Why OpenAI’s Sudden Move Against In-Chat Checkout Could Shake Up Retail Forever
Ever wonder why ChatGPT—this shiny, futuristic shopping buddy—decided it’s not so great at ringing up your purchase? Turns out, navigating the wild web of commerce inside someone else’s sandbox is a high-wire act; one wrong step, and poof—no sale. OpenAI’s recent pivot away from completing purchases within ChatGPT might seem like a stumble, but really, it’s a wake-up call. The AI excels at discovery, comparison, and sparking user intent—but when it comes to sealing the deal, the baton must pass to retailers and their trusty apps. The real magic—and challenge—now lies in how well these handoffs happen. Think deep linking and attribution, folks: without the right tech plumbing, those clicks vanish into the void, and so does your hard-earned revenue. So, if you’re in retail, ask yourself—are you ready to bridge that glaring gap between AI-led discovery and the “cha-ching” moment? Because billions are waiting on the other side. LEARN MORE
Lauren Newman is CRO at Button, an AI-powered commerce platform optimizing creator and affiliate marketing performance.
The thing about building inside someone else’s platform is that you’re always one pivot away from losing your lease. Last month, OpenAI confirmed what many in retail tech had quietly suspected: that completing a purchase inside ChatGPT was never as seamless as the vision suggested. It doesn’t match the behaviors of users, and OpenAI figured this out before they went too deeply into commerce. The official framing, per OpenAI, is that the company is evolving its commerce strategy to better meet merchants and users where they are.
What that really means is users would browse ChatGPT but buy elsewhere. Sound familiar? This is a standard buying behavior exhibited across the web and one that makes attribution difficult. The AI was a brilliant shopping assistant. It was a lousy cash register. That distinction matters enormously.
The Gap Between Discovery And Purchase
OpenAI’s exit from in-chat checkout doesn’t signal the end of AI-powered commerce. It signals a maturation of where AI actually exists in the purchase funnel. Discovery. Comparison. Intent. That is where AI is winning. In fact, it’s taking substantive traffic and influence from traditional search. According to Adobe, AI-driven traffic to U.S. retail sites surged nearly 670% in the first month of the 2025 holiday season. Consumers are using tools like ChatGPT not to check out, but to research and decide. The purchase happens elsewhere, and almost always on mobile, increasingly within a native app. This is the attribution crisis that looms over many retailers.
A shopper asks ChatGPT which running shoe handles trail surfaces best. ChatGPT gives a considered, personalized answer and sends them off. The retailer gets a click. But what happens next? Does that click land on the right product page inside the app? Does it carry attribution? Does it account for whether the shopper is iOS or Android, whether they have the app installed or whether the link resolves properly? Hint: the answer is no.
This Is Exactly What Deep Linking Was Built For
Long before AI assistants were recommending products, mobile commerce had a plumbing problem. Publishers, affiliates and creators were driving enormous shopping intent, but the links that routed shoppers to retailers were fragile. They were at the wrong destination. Missing app context. Broken attribution. Lost commission. The user bounced. The sale evaporated.
Platforms must use AI-powered dynamic deep linking to ensure every tap routes the consumer to the best possible destination. They need to recognize the native app when it’s installed, the mobile web when it isn’t, and the right product page regardless of environment or device.
When ChatGPT routes a shopper to a retailer app rather than hosting checkout itself, the quality of that routing handoff becomes the critical variable. Every link that goes to the right destination in the right app context, with proper attribution and seamless experience, is revenue captured. Every link that doesn’t is revenue walked out the door.
Context Collapse: The Real Risk At The Handoff
ChatGPT is a deeply personalized, conversational shopping advisor. When it tells you a Hoka Speedgoat 6 in wide width is the right shoe for your knee issues, it’s not sending you a generic page. If the retailer lands that shopper on a generic running shoe category page, the context and conversion probability collapse instantly.
This is both a technology and strategy problem. On the technology side, solutions that handle intelligent destination routing based on product-level data, user state and app context are table stakes for any retailer serious about capturing AI-referred traffic. On the strategy side, retailers need to ensure their product feeds, the structured data that surfaces in AI-powered search and recommendation engines, are current, rich and optimized.
The Attribution Problem Doesn’t Disappear; It Gets Bigger
There is a measurement crisis quietly building at the intersection of AI discovery and retail conversion. When a shopper finds a product through ChatGPT and then makes a purchase in a retailer’s app, who gets credit? The AI platform? The affiliate? The creator who originally surfaced the recommendation? The retailer’s own SEO?
OpenAI’s retreat from checkout is not a story about AI losing to retail. It is about AI genuinely adding value where the established infrastructure of retail, trusted apps, saved payment methods, loyalty programs and real-time inventory still has the structural advantage.
The winning model for the next phase of AI commerce looks something like this: AI owns discovery and recommendation. Retailers own the transaction and the relationship. The companies that build the connective tissue between those two layers—dynamic deep linking, intelligent routing, attribution, product feed optimization—own the margins.
If you are a retailer, this is not the moment to observe from the sidelines. Three things deserve immediate attention.
First, audit your deep-linking infrastructure. When AI platforms route shoppers to your app, do those links actually work—across devices, app-installed and non-installed states, and varying user contexts?
Second, invest in your product feed as aggressively as you once invested in search. Structured product data, clean descriptions, accurate pricing and availability, rich attributes, is what gets your products surfaced and recommended by AI systems.
Third, demand attribution clarity from every partner in your AI commerce stack. When ChatGPT, Perplexity or any AI assistant drives a shopper to your app, you need to know it, and you need the data to optimize against it.
OpenAI’s decision to deeplink rather than own checkout is a gift to the retail industry, even if it doesn’t look like one at first glance. The AI discovery layer is mature and scaling fast. The conversion layer is where billions in latent revenue are waiting to be unlocked. The only question is which retailers, and which technology partners, will be ready to capture it.
Forbes Business Development Council is an invitation-only community for sales and biz dev executives. Do I qualify?
