Why Walmart and OpenAI Are Revolutionizing Their Shopping Partnership

The chatbot has been designed to be adaptable, particularly with the latest integrations. “It can undergo minor adjustments in appearance, ensuring it seamlessly blends into various environments,” Danker explains.

Shopping Shift

The updated Walmart experience is part of a larger strategy for OpenAI to enable checkouts within integrated applications, as reported by the Information earlier this month, though no specific reason for this shift was given. Danker discussed this transition at the Morgan Stanley investor conference this month but did not provide supporting data.

OpenAI representative Taya Christianson states that the company aims to enhance user product research while granting merchants greater authority over the checkout process. “We appreciate our partners for learning alongside us,” she added.

Walmart has excluded certain products from Instant Checkout, recognizing that “the single-item checkout experience can be detrimental” in some cases, according to Danker. For example, purchasing a TV often necessitates buying accessories like HDMI cables. On its website, Walmart can suggest bundles to alleviate installation issues, which Danker indicates Sparky will also be able to do in chatbots.

Retailers were keen to collaborate on Instant Checkout since the alternative at that time required directing ChatGPT users to their websites. Walmart believes the Sparky experience will feel more “seamless,” as users can continue chatting and fine-tuning their orders without needing to re-enter their saved payment and delivery details.

Sparky has faced criticism from individuals claiming to work for Walmart on Reddit, and it’s challenging to find positive testimonials on social media. However, the company reports that half of Walmart app users have interacted with it. While users typically search for everyday items like milk and bananas, they turn to Sparky for more exotic products or complex queries. Recently, Walmart US CEO David Guggina noted that Sparky users spend approximately 35 percent more per order than other shoppers.

Danker admits that Sparky is slow and often yields subpar responses, causing some consumers to view it as unreliable. He states that this year’s focus is on training Sparky to be more proactive, enhancing its understanding of individual shoppers, and expanding its usefulness across Walmart’s various departments, including the pharmacy.

While Walmart is promoting Sparky in other areas, it does not intend to restrict other AI tools from shopping on its website. In contrast, Amazon recently secured a temporary court order preventing Perplexity’s automated system from impersonating humans to make purchases. Danker expresses Walmart’s commitment to supporting whatever tools customers choose, provided they ensure a good experience without erroneous orders, unexpected charges, or excessive reliance on customer service.

“We don’t want to dictate the exact journey every customer will take,” he remarks. “We don’t aim to block options based on speculative or hypothetical concerns.”

Regarding consumers’ trust in AI for shopping, Danker is willing to make predictions. “The notion that everything will become automated may be somewhat unrealistic,” he remarks. “People are still enthusiastic about shopping for clothes, home items, and their children.” Walmart seeks to empower users, now with Sparky assisting them in more scenarios.

This is an edition of Will Knight’s AI Lab newsletter. Read previous newsletters here.