When the Chatbot Becomes the Checkout Counter: AI Commerce Is Here Faster Than Anyone Expected

From ChatGPT's Shopify integration to Walmart's AI catalog, the distance between product discovery and purchase just collapsed—and brands have weeks, not years, to adapt

In late 2025, OpenAI quietly did something that should have every brand marketer’s full attention: it turned ChatGPT into a checkout counter. Users in the U.S. can now ask for product recommendations, see real-time inventory from over a million Shopify merchants, and complete purchases without ever leaving the chat. No website visit. No clicking through to Amazon. Just conversation, decision, buy.

Walmart followed weeks later, integrating its entire product catalog into ChatGPT’s shopping experience—270 million weekly customers now discoverable through conversational AI.

This isn’t a product feature announcement. It’s a structural shift in how consumers find and buy products. And the window for brands to prepare is measured in months, not years.

The Mechanics of Agentic Commerce

OpenAI calls this “agentic commerce”—where AI doesn’t just answer questions about products but actively facilitates the transaction. The technology stack powering it is worth understanding.

At the core is the Agentic Commerce Protocol, co-developed with Stripe. It creates a standardized language for AI agents to communicate with merchant systems about product feeds, inventory, checkout, and payments. When someone asks ChatGPT “What are the best eco-friendly yoga mats under $50?”, the system queries merchant databases in real-time, surfaces relevant products, and—if the user wants—processes the order without a single redirect.

Product results are organic and unsponsored, ranked on relevance. ChatGPT acts as the user’s agent—a digital personal shopper—passing information securely between consumer and merchant. Merchants pay a small commission on completed purchases, but neither consumers nor product rankings are affected.

The experience removes friction that has defined online shopping for decades: browsing, comparison shopping, adding to cart, creating accounts, entering payment info. For consumers already comfortable with AI assistants, the appeal is obvious.

Why This Time Is Different

Skeptics might point out that conversational commerce has been promised before. Voice shopping through Alexa never gained meaningful traction. Chatbots on retail sites have mostly been glorified FAQ bots.

Three factors make this different.

Scale. ChatGPT has 700 million weekly users. That’s not a niche audience—it’s a distribution channel that rivals the largest retail platforms. When Walmart announced its integration, it noted that customers would be able to discover and purchase from its inventory within those 700 million conversations.

Behavior shift. According to Adobe research, 39% of U.S. consumers who have used generative AI have already used it for online shopping, and 53% plan to do so. They’re using AI for product research and recommendations—exactly the use cases that lead naturally into transactions.

Infrastructure readiness. Shopify powers millions of merchants. Stripe handles payments globally. The Agentic Commerce Protocol is open-sourced, meaning any platform can build integrations. Microsoft’s Copilot launched its Merchant Program in April 2025. Perplexity introduced one-click purchasing through its search engine. The ecosystem is aligning around in-AI commerce as a standard rather than an experiment.

What Brands Must Do Now

For most brands, this represents a fundamental channel addition—one that requires different optimization than traditional search or paid social. Several priorities emerge.

Optimize product data for AI comprehension. Large language models don’t read product pages the way humans do. They need structured, consistent data: accurate titles, detailed descriptions (ChatGPT’s protocol allows 5,000 characters), complete attribute information, real-time inventory status, and consistent pricing across channels.

As one commerce consultant noted, AI systems need “very deep content, technical content of the product.” If your product feed was built for Google Shopping and hasn’t been updated since, it probably won’t perform well in AI discovery.

Maintain price consistency. AI agents comparison shopping on behalf of consumers will surface the cheapest option. If your pricing varies wildly across channels, you risk being deprioritized or, worse, training AI systems to see your brand as overpriced. One practitioner warned: “If you don’t have a constraint on your pricing model across channels, you run the risk of a future agentic bot...being able to find that and locate that.”

Invest in your own AI experience. Some brands are building AI concierges on their own sites—assistants that have richer context than general-purpose AI can provide. This creates a reason for customers to engage directly rather than transacting entirely within ChatGPT.

AKQA, the agency, has been helping luxury retailers develop exactly this: AI assistants refined with proprietary data that offer more personalized recommendations than what’s available through external platforms. As their CTO put it: “If you’re just building MCPs for the LLMs to access, you might just lose that connection with your end consumers.”

Monitor your AI visibility. Tools are emerging to track brand mentions and product appearances in AI responses. Ahrefs recently launched Brand Radar, which monitors ChatGPT, Perplexity, and soon Gemini. Understanding where you appear—and for what queries—is the new SEO.

The Publisher Problem Comes for Commerce

There’s a darker angle worth acknowledging. Publishers have spent the past two years worrying that AI will eliminate the need for consumers to visit their websites—the “zero-click search” phenomenon. That fear is now arriving for commerce.

If ChatGPT can answer “What running shoes should I buy?” and complete the transaction in the same interface, does the brand website matter? Does the carefully designed product detail page serve any purpose?

The early evidence is mixed. USA Today’s experience with its DeeperDive AI chatbot suggests that in-AI ads are possible but not yet delivering strongly contextual results. Taboola-powered recommendations in that chatbot often showed irrelevant sponsored content—a flashlight ad after a lipstick query, for example.

For brands with strong direct customer relationships, in-AI commerce might actually be good: another distribution point without the need to build new infrastructure. For brands that differentiated through website experience and customer service, the disintermediation is a genuine threat.

The Unit Economics Question

OpenAI’s financial situation complicates this story. The company projects losses of $44 billion through 2029 before reaching profitability. Revenue is growing rapidly—toward $200 billion by 2030, the company says—but costs are growing just as fast. Only 5% of ChatGPT’s 800 million users pay for subscriptions.

Commerce commissions could become a meaningful revenue stream, but the pressure to monetize raises questions about how product rankings will evolve. OpenAI says results are currently unsponsored and organic. Will that hold as the company faces pressure to close its cash burn?

For brands, this uncertainty means hedging. Build for AI commerce as a channel, but don’t bet everything on it remaining open and neutral. The history of platforms—from Facebook’s organic reach decline to Amazon’s pay-to-play search results—suggests that early openness often gives way to monetization that favors larger advertisers.

What Happens to the Funnel?

Traditional marketing funnels assumed distinct phases: awareness, consideration, intent, purchase. AI commerce collapses these into a single interaction. Someone asks “What’s a good anniversary gift for my wife who likes gardening?” and, in the same conversation, selects a product and buys it.

This changes the role of brand advertising. If the consideration and purchase phases happen inside an AI conversation, awareness-building becomes both more important (you need to be in the AI’s knowledge base) and harder to measure (how do you attribute a sale to brand advertising when the transaction happened entirely in chat?).

Content strategy shifts too. The question isn’t just whether your product page is optimized for Google—it’s whether your brand’s presence across the web has trained AI systems to recommend you for the right queries. That’s a much harder problem to solve.

The Honest Assessment

AI commerce is real, it’s here, and it will affect how consumers discover and buy products. But the hype should be tempered by practical realities: AI shopping still represents a small fraction of total commerce, the technology has friction points, and the platforms are still figuring out monetization.

The brands that will navigate this best are those that treat AI commerce as a new channel requiring specific optimization—not a replacement for everything else, and not something to ignore until it’s too big to catch up on.

The checkout counter has moved into the conversation. Whether that’s opportunity or threat depends entirely on how quickly you adjust.