Walmart and OpenAI are abandoning their initial approach to AI-powered shopping after the “Instant Checkout” feature within ChatGPT proved largely ineffective. Conversion rates were three times lower for purchases made directly inside the chatbot compared to those requiring a click-out to Walmart’s website, according to Walmart’s design and product lead, Daniel Danker.
The early experiment, launched in November, aimed to integrate ecommerce directly into OpenAI’s chatbot interface, allowing users to order products without leaving ChatGPT. The idea was that AI agents could drive revenue by earning commissions on purchases. However, the clunky experience – forcing individual checkouts for each item – deterred customers who feared receiving multiple shipments for a single order.
The Shift to Embedded Chatbots
Rather than fixing the broken “Instant Checkout,” Walmart and OpenAI are moving toward embedding Walmart’s own chatbot, Sparky, directly within ChatGPT and Google’s Gemini. This solves the core problem: consumers prefer consolidated checkouts. Sparky will sync with existing Walmart shopping carts, allowing for a more natural shopping flow.
The move comes as OpenAI refocuses on improving product research while giving merchants greater control over the checkout process. OpenAI spokesperson Taya Christianson stated that the company wants to “focus on improvements to help users research products, while giving merchants more control over checkout.”
Why This Matters: The Future of AI Commerce
The failure of Instant Checkout underscores the challenges of fully automated ecommerce. While AI can assist with shopping, it cannot replace the human desire for control and convenience. The new model addresses this by bringing the Walmart experience into existing chatbots, rather than forcing customers into a fragmented system.
Walmart has excluded some products from Instant Checkout because it knew “the single-item checkout experience is detrimental” in some cases. For instance, when someone buys a TV, they likely need to buy accessories like HDMI cables. On its website, Walmart can nudge shoppers to buy a bundle to avoid a frustrating installation experience. Through Sparky, Walmart will be able to replicate that in chatbots.
Walmart’s Strategic Advantage
Walmart sees ChatGPT as a significant source of new customers, doubling the rate compared to traditional search engines. The chatbot attracts a different audience than typical Walmart shoppers, but the retailer’s broad selection and pricing mean its products are frequently recommended within AI-driven conversations.
The company’s investment in Sparky, developed using both open-source and proprietary AI models, is also paying off: half of Walmart app users have engaged with the chatbot, with Sparky users spending 35 percent more per order than average shoppers.
Despite acknowledging Sparky’s current limitations (slow responses, occasional unreliability), Walmart is prioritizing its improvement. The goal is to train the chatbot to proactively learn individual shopping habits and expand its utility across departments, including pharmacy services.
Walmart remains open to other AI agents shopping on its website, unlike Amazon, which recently sought a court order to block automated purchasing tools. Walmart’s stance reflects a willingness to adapt to evolving consumer behavior, even if full automation remains distant.
“This idea that it will all become automated might be a little bit far-fetched,” says Daniel Danker. “People do get excited about shopping for clothes, for their home, for their children.”
The shift from failed instant checkout to chatbot integration signals a pragmatic approach to AI commerce. Walmart is not trying to replace shopping; it’s adapting to meet customers where they are, in a world increasingly shaped by conversational AI.
