Agentic checkout grows across major AI platforms

Agentic checkout is no longer an abstract idea discussed in demos. Over the past several months, major technology companies have begun rolling out tools that allow AI systems to guide, assist, or complete purchases inside chat and assistant interfaces. The growing sophistication of AI agents has made this possible, influencing the rise of agentic commerce. Companies like OpenAI, Microsoft, and Google have invested heavily in systems that can execute multi-step tasks rather than respond to single prompts.

AI-driven shopping is becoming part of mainstream commerce. Before agentic commerce, platforms already played a role in shopping. Consumers used Google Shopping ads to compare products before clicking through to retailer websites. Amazon’s marketplace checkout centralizes discovery and payment inside one platform. Social apps experimented with in-app checkout and buy buttons, allowing users to purchase without leaving feeds. What those systems shared was a clear separation between discovery and decision-making. Users still had to navigate pages, filters, and checkout forms themselves.

Agentic checkout changes that flow. Instead of directing users to storefronts or product pages, AI systems now manage parts of the decision process in conversation. This article brings those developments together to show who is offering what today, and how checkout actually works across platforms.

What agentic checkout means

Agentic checkout refers to AI systems that can handle multiple steps of a shopping journey. Rather than stopping at product discovery, these tools can compare options, surface prices, and move users closer to payment or complete the purchase inside the same interface.

This builds on earlier efforts to reduce friction, such as saved payment credentials and one-click checkout. The difference is that the AI agent becomes the interface itself. There is no separate product page acting as the center of the experience.

ChatGPT Instant Checkout

OpenAI was among the first to move in this space, partnering with Shopify in September 2025 to enable direct product sales inside ChatGPT through Instant Checkout. ChatGPT uses Instant Checkout through the Agentic Commerce Protocol (ACP), centralizing transactions within the chat interface. Merchants have limited post-purchase control, and OpenAI retains most customer data, including emails and behavioral signals. Payments and fulfillment are handled through partners such as Stripe and Shopify, but access to customer information for remarketing or loyalty efforts depends on what the platform chooses to share.

For example, a small clothing retailer integrated with ChatGPT could allow a customer to browse t-shirts, compare sizes, and complete a purchase entirely in-chat, but the retailer would not automatically capture the customer’s email for follow-up marketing unless the platform provides it.

Google’s Universal Commerce Protocol takes a different path

Google takes a different approach with its Universal Commerce Protocol (UCP), which allows merchants to maintain greater control over the shopping experience and retain first-party data, including emails and loyalty program information. While shopping interactions can start in Google Search, Gemini, or AI Mode, merchants continue to control checkout flows, customer data, and loyalty programs. 

This setup allows brands to retain first-party data and maintain direct relationships with shoppers, but it also requires merchants to invest in compatible infrastructure to support the protocol. For instance, a bookstore could let a customer ask an AI for recommendations, complete the checkout, and still own the customer relationship and loyalty points.

Microsoft’s Copilot Checkout expands ACP adoption

Microsoft is also competing for a share of this emerging market with Copilot Checkout, introduced in early January. Like Google, Copilot Checkout also relies on ACP, enabling in-chat transactions with partners such as PayPal, Shopify, Stripe, and Etsy. While merchants retain partial control, key data and workflow management remain platform-managed. The agentic is currently available in the U.S.

Perplexity’s Instant Buy follows the in-chat model

Perplexity has introduced its own variation with Instant Buy built into its conversational search interface. Like the other ACP-based experiences, Perplexity lets users discover products, compare options, and complete purchases without leaving the chat environment. Payments are handled through PayPal, and merchants integrated with Instant Buy may only see limited customer information, relying on the platform for fulfillment and post-sale engagement.

Amazon’s assistants extend an existing model

Amazon has been testing AI shopping assistants that help users discover products, compare options, and complete purchases within its ecosystem. While not explicitly branded as agentic checkout, the functionality aligns with the concept. Checkout remains fully controlled by Amazon, with transactions completing entirely inside its marketplace. This approach continues Amazon’s long-standing model of vertically integrating discovery, payment, and fulfillment.

These approaches taken by these companies highlight a key tradeoff in agentic commerce. On one hand, protocols built around full in-chat experiences, like ACP, reduce friction for shoppers, allowing them to buy without leaving the interface, but they shift control and valuable customer data toward the platform. 

On the other hand, protocols that emphasize merchant control, like Google’s UCP, require more technical integration but let merchants maintain first-party relationships, loyalty programs, and marketing data. UCP prioritizes merchant ownership and flexibility.

Where the industry stands now

Agentic checkout has moved beyond early experiments into live implementations across major AI platforms. OpenAI, Google, Microsoft, Perplexity, and Amazon each offer tools that manage part of the shopping journey within AI chats.

The industry is still fragmented. Some platforms, like OpenAI, allow checkout entirely inside conversational experiences. Others, like Google, focus on routing users to merchant sites while standardizing the technical connection between AI agents and sellers.

Agentic checkout remains in an early stage, building on earlier digital commerce developments. Platforms are currently testing these features and gradually expanding availability as they compete for a position in the growing e-commerce market.

‍