How It WorksFeaturesPricingBlog
Get Started
Back to all posts
chatgpt-shoppingai-shoppingguide

ChatGPT Shopping in March 2026: What Actually Works After OpenAI Killed Native Checkout

5 min read
Share:
ChatGPT Shopping in March 2026: What Actually Works After OpenAI Killed Native Checkout

OpenAI pulled native checkout from ChatGPT in early March 2026. Forrester confirmed only about 30 Shopify merchants ever had the feature live. And yet 900 million people use ChatGPT every week, many of them actively asking for product recommendations.

The checkout model changed. The discovery opportunity got bigger. Here's what ChatGPT shopping actually looks like today and how to make sure your products show up.

How Does ChatGPT Shopping Work in 2026?

ChatGPT shopping now operates as a referral channel. When someone asks for product recommendations, ChatGPT shows product cards inside the conversation - images, prices, ratings, short descriptions. Clicking a card takes the user to the retailer's website to buy.

This replaced the native checkout model OpenAI tested with Shopify, where transactions happened entirely inside ChatGPT. The referral model is actually better for most retailers: you keep the checkout experience, the customer data, the attribution, and the upsell opportunity. Under native checkout, all of that lived inside OpenAI's interface.

The question for retailers isn't about the checkout flow anymore. It's whether your products appear in the recommendation at all. That moment - when ChatGPT decides to show your product instead of a competitor's - is where the value lives.

What Determines Which Products ChatGPT Recommends?

ChatGPT pulls from multiple data sources to build product recommendations, and understanding these sources is the key to showing up.

Product feed quality. ChatGPT's product index draws from feeds and web-crawled data. Products with rich, structured attributes - 30 or more per product - consistently appear over products with only basic title, price, and description. Think material composition, exact dimensions, compatibility information, care instructions, and certifications. The more attributes, the more queries your product can match.

Bing indexing. ChatGPT uses Bing for its web search results. This is the most commonly overlooked gap in ChatGPT visibility. Most ecommerce teams optimize exclusively for Google and have never checked their Bing Webmaster Tools. If Bing hasn't indexed your product pages, ChatGPT can't surface them through search.

JSON-LD structured data. Rich Product schema on your pages gives ChatGPT's crawlers structured information to parse directly. additionalProperty fields for specifications, aggregateRating for social proof, offers with real-time pricing and shipping details, and MerchantReturnPolicy for purchase confidence. Basic schema isn't enough - the expanded attributes are what separate surfaced products from invisible ones.

Third-party trust signals. Review volume, editorial mentions, and community discussions all feed ChatGPT's confidence in recommending a product. A product with 500 verified reviews across platforms will consistently outrank one with 10 reviews, all else equal.

Data freshness. Stale pricing, out-of-stock products showing as available, outdated descriptions - these erode trust signals quickly. ChatGPT deprioritizes products where the data looks unreliable.

Why Is Bing Indexing Critical for ChatGPT Visibility?

This deserves its own section because it's the single most impactful and least understood lever for ChatGPT shopping visibility.

ChatGPT's web search is powered by Bing. When a user asks "best waterproof hiking boots under $200," ChatGPT queries Bing, processes the results, and synthesizes a recommendation. If your product pages aren't in Bing's index, they don't exist in that pipeline.

Here's a concrete example: we recently audited a site with 27 published content pages and found Bing had only indexed 6 of them. That's 78% of the site's content completely invisible to ChatGPT's search. After submitting all URLs through Bing's Webmaster API, new pages started appearing in Bing within 24 hours.

Go to Bing Webmaster Tools today. Check how many of your pages are actually indexed. Submit your sitemap. Use the URL Submission API for your highest-value product pages. This single action can materially improve your ChatGPT visibility faster than any other optimization.

What Changed for Retailers When OpenAI Killed Native Checkout?

Three shifts happened, and two of them are actually good for retailers.

Your website is back in the purchase flow. Under native checkout, users completed transactions without ever seeing your site. Now they click through to your storefront. That means your conversion rate, page speed, and checkout experience directly impact whether a ChatGPT referral turns into revenue. It also means you can retarget visitors, build email lists, and cross-sell - none of which was possible under native checkout.

Attribution works. Look for chatgpt.com in your referral traffic. You can now track which products ChatGPT recommends, which get clicked, and which convert. Under native checkout, you had almost zero visibility into the buyer journey.

Platform exclusivity ended. Native checkout required Shopify Payments. The referral model works with every ecommerce platform. Adobe Commerce, SFCC, BigCommerce, custom builds - any retailer can benefit from ChatGPT shopping recommendations now, not just Shopify merchants.

How to Get Your Products Into ChatGPT Shopping Results

Seven concrete steps, ordered by impact:

1. Get indexed in Bing. Register with Bing Webmaster Tools. Submit your sitemap. Push your top product pages through the URL Submission API. This is the highest-impact action because it addresses the most common blind spot.

2. Enrich product data to 30+ attributes. Most retailers have 5-8 structured attributes per product. Target at least 30: material, dimensions, weight, care instructions, compatibility, certifications, use-case tags, color variants, and sizing details. Every additional attribute is another potential query match.

3. Implement expanded JSON-LD. Go beyond basic Product schema. Use additionalProperty for every specification. Include aggregateRating, detailed offers with shipping, and MerchantReturnPolicy. This is the structured data AI crawlers parse most effectively.

4. Optimize your Google Merchant Center feed. Use product_detail attributes for everything beyond standard fields. This feeds Google AI Mode directly and indirectly influences ChatGPT's product index through cross-referenced data.

5. Add llms.txt to your domain. A single text file at your domain root that tells AI crawlers what your site offers and where to find key information. Takes an hour to implement, costs nothing, and improves discoverability across every AI engine - not just ChatGPT.

6. Build review volume systematically. AI agents treat reviews as a confidence signal. Manage your review profile across Google, your own site, and relevant vertical platforms. A product with thin reviews will lose to a well-reviewed competitor even if your product data is richer.

7. Monitor your AI visibility weekly. Search for your products using natural customer language. "Best wireless earbuds for running." "Organic skincare for sensitive skin." Track which products surface and which competitors appear instead. Do this across ChatGPT, Perplexity, and Google AI Mode. Build a simple spreadsheet and update it weekly.

How Big Is the ChatGPT Shopping Opportunity?

The numbers behind the opportunity are substantial and growing:

The shift from native checkout to referral didn't reduce the opportunity. It actually made it more accessible. Any retailer on any platform can now show up in ChatGPT's product recommendations. The differentiator is whether your product data, your Bing presence, and your structured markup give ChatGPT enough information to confidently recommend you over a competitor.

Retailers who treat AI visibility as a channel - with dedicated monitoring, data enrichment, and optimization cycles - will capture a growing share of those 900 million weekly users. The rest will keep wondering why their products don't show up.