OpenAI is reportedly building a payment and shopping system directly inside ChatGPT. It sounds big. But here’s the problem: no one has cracked native checkout — not even the trillion-dollar platforms.

Meta spent years pushing Instagram and Facebook Shops and is now quietly scaling it back. TikTok Shop is flooded with low-quality goods and hasn’t shifted long-term buying habits in Western markets. Google tried “Buy on Google” and it flopped. Even Pinterest — one of the highest product-intent platforms in existence — still struggles to convert inspiration into actual purchases.

The data backs this up. According to Insider Intelligence, just 7% of U.S. adults have ever completed a purchase via social commerce. Even fewer do it regularly.

I’ve worked inside these ecosystems at BigCommerce, Pinterest, and Meta. I’ve seen firsthand how hard it is to shift user behavior at scale. People may ask ChatGPT what to buy, but most still want to read reviews, compare prices, and check out on platforms they already trust.

The pattern is consistent across every platform that has tried this: the discovery experience is strong, the checkout experience bleeds users. Trust, familiarity, loyalty programs, branded credit cards, customer service — these are the reasons people return to retailer-owned checkout flows. A chatbot sitting between the user and the purchase doesn’t solve any of those problems.

OpenAI may want to be the next Google and Amazon in one. But if platforms with billions in resources and hundreds of millions of daily active users couldn’t scale native checkout, it’s hard to see how a chat interface gets there first.

This feels less like product-market fit and more like AI hype on autopilot.

Originally published on LinkedIn, October 2025.