ChatGPT Shopping Features: How 50 Million Daily Queries Are Reshaping Product Discovery

On November 24, 2025, OpenAI launched ChatGPT Shopping Research -- a dedicated shopping experience that transformed how hundreds of millions of users discover and evaluate products. Within weeks of launch, ChatGPT was processing 50 million shopping-specific queries daily, roughly 2% of its 2.5 billion total daily prompts. That single feature turned a chatbot into the fastest-growing product discovery channel in ecommerce history.

For ecommerce merchants, the implications are enormous. ChatGPT shopping recommendations are organic and unsponsored -- relevance, not ad spend, determines which products appear. This means every merchant, regardless of budget, competes on the same terms. But understanding how the system works, what data it pulls, and how its features function is essential to capturing this traffic.

This guide breaks down every major ChatGPT shopping feature, the infrastructure behind it, and the specific mechanics that determine which products get recommended.

The Scale of ChatGPT Shopping

Before examining individual features, the numbers establish why this matters:

  • 50 million shopping queries daily -- confirmed by research from OpenAI's Economic Research team and Harvard economist David Deming
  • 900 million weekly active users as of February 2026, more than doubling from 400 million in February 2025
  • 84 million shopping queries per week in the US alone, representing roughly 8% of Amazon's weekly search volume (Stackline)
  • 1,079% session growth in ChatGPT ecommerce traffic during 2025 (Adobe Analytics)
  • 752% year-over-year spike in AI referrals to ecommerce during the 2025 holiday season (Adobe Analytics)
  • 1 in 5 of Walmart's referral clicks now originates from ChatGPT (Similarweb)

These are not projections. This is documented traffic flowing through a system that did not exist 18 months ago. And because Amazon blocks all OpenAI crawlers -- making 600 million Amazon product listings invisible to ChatGPT Shopping -- there is an enormous vacuum for independent retailers and DTC brands to fill.

Product Cards: The Visual Shopping Interface

When ChatGPT detects a shopping-related prompt, it can display visual product cards directly within the conversation. These cards are the primary shopping interface and include:

  • Product images pulled from merchant feeds and structured data
  • Current pricing with currency and availability status
  • Star ratings aggregated from review data
  • Key specifications relevant to the query context
  • Direct purchase links to merchant product pages

Product cards are not static search results. They are dynamically generated based on the specific query context, user preferences expressed during the conversation, and real-time product data. When a user asks "What's the best espresso machine for a small kitchen under $400?", the product cards that appear are filtered by price, sized for kitchen counters, and ranked by review quality -- all within the conversational flow.

How Product Card Data Is Sourced

ChatGPT pulls product card data from multiple sources in a specific priority hierarchy:

  1. Microsoft Merchant Center feeds -- This is the primary product data source. A feed maintained in Microsoft Merchant Center is the main source of product data powering ChatGPT's shopping features. Feeds must be updated regularly and maintained in full.

  2. Shopify and Etsy catalog APIs -- Products from these platforms are automatically discoverable in ChatGPT via Agentic Storefronts, requiring no separate integrations, no apps, and no additional transaction fees beyond standard processing rates.

  3. Structured data from product pages -- Product schema markup (JSON-LD) provides fallback data when direct feed integrations are not available. Merchants with comprehensive Product schema see a 34% higher rate of inclusion in AI shopping features.

  4. Google Shopping data -- Semrush research confirmed that ChatGPT searches Google Shopping to create its recommendations, adding another data layer beyond Bing.

The practical implication: if your product data exists only on your website without structured markup, without a Merchant Center feed, and without a Shopify or Etsy integration, your products are significantly less likely to appear in product cards.

Product Card Schema Requirements

The minimum structured data that feeds product card generation includes:

{
  "@type": "Product",
  "name": "Product Name",
  "image": "https://yourstore.com/images/product.jpg",
  "description": "Detailed product description",
  "brand": { "@type": "Brand", "name": "YourBrand" },
  "offers": {
    "@type": "Offer",
    "price": "199.99",
    "priceCurrency": "USD",
    "availability": "https://schema.org/InStock",
    "url": "https://yourstore.com/product"
  },
  "aggregateRating": {
    "@type": "AggregateRating",
    "ratingValue": "4.5",
    "reviewCount": "234"
  }
}

Missing fields directly reduce the richness of your product cards. A product without an aggregateRating will display without stars, making it visually less compelling next to competitors who include review data.

Price Comparison: Cross-Merchant Analysis

One of the most commercially significant shopping features is ChatGPT's ability to compare prices across merchants for the same or similar products. When a user asks "Where can I get the Sony WH-1000XM5 for the best price?", ChatGPT:

  1. Searches across multiple merchant sources including direct retailer sites, marketplace listings, and feed data
  2. Presents comparison tables showing price, availability, shipping options, and seller ratings side by side
  3. Identifies price ranges so users can see the spread between the cheapest and most expensive options
  4. Flags deals and discounts when pricing data indicates a sale or promotional pricing

This has profound implications for pricing strategy. Your product page is no longer competing only against the ten blue links on a Google results page. It is being directly compared, in a structured table format, against every merchant selling the same product that ChatGPT can find through its data sources.

The Pricing Data Pipeline

ChatGPT assembles pricing data from:

  • Merchant Center feeds where price fields are explicitly structured
  • Structured data on product pages where Offer schema includes current pricing
  • Real-time page fetching where ChatGPT-User visits product pages and reads prices from HTML

The critical requirement is that prices must be present in server-rendered HTML. Prices loaded via client-side JavaScript API calls after page load are largely invisible to OAI-SearchBot. If your product page renders pricing through a React or Vue component that fetches prices from an API endpoint, ChatGPT may see a product page with no price -- and a product without a price does not appear in comparison tables.

Visual Search: From Photo to Purchase

ChatGPT's visual search capability allows users to upload a photo of a product and find similar or identical items available for purchase. A user can photograph a pair of shoes they see on the street, upload the image, and ChatGPT will return specific products with images, prices, and a tabular comparison of fabric, fit, and price range.

This feature leverages OpenAI's multimodal capabilities -- the same vision system that powers image understanding across ChatGPT. For shopping, it means:

  • Reverse product identification -- Upload a photo, get the exact product name and available retailers
  • Style matching -- Upload a photo of a couch and find visually similar options within a specified budget
  • Material and color matching -- The system can identify fabric types, color shades, and design patterns to find close matches

Visual search relies heavily on product image quality and metadata. To maximize visibility:

  • Use high-resolution images with clean backgrounds on product pages
  • Include multiple angles -- front, side, detail shots, and lifestyle context
  • Add descriptive alt text that includes material, color, style, and use case
  • Ensure image URLs are stable and accessible to crawlers (not behind authentication or lazy-loading that prevents server-side rendering)
  • Implement ImageObject schema with descriptive name and description properties

Products with rich visual assets are more likely to be matched when users perform visual searches, because the system has more visual data points to compare against.

Interactive Shopping Research Mode

The November 2025 launch introduced a full interactive shopping research experience that goes beyond simple product cards. When ChatGPT identifies a complex shopping query -- one that requires understanding preferences, budget constraints, use cases, and trade-offs -- it activates shopping research mode.

The Research Flow

The shopping research process follows a specific pattern:

  1. Clarifying questions -- ChatGPT asks about budget, intended use, feature priorities, brand preferences, and deal-breakers. For a laptop query, it might ask about screen size needs, whether gaming performance matters, portability requirements, and operating system preference.

  2. Multi-source research -- The system then researches across the web, reviewing multiple quality sources. OpenAI's testing showed this process takes 3-5 minutes for complex queries as the system evaluates dozens of sources.

  3. Personalized buyer's guide -- ChatGPT delivers results as a structured guide with product images, current pricing, availability, specifications, and aggregated review sentiment. This is not a list of links -- it is a synthesized evaluation.

  4. Interactive refinement -- Users can mark products as "Not interested" or "More like this" in a Tinder-style interface, and the model updates suggestions in real-time based on the feedback. This creates a personalized discovery loop that traditional search cannot replicate.

  5. Ongoing conversation -- Users can ask follow-up questions like "What about battery life specifically?" or "Is there a version of this under $300?" and the system refines its recommendations without starting over.

Accuracy and Performance

OpenAI's internal testing measured Shopping Research at a 64% accuracy rating -- defined as how well product recommendations matched user expectations on details like price, color, material, and feature requirements. While that figure leaves room for improvement, it represents performance from initial launch. The system uses a specialized model -- a version of GPT-5 mini post-trained specifically for shopping use cases -- to deliver more accurate product matching.

Instant Checkout: Purchasing Without Leaving Chat

OpenAI introduced Instant Checkout, allowing users to complete purchases directly within the ChatGPT interface for participating merchants. When Instant Checkout is available for a recommended product, a "Buy" button appears directly in the chat.

How Instant Checkout Works

  1. User receives a product recommendation with Instant Checkout availability
  2. User clicks the "Buy" button within the chat interface
  3. Payment is processed through the merchant's existing payment infrastructure
  4. Order confirmation appears within the chat

Merchant Requirements and Fees

  • Shopify merchants are automatically connected via the Agentic Commerce Protocol (ACP). Products become discoverable in ChatGPT by default through Agentic Storefronts, requiring no separate app installation.
  • Etsy sellers were early partners for Instant Checkout integration
  • Transaction fee: OpenAI charges merchants a 4% transaction fee on every completed Instant Checkout purchase
  • Product Feed Spec: The ACP requires merchants to supply complete, accurate, and fresh product data -- updated as often as every 15 minutes -- to power ChatGPT's product search and Instant Checkout

The 4% fee is a meaningful cost, but the context matters. There is no advertising spend required to appear in ChatGPT shopping results. The recommendations are organic. For merchants paying 15-30% of revenue on Google Ads or Amazon advertising fees, a 4% transaction fee on incremental sales from an organic channel is a favorable trade.

The Agentic Commerce Protocol

The Agentic Commerce Protocol (ACP) is the underlying infrastructure that makes Instant Checkout possible. It standardizes how merchants feed product catalogs into ChatGPT and how transactions are processed. Key elements:

  • Real-time product data sync with updates as frequent as every 15 minutes
  • Inventory and availability tracking so ChatGPT does not recommend out-of-stock products
  • Price synchronization ensuring the checkout price matches the displayed price
  • Order routing back to the merchant's existing fulfillment system

For Shopify merchants, ACP integration is automatic. For other platforms, integration requires submitting product feeds through Microsoft Merchant Center or direct API connections.

The "No Ads" Advantage

Perhaps the most disruptive aspect of ChatGPT Shopping is what it does not include: advertising. Every product recommendation is organic, ranked purely on relevance to the user's query and preferences. OpenAI has stated this explicitly -- product results are unsponsored.

This creates a fundamentally different competitive dynamic than Google Shopping or Amazon, where paid placement heavily influences which products users see first. On ChatGPT:

  • A small DTC brand with excellent reviews and well-structured product data competes equally against Fortune 500 retailers
  • Products are recommended based on how well they match the user's stated needs, not how much the merchant bid on keywords
  • Review quality and authentic third-party validation become the primary competitive moat

For merchants who have been priced out of Google Ads or cannot compete with Amazon's advertising flywheel, ChatGPT Shopping represents a level playing field that has not existed in ecommerce for over a decade.

Will Ads Come Eventually?

OpenAI has not announced advertising plans for Shopping Research, but the revenue potential is obvious. The Agentic Commerce Protocol's 4% transaction fee is one monetization path. If advertising does arrive, early merchants who have already established organic visibility and built trust signals will have a significant first-mover advantage -- the same way early Google SEO practitioners benefited before paid search dominated the landscape.

ChatGPT Shopping Versus Other AI Shopping Platforms

How does ChatGPT's shopping implementation compare to competitors?

| Feature | ChatGPT Shopping | Perplexity Buy with Pro | Google AI Mode | Microsoft Copilot Checkout | |---|---|---|---|---| | Daily shopping queries | 50 million | Not disclosed | Not disclosed | Not disclosed | | In-chat checkout | Yes (Instant Checkout) | Yes (Buy with Pro) | Coming soon | Yes (Copilot Checkout) | | Product cards | Yes, with images and pricing | Yes, with pricing | Yes, via Shopping Graph | Yes, via Bing Shopping | | Ad-free results | Yes (currently) | Yes | No (ads appearing) | No (ads integrated) | | Visual search | Yes (photo upload) | Yes (Snap to Shop) | Yes (Circle to Search) | Limited | | Price comparison | Yes | Limited | Yes (Shopping Graph) | Yes (Edge sidebar) | | Merchant fee | 4% on Instant Checkout | Zero fees | Standard Google fees | Standard Microsoft fees | | Primary data source | Microsoft Merchant Center + Shopify | Own merchant program | Google Merchant Center | Bing Merchant Center |

What Merchants Should Do Now

The window for establishing ChatGPT shopping visibility while the platform is still in its organic, ad-free phase is closing. Merchants who act now will build the trust signals and data infrastructure that become increasingly difficult to replicate once the market matures.

Immediate Actions

  1. Submit product feeds to Microsoft Merchant Center -- This is the single most impactful step. ChatGPT's primary product data source is Microsoft's feed infrastructure. If your products are not in Merchant Center, they are significantly less likely to appear in shopping results.

  2. Implement complete Product schema on every product page with price, availability, reviews, images, brand, and SKU. Merchants with comprehensive schema see 34% higher inclusion rates in AI shopping features.

  3. Ensure server-side rendering of all product data. Prices, reviews, and specifications must be present in the initial HTML response, not loaded via client-side JavaScript.

  4. Register in Bing Webmaster Tools and verify your site is being crawled. Bing indexing is the gateway to ChatGPT visibility.

  5. Allow OAI-SearchBot and ChatGPT-User in your robots.txt while blocking GPTBot if you want to prevent training data usage.

Strategic Actions

  1. Build product data freshness by updating product pages regularly. Content updated within 30 days receives significantly more citations. Add visible "last updated" dates to product and category pages.

  2. Create use-case-specific content -- not generic product descriptions, but detailed buyer's guides that match how people ask ChatGPT for recommendations. "Best espresso machine for beginners under $300" as an H2 heading directly matches natural language queries.

  3. Develop third-party validation -- Get products reviewed on Wirecutter, niche review sites, Reddit communities, and YouTube channels. ChatGPT cross-references your claims against external sources before recommending products.

  4. Maintain price competitiveness -- With cross-merchant price comparison built into the shopping interface, pricing outliers are immediately visible. Ensure your pricing is competitive or that your value proposition (faster shipping, better warranty, exclusive bundles) is clearly articulated in your product data.

  5. Monitor AI referral traffic in your analytics. Set up tracking for ChatGPT referral traffic (referrers include chatgpt.com and chat.openai.com) to measure the channel's growth and conversion performance for your store.

The Trajectory

ChatGPT Shopping launched in November 2025. Within months, it reached 50 million daily shopping queries. AI-referred traffic to ecommerce grew 1,079% during 2025. Shopify stores saw AI-attributed orders increase 11x between January 2025 and early 2026.

The trajectory is not linear -- it is exponential. And because ChatGPT shopping results are organic, the merchants who build visibility now are building a competitive moat that will compound over time. Every product recommendation ChatGPT makes that leads to a positive user experience reinforces that product's position in future recommendations.

The question is not whether ChatGPT Shopping will matter for your business. The question is whether you will be visible when your customers start asking.

Up next

Optimizing for ChatGPT

Specific tactics to get your products recommended by ChatGPT.