OpenAI has officially killed native checkout inside ChatGPT and pivoted entirely to product discovery. The instant checkout feature, launched with considerable fanfare in September 2025, lasted just six months before OpenAI pulled the plug. This is not a minor UX tweak. It is a fundamental strategic shift that redefines how 900 million weekly ChatGPT users interact with products and brands inside the platform.
For brands that built their AI commerce strategy around ChatGPT as a transaction platform, this pivot demands an immediate rethink. For brands that have not yet started, this is actually the better entry point.
Here is everything that changed, why it matters, and what to do about it.
What Happened to ChatGPT Checkout?
OpenAI launched native checkout inside ChatGPT in September 2025 with the promise of frictionless AI-powered commerce. Users could discover a product, add it to a cart, and complete a purchase without ever leaving the chat interface. On paper, it was the future of shopping.
In practice, it did not work.
The core problems were trust and conversion. Users were willing to ask ChatGPT for product recommendations, but the moment they were asked to enter payment information inside a chat window, conversion rates cratered. Internal data reportedly showed that ChatGPT checkout converted at less than 1.2% of sessions where a purchase intent was detected, compared to 3.5-4.8% when users were redirected to familiar retailer websites.
The math was simple: users trust Amazon's checkout page. They trust Nordstrom's checkout page. They do not yet trust a chatbot's checkout page, regardless of how sophisticated the underlying AI is.
OpenAI made the right call. Rather than fighting consumer behavior, they leaned into what was already working: product discovery.
What Are the New Product Discovery Features?
ChatGPT's shopping experience has been rebuilt around three core discovery capabilities that launched in March 2026.
Side-by-Side Product Comparisons
Users can now ask ChatGPT to compare products across any attributes, and the response renders as a structured comparison table directly in the conversation. This is not the old text-based list of pros and cons. It is a dynamic, interactive comparison that pulls real-time data from product feeds.
| Feature | Old ChatGPT Shopping | New Discovery Experience |
|---|---|---|
| Product display | Text descriptions only | Visual cards with images, pricing, ratings |
| Comparison capability | Unstructured paragraphs | Structured side-by-side tables |
| Price data | Static, often outdated | Real-time from retailer feeds |
| Purchase path | Native checkout (removed) | Direct retailer handoff |
| Filtering | Basic text prompts | Conversational multi-attribute filtering |
| Image search | Not available | Upload photo, find similar products |
Image-Based Product Search
Users can now upload a photo and ask ChatGPT to find similar products, identify the exact product, or suggest alternatives at different price points. This feature leverages GPT-5's multimodal capabilities and is already being used in roughly 18% of product-related conversations.
The implications for brands are significant. Product images are now a discovery vector, not just a conversion asset. Brands with high-quality, distinctive product photography have a material advantage in image-based search.
Conversational Filtering
Instead of rigid category filters, users can now refine product recommendations through natural conversation. A query like "Show me running shoes under $150 that work for flat feet and come in wide sizes" returns filtered results that would have required navigating six different dropdown menus on a traditional e-commerce site.
Early data shows that conversational filtering sessions result in 34% higher click-through rates to retailer sites compared to standard product recommendation queries.
How Has the Agentic Commerce Protocol Expanded?
The Agentic Commerce Protocol (ACP) is OpenAI's framework for connecting retailer ecosystems directly into ChatGPT's discovery layer. It launched in late 2025 with a handful of retail partners and has expanded significantly in Q1 2026.
Confirmed ACP Retailers as of March 2026
| Retailer | ACP Status | Product Categories | Integration Depth |
|---|---|---|---|
| Amazon | Full integration | All categories | Product data, pricing, reviews, availability |
| Walmart | Custom integration (Sparky) | All categories | Product data, pricing, store availability |
| Target | Full ACP integration | All categories | Product data, pricing, same-day delivery status |
| Sephora | Full ACP integration | Beauty, skincare | Product data, reviews, shade matching |
| Nordstrom | Full ACP integration | Fashion, beauty | Product data, styling recommendations |
| Best Buy | Full ACP integration | Electronics, appliances | Product data, pricing, Geek Squad availability |
| Home Depot | Full ACP integration | Home improvement | Product data, project-based recommendations |
This expansion means that ChatGPT's product discovery now covers the vast majority of retail categories. When a user asks for a product recommendation, ChatGPT can pull real-time data from these retailer feeds, display accurate pricing and availability, and link directly to the retailer's product page.
For brands that sell through these retailers, ACP participation is automatic but optimization is not. Your product data in these retailer feeds is now a direct input to ChatGPT's recommendation engine.
Why Does This Pivot Matter for 900M Weekly Users?
ChatGPT's 900 million weekly active users make it the largest AI platform in the world by a wide margin. When OpenAI makes a strategic shift in how those users interact with products, the entire AI commerce landscape moves with it.
Here are the numbers that matter:
- 37% of consumers now start at least some product searches with AI tools rather than Google or Amazon
- AI-influenced shopping is projected to reach $20.9 billion in 2026, a 4x increase year over year
- 60% better conversion rates are observed when shoppers engage with AI-powered product assistants before purchasing
- Product discovery queries on ChatGPT increased 215% in Q1 2026 compared to Q4 2025
The pivot from checkout to discovery means ChatGPT is positioning itself as the top of the shopping funnel, not the bottom. It wants to be the place where consumers decide what to buy, then hands them off to retailers to complete the transaction.
This is a much larger addressable market. Checkout is one moment. Discovery is an entire journey.
What Should Brands Do Right Now?
The pivot creates specific, actionable opportunities. Here is a prioritized list of what to do in the next 30 days.
1. Audit Your Product Data in Retailer Feeds
If you sell through any ACP-connected retailer, your product data in their feeds is now a direct input to ChatGPT's discovery engine. Audit every field:
- Titles: Include the attributes users compare on (size, material, key specs)
- Descriptions: Write for AI parsing, not just human scanning. Use clear, factual language with specific measurements and specifications
- Images: Ensure you have at least 4-6 high-quality images per product, including lifestyle shots that work for image-based search
- Specifications: Fill in every available attribute field. Incomplete spec sheets mean your product gets excluded from comparison tables
2. Optimize for Comparison Queries
ChatGPT's new side-by-side comparison feature means your product will frequently appear next to competitors. Prepare for this by:
- Identifying the top 3-5 attributes users compare in your category
- Ensuring your product data clearly communicates your advantages on those attributes
- Creating content on your own site that provides honest, detailed comparisons (this feeds ChatGPT's training data and citation sources)
3. Invest in Product Photography for Visual Search
Image-based search is now a real discovery channel. Every product needs:
- Clean, white-background product shots from multiple angles
- Lifestyle/context images showing the product in use
- Detail shots of key differentiating features
- Consistent visual branding that makes your products recognizable across search results
4. Build Your Review Profile Across Platforms
ChatGPT synthesizes reviews from multiple sources. A brand with 500 reviews on Amazon but zero reviews on Google, Best Buy, or independent review sites has a weaker signal than a brand with 200 reviews distributed across five platforms.
Prioritize review generation on the specific retailers where you sell through ACP-connected channels.
5. Monitor Your ChatGPT Discovery Presence
Run weekly queries for your product category and track:
- Whether your products appear in comparison tables
- How your products are described versus competitors
- Whether your product images are being surfaced in visual search results
- Which retailer links ChatGPT directs users to for your products
How Does This Affect the Broader AI Commerce Landscape?
OpenAI's pivot validates a pattern we have been tracking across every major AI platform: AI is winning the discovery layer, not the transaction layer.
Amazon Rufus, with 250 million users and an estimated $10 billion in incremental sales, succeeds because it is embedded inside Amazon's existing checkout infrastructure. Macy's AI chatbot, which drives 4.75x more spending per user, works because it enhances an existing shopping experience rather than replacing it.
The AI platforms that try to own the entire transaction lose. The AI platforms that own discovery and hand off to trusted purchase environments win.
For brands, this means the critical optimization target is the discovery moment: the conversation where a consumer asks "What should I buy?" Your product needs to appear in that conversation with accurate data, compelling attributes, and strong visual assets.
The checkout will take care of itself. The discovery is where the battle is won or lost.
What Comes Next
OpenAI has signaled that further shopping features are coming in Q2 2026, including personalized product alerts, AI-curated shopping lists, and deeper integration with loyalty programs through ACP partners.
The brands that invest now in product data quality, visual assets, and multi-platform review profiles will be positioned to capture disproportionate value as these features roll out.
Need help optimizing your product discovery presence across ChatGPT and other AI platforms? Schedule a free consultation with our AI search team. We will audit your current visibility across every major AI shopping surface and identify the highest-impact improvements.