Why agentic shopping adoption remains low despite the hype

The craze around AI in retail is real. And anyone looking for clear answers on how to execute an AI playbook needs to realize that agent-assisted shopping will take time to catch on.

Take OpenAI’s instant checkout feature, which didn’t land, primarily because people weren’t ready for it. However, it wasn’t necessarily a failure.

OpenAI launched its Instant Checkout with a handful of merchants on Etsy and Shopify in September 2025, and then rolled it back, letting retailers handle their own checkout, in March.

“We’ve found that the initial version of Instant Checkout did not offer the level of flexibility that we aspire to provide, so we’re allowing merchants to use their own checkout experiences while we focus our efforts on product discovery,” OpenAI said in a blog post.

During Retail Brew’s recent virtual event, Andy Szanger, director of strategy at CDW, said the issue wasn’t that the OpenAI checkout feature was broken or unsuccessful; it had more to do with the timing of the rollout. “It is really more of a failure of sequence, and more precisely, a failure to understand the difference between what technology can do and what consumers are ready for it to do,” Szanger said.

Szanger referenced a recent book by former OpenAI executive Zack Kass called The Next Renaissance: AI and the Expansion of Human Potential, which delves into the impact of AI on society at large, in unpacking OpenAI’s shopping pivot.

“One of the things he spoke about in the book is what he calls the gap between the technological threshold and the societal threshold,” Szanger said. “He argues that the path to the future will be driven not by scientific progress, but by the culture and societal progress. I think that framing explains the change in approach to checkout with OpenAI pretty well.”

Data from logistics platform Radial paints a similar picture. Though nearly 6 in 10 consumers (58%) said they’d be open to ordering through an AI assistant, only 6% have actually tried it, per Radial.

Insert agent: AI-powered shopping could hit $1 trillion in US consumer retail sales by 2030, according to a report by ICSC and McKinsey & Company.

Department store chain Nordstrom, Szanger said, is using agentic technology to support employees, not replace them. For instance, Nordstrom is using the technology to provide accurate delivery times, which is the top reason customers reach out to Nordstrom customer service.

“Human interaction is not an overhead; it is the product,” Szanger said.

AI agents make the most sense for simple, repeat purchases, like everyday items you buy over and over again because they’re low-risk and easy to automate, Szanger added. But for bigger, more personal moments like a first purchase, a luxury item, or helping a loyal customer with a problem or return, people still want real human interaction and support.

Retailers need to make it easy for customers to find products and feel confident when buying them, while using AI to improve the experience without taking humans out of the equation, Szanger said.

“[Nordstrom has] explicitly stated multiple times that their people are their competitive advantage,” he said. “That’s a strategic choice. And I think that we need to keep that in mind as we think about where agents sit in this process.”

Ultimately, agentic commerce is shaping up as a connected experience, Szanger said.

“Where you have that transaction, you want to then make sure that the retailer owns that process,” he said. “Discovery happens in the AI platform. The checkout happens on the commerce platform.”