Splitit said it is backing Google’s Universal Commerce Protocol (UCP), an open standard intended to support agentic commerce, through which AI systems can discover products and complete purchases on behalf of consumers. While autonomous shopping via AI agents is advancing quickly, though, some initiatives show adoption may be uneven.
The card-linked installment payments provider said supporting the protocol, which Google unveiled in January, could allow merchants to incorporate flexible payment options into AI-driven shopping flows. UCP was developed with participation from several major retailers and commerce platforms, including Shopify, Target, Walmart, Etsy and Wayfair, and is designed to standardize how AI agents interact with merchant systems across the shopping lifecycle. Many payments providers are also on board, including Visa, Mastercard, Paypal and Splitit competitor Klarna.
Splitit framed installment payments as a key component of automated purchasing environments, where AI systems may handle product discovery and transaction execution.
“Payment flexibility is quickly becoming a core consideration in AI-powered commerce,” said James Wray, head of business development at Splitit. “As AI agents increasingly discover products for consumers, they need payment solutions that deliver certainty. Splitit’s card-linked installments remove approval uncertainty and friction, allowing merchants to turn browsers into buyers.”
The company’s support for UCP reflects a broader industry push to create infrastructure for AI-driven commerce. Technology providers, retailers and payment firms are experimenting with standards and protocols intended to enable autonomous purchasing workflows that span product search, payment authorization and fulfillment.
At the same time, recent developments suggest adoption may unfold unevenly.
Just months after launching a high-profile initiative with Walmart, Etsy and Shopify to enable purchases directly inside ChatGPT, OpenAI has paused its native “Instant Checkout” capability and shifted transactions back to retailer apps and websites. According to a report in The Information ($), users and merchants both displayed discomfort with the product.
Most users were comfortable asking ChatGPT to recommend products or compare options but rarely completed purchases within the chat interface itself, the report indicated. The findings suggest that AI tools may play a stronger role in product discovery than in the final transaction stage, at least in the near term.
Operational and infrastructure challenges also contributed to the shift. The complexity of synchronizing product catalogs, managing taxes and preventing fraud across large ecommerce ecosystems made fully automated checkout within ChatGPT difficult to scale. In addition, only a small number of merchants had implemented the necessary integrations during the early rollout, limiting the feature’s reach.
These challenges illustrate the broader dynamics likely to shape agentic commerce adoption. While many retailers and technology firms are experimenting with AI-driven shopping, the transition from recommendation engines to autonomous purchasing systems requires standardized product data, reliable payment infrastructure and consumer trust in automated transactions.
Industry observers increasingly view this progression as iterative rather than linear. Early experiments—such as conversational checkout—may be scaled back or redesigned while companies refine the underlying infrastructure needed to support AI-mediated commerce.
For companies like Splitit, that evolution underscores the importance of building foundational components such as payment flexibility and interoperability. But the recent pullback from in-chat checkout also highlights that the path toward fully automated retail will likely involve cycles of experimentation, retrenchment and renewed development before agentic commerce becomes a mainstream channel.
