Agentic commerce may have been oversold

OpenAI’s instant checkout experiment shows how far AI still is from autonomous shopping

The promise of “agentic commerce” has captured the imagination of both Silicon Valley and the retail industry. The idea is simple but powerful: AI agents that not only recommend products but also complete purchases on behalf of users. Yet OpenAI’s recent decision to discontinue its Instant Checkout experiment suggests the technology may not be as transformative as early headlines implied. The feature worked technically, but in practice it relied heavily on user prompts, confirmations, and traditional checkout mechanics — raising questions about whether true agentic shopping is still a distant goal.

Key takeaways

• Agentic commerce expectations ran ahead of reality: The Instant Checkout feature was often described as a breakthrough in autonomous shopping, but the actual experience remained largely manual and dependent on user instructions.
• Users were still doing most of the work: Rather than delegating purchasing decisions to an AI agent, users still had to guide the process, confirm choices, and complete key steps — making the interaction feel closer to assisted search than autonomous commerce.
• The interface change was incremental, not revolutionary: Completing a purchase inside a chat interface did not significantly improve the buying experience compared with established e-commerce platforms, reducing its perceived value.
• Retail infrastructure remains a major barrier: Payments, returns, logistics, and customer service still rely on existing retailer systems, limiting how autonomous an AI purchasing agent can realistically become.
• Trust remains unresolved: Many consumers are not yet comfortable allowing an AI system to make purchasing decisions on their behalf, particularly for higher-value or more complex purchases.
• The long-term idea is still viable: Despite the failed experiment, the concept of AI agents managing discovery, comparison, and eventually transactions remains strategically important for the future of digital commerce.
• Real agentic commerce will require deeper integration: Truly autonomous purchasing would require AI systems to integrate tightly with payment networks, retailer APIs, logistics systems, and personal preference data — infrastructure that is still evolving.

Source: The Drum

Enjoying this analysis? Hospitality.today delivers daily insights on hotel distribution, AI trends, and travel commerce — straight to your inbox. Subscribe for free at Hospitality.today →