Key takeaways
- ChatGPT Shopping carousels pull 83% of their product listings directly from Google Shopping -- so your Google Merchant Center feed is the single most important lever to pull
- ChatGPT does NOT rank products by price or shipping cost, but those factors influence whether your product gets recommended at all
- You need to explicitly allow OAI-SearchBot to crawl your site -- blocking it means ChatGPT can't discover your pages
- Product schema markup, high-quality images, and review data are the three on-page factors that most directly affect carousel inclusion
- Tracking your ChatGPT Shopping appearances requires dedicated tooling -- standard Google Analytics won't show you this
Why ChatGPT Shopping matters right now
Google has dominated product discovery for two decades. Then Bing added shopping carousels. Now ChatGPT does it too -- and unlike Google Shopping ads, the carousels are currently free to appear in.
That last point deserves a moment. You're not bidding for placement. There's no cost-per-click. ChatGPT surfaces products it thinks are the best match for a user's query, and if your product data is in order, you can show up next to brands spending millions on paid search.
The scale is real. ChatGPT crossed 400 million weekly active users in early 2026. A meaningful chunk of those users are asking shopping questions: "best running shoes under $150," "which espresso machine is worth buying," "compare noise-cancelling headphones." When they do, they see a visual carousel with product images, prices, ratings, and a direct link to buy. Walmart has already partnered with OpenAI to build an AI-first shopping assistant. The channel is maturing fast.
The window for early-mover advantage is still open, but it's closing.
How ChatGPT Shopping actually works
Before optimizing for something, it helps to understand the mechanics.
Where the product data comes from
According to a study by ALM Corp, ChatGPT sources 83% of its product carousel items directly from Google Shopping. This is the most important data point in this entire guide. It means the infrastructure you've already built for Google -- your Merchant Center feed, your product listings, your review signals -- is the foundation for ChatGPT Shopping too.
The remaining ~17% comes from direct web crawling. ChatGPT's crawler (called OAI-SearchBot) visits product pages, reads structured data, and indexes what it finds. So there are two distinct paths to appearing in carousels: through Google Shopping, and through direct crawl.
What ChatGPT evaluates
ChatGPT's shopping results are shaped by several factors:
- Relevance to the user's query (the most important factor)
- Product title and description quality
- Image quality and completeness
- Review data and ratings
- Availability and in-stock status
- Structured data markup on your product pages
What it does NOT rank on: price and shipping costs. Those factors don't determine your position in the carousel -- but they do determine whether your product gets recommended. If a user asks for "espresso machines under $200" and your product is $350, ChatGPT won't show it regardless of how well-optimized your listing is. That's not a flaw; it's the system working correctly.
The user experience
When a user asks a shopping question, ChatGPT displays a horizontal carousel with product cards. Each card shows a product image, name, price, and star rating. Users can click to expand a card and see more details: additional images, where to buy, why ChatGPT thinks they'd like it, and a summary of what other people are saying. They can also ask follow-up questions about a specific product directly in the chat.
OpenAI is also rolling out ChatGPT Instant Checkout via the Agentic Commerce Protocol, which will let users buy without leaving the chat. That's still in early stages, but it signals where this is heading.
The Google Shopping foundation
Since 83% of ChatGPT carousel items come from Google Shopping, your first priority is getting your Google Merchant Center feed in excellent shape.
Audit your product feed
A sloppy feed hurts you on Google and now on ChatGPT. Run through these basics:
- Product titles should be descriptive and include the key attribute buyers search for (brand, material, size, color where relevant). "Blue Running Shoes" is weak. "Nike Air Zoom Pegasus 41 Men's Running Shoes -- Blue/White, Size 10" is better.
- Descriptions should be substantive. Don't just repeat the title. Explain what the product does, who it's for, and what makes it worth buying.
- Images need to be high-resolution, clean backgrounds, and show the product clearly. ChatGPT's carousel is visual-first -- a blurry or cluttered image will hurt click-through even if you get listed.
- GTINs (Global Trade Item Numbers) should be populated wherever they exist. Missing GTINs reduce feed quality scores.
- Product categories should be as specific as possible using Google's taxonomy.
Keep inventory signals accurate
ChatGPT won't recommend out-of-stock products. Make sure your feed reflects real-time inventory. If you're running on a platform like Shopify or WooCommerce, most feed tools sync automatically -- but check the sync frequency. A 24-hour lag can mean showing up for products you can't actually sell.
On-page optimization for direct crawling
The other 17% of ChatGPT carousel appearances come from OAI-SearchBot crawling your site directly. This is where standard e-commerce SEO and structured data work together.
Allow OAI-SearchBot
This is step zero. If your robots.txt file blocks OAI-SearchBot, ChatGPT cannot crawl your pages. Check your robots.txt and make sure you're not inadvertently blocking it. The user-agent string to allow is OAI-SearchBot.
If you're using a CDN or WAF (like Cloudflare), double-check that bot management rules aren't blocking it at the infrastructure level.
Implement product schema markup
Structured data is how you communicate product details directly to crawlers in a machine-readable format. Use Product schema from Schema.org and include:
namedescriptionimage(multiple angles if possible)offers(withprice,priceCurrency,availability,url)aggregateRating(withratingValueandreviewCount)brandskuandgtinwhere available
If you're on Shopify, most themes include basic product schema by default -- but it's worth auditing whether the output is complete. Many default implementations skip aggregateRating entirely.
Optimize product page content
Beyond schema, the actual text on your product pages matters. ChatGPT reads and interprets page content, not just structured data. A few things that help:
- Write product descriptions that answer the questions buyers actually ask. Not just "features" but "who is this for" and "what problem does this solve."
- Include an FAQ section on product pages. Questions like "Is this machine washable?" or "Does this work with X?" are exactly the kind of queries ChatGPT users ask before buying. If the answer is on your page, ChatGPT can surface it.
- Shipping and return policy information should be accessible and clearly written. Even though these don't directly affect ranking, ChatGPT reads them to determine whether your product is a good recommendation for a specific query.
Product images
High-quality images aren't just for human visitors. ChatGPT's carousel is image-forward, and the image pulled from your listing is often the first thing a user sees. Use images that are:
- At least 800x800px (larger is better)
- On a clean or white background for the primary shot
- Showing the product clearly without clutter
- Properly named (descriptive filenames help crawlers understand what they're looking at)
Reviews and social proof
Review data is one of the clearest signals ChatGPT uses to evaluate product quality. A product with 4.7 stars and 340 reviews will almost always outperform a similar product with no reviews in the carousel.
A few practical moves:
- Set up automated post-purchase review request emails. Most e-commerce platforms have this built in or available via app.
- Respond to negative reviews. ChatGPT summarizes "what people are saying" -- a pattern of unaddressed complaints will show up in that summary.
- Make sure your review data is included in your product schema. If reviews live on a third-party platform and aren't surfaced on your own product pages, ChatGPT may not see them.
Category pages, FAQs, and supporting content
Product pages aren't the only content ChatGPT reads. Category pages and supporting content like buying guides, comparison articles, and FAQs all contribute to your overall visibility.
If someone asks "what's the best budget espresso machine," ChatGPT might pull from your buying guide rather than (or in addition to) your product pages. A well-structured buying guide that mentions your products, explains trade-offs, and answers common questions can drive carousel appearances indirectly.
This is also where the content gap between you and competitors often lives. Your competitors might be getting cited in ChatGPT responses because they have a "best espresso machines under $200" article that you don't. Identifying those gaps is worth the effort.
Tracking your ChatGPT Shopping visibility
Here's the frustrating part: standard analytics tools don't show you ChatGPT Shopping traffic clearly. Referral traffic from ChatGPT shows up as chatgpt.com in your referral report, but that doesn't distinguish between carousel clicks and regular chat link clicks.
To actually understand your ChatGPT Shopping performance, you need tools built for this.
Promptwatch tracks ChatGPT Shopping appearances specifically -- it monitors when your brand shows up in product recommendation carousels, which prompts trigger those appearances, and how your visibility compares to competitors. That kind of page-level tracking is hard to replicate with general analytics.
For broader AI search visibility monitoring, a few other tools are worth knowing about:
Azoma is built specifically for AI shopping optimization -- covering ChatGPT, Amazon's Rufus, and other AI-powered shopping surfaces. If your business is heavily e-commerce focused, it's worth a look.
Profound tracks brand visibility across AI search engines with solid monitoring features, though it skews more toward brand tracking than product-level shopping data.
Here's a quick comparison of how these tools approach ChatGPT Shopping tracking:
| Tool | ChatGPT Shopping tracking | Product-level data | Content gap analysis | AI crawler logs |
|---|---|---|---|---|
| Promptwatch | Yes | Yes | Yes | Yes |
| Azoma | Yes | Yes | No | No |
| Profound | Partial | No | No | No |
| Standard GA4 | No | No | No | No |
The takeaway: if ChatGPT Shopping is a meaningful channel for your business (or you want it to be), you need dedicated tooling. GA4 alone won't give you the signal you need.
Platform-specific notes
Shopify
Shopify has a native Google & YouTube channel app that syncs your product catalog to Google Merchant Center. If you're using it, your products are already in the pipeline for ChatGPT Shopping via the Google Shopping route. The main things to check:
- Make sure the sync is active and error-free (the app shows sync errors in the dashboard)
- Audit your product titles and descriptions in Shopify -- these feed directly into your Merchant Center listings
- Check that your product schema is complete (Shopify's default schema is decent but often missing review data)
WooCommerce
WooCommerce requires a bit more manual setup. Use a plugin like WooCommerce Google Listings & Ads to connect to Merchant Center. For schema, the Yoast SEO or Rank Math plugins handle product schema reasonably well, but review the output to make sure aggregateRating is populated.
Custom platforms
If you're on a custom platform, you're responsible for both the Merchant Center feed (usually via a data feed file or API) and the structured data implementation. Prioritize the feed first -- it covers the 83% Google Shopping pathway -- then work on schema.
Common mistakes to avoid
A few things that consistently hold brands back from ChatGPT Shopping visibility:
Blocking OAI-SearchBot. Surprisingly common, especially on sites that added aggressive bot-blocking rules during the AI crawler wave of 2024-2025. Check your robots.txt.
Thin product descriptions. A three-sentence description that just lists specs won't give ChatGPT enough to work with. Write for the buyer, not the spec sheet.
Missing or outdated review data. If your product schema shows 0 reviews but your product has 200 reviews on your site, something is broken in your schema implementation.
Inconsistent product data across channels. If your product is called one thing in your Merchant Center feed and something slightly different on your website, you create confusion for crawlers trying to reconcile the data.
Ignoring the FAQ opportunity. Product page FAQs are one of the highest-leverage, lowest-effort improvements you can make. They directly answer the conversational queries ChatGPT users ask.
Putting it together: a practical checklist
Work through this in order -- the Google Shopping items come first because they cover the majority of ChatGPT carousel appearances:
Google Shopping (highest priority)
- Merchant Center feed is active and error-free
- Product titles are descriptive and include key attributes
- Product descriptions are substantive (100+ words)
- Images are high-resolution and clean
- GTINs are populated where available
- Inventory signals are accurate and syncing frequently
On-page and technical
- OAI-SearchBot is allowed in robots.txt
- Product schema includes name, description, image, offers, aggregateRating, brand, sku
- Product pages have FAQ sections addressing common buyer questions
- Shipping and return policies are clearly accessible
Content and reviews
- Post-purchase review request emails are active
- Review data is surfaced in product schema
- Buying guides and comparison content exist for your key product categories
Tracking
- You have a tool in place to monitor ChatGPT Shopping appearances
- You're tracking referral traffic from chatgpt.com in your analytics
The bigger picture
ChatGPT Shopping is not a separate channel that requires a completely different strategy. It's an extension of the e-commerce SEO and product data work you should already be doing. The brands that will win here are the ones with clean product feeds, complete structured data, genuine review signals, and content that actually answers buyer questions.
The free placement window won't last forever. OpenAI will almost certainly introduce paid placement options as the channel matures -- that's what Google did, what Bing did, and what every major platform eventually does. Getting your organic presence established now means you'll have a baseline to build from when that shift happens.