In September 2025, OpenAI made a bet on agentic commerce. People were already coming to ChatGPT to figure out what to buy. They asked for product ideas, researched features, and compared their options. 

The behavior was there, so OpenAI added a checkout button. The promise was simple: AI assistants that don't just help people find products but buy them too. No browser tabs or checkout forms. Just a conversation ending with a purchase. 

OpenAI launched Instant Checkout with Etsy, Shopify, and Walmart among the brands announced as partners. But just six months later, OpenAI quietly stepped back entirely.  

And here's the thing: they're not the first platform to promise native checkout and retreat. This is becoming a pattern, and brands that understand it will be in the best position to win what comes next. 

So, what happened? And what does it tell us about where agentic commerce is headed? 

Where OpenAI Went Wrong with Agentic Commerce

The failure wasn't a lack of shopper intent. People genuinely want help buying things and are using AI to do it. The problem was a gap between user behavior and platform design. 

When you give users access to a powerful AI tool that can answer any question, research any product, and compare any option, they use it for exactly that. They research, explore, and compare. Very few users were actually ready to complete transactions inside ChatGPT. 

On the merchant side, the complexity was even more daunting. Only about a dozen of Shopify's million-plus sellers built the technical infrastructure needed to support in-chat transactions. From real-time inventory sync and fraud detection to tax compliance across jurisdictions and privacy regulations, it turns out OpenAI was not ready for the unglamorous operational infrastructure needed for running a checkout experience at scale. 

So, they made a pragmatic call. Now, ChatGPT hands off customers to the brand, either their app or website, with cart items already loaded. 

History Is Repeating: Meta and TikTok's Native Checkout Attempts

This isn't the first time a major platform promised seamless in-platform checkout and then retreated. Meta spent years building infrastructure for native checkout in Instagram and Facebook. The idea was similar to OpenAI's promise: keep the customer in platform, reduce friction, and own the full experience. 

Then in 2025, Meta walked back native checkout too. 

The new model: platforms own discovery while brands own transactions. Brands are setting up full shop experiences inside social, but when the customer is ready to buy, they land on the brand's website. Cart intact, transaction on brand-owned infrastructure.  This benefits brands since they can collect visit and checkout data in order to own and nurture their customers. 

TikTok has been pushing their native checkout and fulfillment for years. But as the feed becomes more crowded with sellers, TikTok mandates a pay-to-play model through paid ads, and inconsistent TikTok-owned logistics leading to post-purchase friction, brands are starting to question this model. 

The same model is playing out in AI commerce right now, just on a faster timeline. 

Google's Agentic Commerce Advantage Over OpenAI

While OpenAI was stepping back from agentic commerce, Google launched agentic checkout through its Universal Commerce Protocol. Google had two things OpenAI didn't:

1. Merchant Center to standardize product data at scale.
2. Google Pay to handle transactions. 

The commerce system was already there. The agentic layer just needed to connect to it. 

That advantage matters. It's proof that agentic commerce isn't a feature any platform can add overnight. It requires data infrastructure, payment guardrails, and consumer trust. Google had years of investment in all three before anyone was even discussing agentic commerce. 

How to Prepare for Agentic Commerce

Agentic commerce is not a trend to watch from a distance. Platforms are building the infrastructure right now, and the window for brands to get ahead is open. Prepare for agentic commerce by focusing on these areas: 

1. Make Your Products Discoverable by AI

AI tools are already surfacing product recommendations. If your product data isn't structured, accurate, and accessible to these platforms, you're invisible. To make your brand discoverable by AI, we recommend setting up clean feeds, strong product descriptions, and proper schema markup. 

2. Set Up a Seamless Checkout Handoff

The Meta Shops model tells us where this is headed in the short- to medium-term: AI platforms will own discovery, but brands will own transactions. Make sure your commerce platform is ready to receive customers from external sources with cart data intact. Brands on Shopify are largely in good shape, but brands on Salesforce Commerce Cloud (SFCC) or another enterprise platform should scope the technical lift now, not later. 

3. Start Testing Agentic Commerce Tools Now 

OpenAI is rebuilding. Google is expanding its pilot. Microsoft is integrating AI shopping into Bing and Copilot. Each platform is moving at a different pace with a different approach. But this isn't a "wait and see" situation. The tools available to brands will change fast, and the brands that test early will have a real edge. 

4. Build a Measurement Framework That Matches the Channel

This is where most brands will fall short. Agentic commerce is a new acquisition channel and won't show up cleanly in your existing ecommerce analytics. Last-click attribution and site traffic metrics won't cut it, but enterprise-level measurement will. 

You need a measurement framework built to isolate the revenue, customer acquisition, and brand impact of these new sales channels. Most importantly, your measurement should connect emerging channel activity to enterprise-level business outcomes, not just site traffic. 

Is Your Brand Agentic Commerce Ready?

OpenAI's retreat from agentic commerce signals that it's more complex, infrastructure-dependent, and brand-controlled than anyone initially pitched. Google is proving that the concept works when the infrastructure is right. Meta already showed us the checkout handoff model can scale. The platforms are converging on a model where they own the top of the funnel and brands own the close. 

The brands that will win aren't the ones waiting for a turnkey solution. They're the ones building the technical connections and measurement frameworks now before agentic commerce becomes table stakes. 

At Ovative, we're actively working with our clients to navigate this emerging channel in real time, helping brands build for what's next. If you're ready to build your agentic commerce strategy, our media and measurement experts are ready to help. Contact us to get started.