden on OpenAI to maintain complex stateful shopping carts while leveraging partner infrastructure for the final mile.

Shopify is aligning with this shift, upgrading its ChatGPT integration to allow storefront connections via an in-app browser. They also introduced an Agentic Plan, enabling merchants without dedicated storefronts to surface products across AI platforms using Shopify's tools. This suggests a broader industry move toward specialized agents rather than monolithic commerce models.

The pivot acknowledges the engineering complexity of unified commerce. By separating discovery from transaction execution, OpenAI aims to deliver reliable recommendations while letting retailers manage the fragile logic of checkout. It signals a maturing approach to agentic workflows where the AI orchestrates rather than owns the transaction.

Source: CNBC