ChatGPT recommends products — complete with photos, pricing, and purchase links — to its 700 million weekly users.
And now customers can complete purchases without leaving the chat.
BIG deal.
But will ChatGPT recommend your products?
That’s not automatic. And you can’t pay for placement.
What you can do is optimize your site so ChatGPT understands what you sell, trusts your brand, and surfaces your products when buyers search. This guide shows you how.
You’ll learn the eight-step framework for getting featured in ChatGPT Shopping.
I also spoke with Leigh McKenzie, Backlinko’s Head of Growth and founder of the ecommerce brand UnderFit, to get his insights on what’s actually working.
First, let’s look at how ChatGPT decides which products make the cut.
How ChatGPT Shopping Works
ChatGPT Shopping kicks in automatically for some shopping intent prompts.
While it doesn’t fire every time, I found it appears more often than not after testing 100+ prompts.
The key? Typing a prompt with clear buying intent.
Like “e-bikes that can handle potholes.”
Instead of just explaining things or offering advice, ChatGPT Shopping recommends specific products.
This includes product images, pricing, and links to online stores and websites where users can make a purchase.
Side note: The ChatGPT Shopping experience isn’t consistent. Even with the same prompt, the carousel may (or may not) show. It can also appear at the top, middle, or bottom of the chat. This variability suggests the feature is still evolving.
If your store gets recommended, countless high-intent shoppers will see your products.
For example, when I tested the e-bike query, ChatGPT gave me a brief explanation of what features to prioritize.
But it also provided a visual product carousel with eight products, each in its own card with key details.
(It looks similar to Google Shopping ads, except you don’t have to pay for them.)
Clicking on any card opens a side panel with:
Additional product photos
A list of stores, prices, and direct links
A short “why you might like this” summary
Sentiment pulled from reviews and forums
From there, users simply click “Visit” to reach the merchant’s product page.
But this experience is changing.
As of September 2025, OpenAI is rolling out Instant Checkout — a feature that lets shoppers buy directly inside ChatGPT.
This is a huge shift.
ChatGPT is no longer just a product discovery tool. It’s a full shopping destination.
Right now, Instant Checkout is only available to Etsy sellers in the United States.
But OpenAI plans to expand this feature to Shopify merchants and other countries soon.
Not on either platform?
They’re also accepting applications for merchants to build their own integrations. (More on this in Step #7.)
How ChatGPT Selects Products to Recommend
A shopper describes what they’re looking for (“running shoes with arch support under $150”), and ChatGPT’s AI goes to work.
It scans the web for the most relevant products based on that request.
And weighs details like product names, descriptions, features, reviews, brand authority, and other signals to find the best matches.
If your product checks the right boxes — and the information on your site is clear and crawlable — it has a chance to be recommended.
ChatGPT may also consider the user’s location and preferences when making recommendations.
Ultimately, all product recommendations must also pass through OpenAI’s safety systems.
This filters out low-quality, misleading, or unsafe products.
So, what does all of this mean for you?
ChatGPT Shopping is evolving fast — and the brands that keep up will win the most visibility.
Here’s how to ensure ChatGPT can understand, trust, and recommend your products.
1. Add Structured Schema Markup to Your Site
ChatGPT needs structured data to understand what you sell.
Schema markup is code that labels key details on your product pages (and website as a whole): name, price, description, availability, reviews, and more.
It turns raw HTML into data AI tools can parse instantly.
Without it, ChatGPT (and other AI systems) have to guess what’s on your page.
With it, they see clean, structured information they can confidently include in product recommendations.
At a minimum, your product schema should include:
Product: Name, description, brand, image, and identifiers (GTIN, SKU, MPN)
Offer: Price, currency, availability, and URL
Review: Individual reviews with reviewer names and ratings
It may look intimidating, but many content management systems — like WordPress, Shopify, and Wix — offer plugins or built-in tools that generate the markup for you automatically.
Once your markup is in place, test that it’s working correctly using Google’s Rich Results Test or Schema.org’s validator.
These tools make it easy to check that your structured data is valid, visible, and error-free.
Pro tip: Go beyond the schema basics. Add AggregateRating for average review scores or FAQPage markup to answer common buyer questions. The more context you provide, the easier it is for AI to surface your product in response to specific prompts.
2. Create and Maintain a High-Quality Product Feed
A product feed is a structured file that packages up your product details and sends them to platforms like Google Merchant Center, Shopify, and Etsy.
It includes details like titles, prices, availability, images, links, and more.
ChatGPT may use data from major platforms like Google to decide which products to recommend.
Pro tip: Want to add your product feed directly to ChatGPT? OpenAI will notify interested merchants when this feature is available. Fill out the Merchant Application form for consideration.
For example, if your Google Shopping feed is outdated, incomplete, or inaccurate, ChatGPT may return bad information about your products.
Or skip recommending them entirely.
That’s why a high-quality, up-to-date product feed is critical.
Side note: If you’re on an ecommerce platform like WooCommerce or Shopify, feeds are usually created automatically.
But keeping feeds accurate is easier said than done.
There are a lot of moving parts, like site updates, refresh schedules, and third-party tools.
And it only takes one slip for mismatches to creep in.
Here are a few common product feed issues — and how to fix them:
Product Feed Problem
Why It Happens
Fix
Price Mismatch
Feed not refreshed, sync delay
Enable daily/real-time feed updates. Use one consistent pricing source.
Inaccurate availability
Inventory updates on site, but feed refresh lags
Sync stock levels in real-time whenever possible. Double-check before campaigns.
Wrong or Truncated Title
Feed title auto-truncated or different from H1/meta
Align feed titles with on-page H1/meta. Keep product names consistent.
Incorrect image
Feed defaults to first gallery image
Set hero/product image as primary in CMS and feed
Missing reviews
Reviews hidden in JS or not in schema
Add Review and AggregateRating schema in HTML
Conflicting schema
Multiple apps/plugins overwrite each other
Use one schema source. Validate with Schema.org or Google’s Rich Results test
Automation keeps most updates in sync. And manual checks before major launches or sales help catch anything that may slip through.
Here’s how Leigh maintains a balance of the two for his ecommerce store:
“I keep all my product data in a spreadsheet. Whenever I change a product detail, I update it there first. WooCommerce uses that data to update my site’s pages and schema automatically. Then, Channable takes the same spreadsheet and syncs those updates into my product feeds. That way, my site and my feeds are always pulling from the same source, so everything stays consistent.”
3. Make Sure AI Bots Can Read Your Site
If ChatGPT can’t read your site, it can’t recommend your products.
Two simple technical issues block many ecommerce sites from showing up: hidden content and restricted crawlers.
Check for JavaScript
Many AI bots — including ChatGPT — still struggle with content that only loads via JavaScript.
If key details aren’t in the page’s raw HTML, the bot might never see them.
This includes your product descriptions, prices, and images.
Eek.
Here’s how to check if that’s happening:
Pull up a product page on your website or online store
In Google Chrome, go to “Settings” > “Privacy and security” > Site Settings
Under “Content,” click “JavaScript” and toggle “Don’t allow sites to use JavaScript”
Reload the product page you’re testing
If your product details disappear, it means they’re only loading through JavaScript.
To fix this, work with a developer to ensure all essential information is in your site’s raw HTML.
They’re onboarding merchants on a rolling basis and will reach out when you’re accepted.
Once you’re in the pipeline, you’ll need to:
Provide a structured product feed that meets OpenAI’s product feed specs. Leigh recommends starting with your existing Google feed and updating it as needed to meet OpenAI’s requirements.
Enable ACP checkout. ACP lets ChatGPT place and complete orders in your system. If you’re on Stripe, setup can be as simple as one line of code. If not, you can still integrate using Stripe’s Shared Payment Token API or the Delegated Payments Spec — no provider switch required.
Connect your payment provider. You’ll still process transactions and remain the merchant of record.
Pass certification requirements. OpenAI requires sandbox testing and end-to-end checks before you go live.
Pro tip: Even if ChatGPT Instant Checkout isn’t available for your store yet, preparing your product data, feeds, and backend now will help you move faster when it is. This should give you a head start as this feature gains popularity.
8. Track Your ChatGPT Visibility
It’s not enough to show up in ChatGPT Shopping.
You also need to measure how well you’re performing.
Start with tracking traffic.
The easiest way is through OpenAI’s built-in UTM tag.
utm_source=chatgpt.com
This is code that OpenAI automatically adds to all outbound links. And looks like this:
Set up a custom segment in Google Analytics to track and analyze ChatGPT traffic to your site.
Once that’s done, look for patterns:
Is ChatGPT traffic increasing month over month — or slowing down?
How does the conversion rate compare to other channels?
Do visitors stick around or bounce right away?
Side note: Not every ChatGPT mention will be traceable. Some users see your product in a chat and search your brand directly on Google instead of clicking. Look for spikes in branded search traffic or direct visits to gauge the broader impact of LLMs.
But traffic only tells you what happens after people click.
You also need to measure what happens before — specifically, which prompts surface your products.
To do this, it helps to understand the kinds of prompts shoppers type.
Most fall into four buckets.
Price-based: “Best dog food bowl under $20,” “luxury ceramic dog bowl”
Use-case: “Dog bowl for messy eaters,” “raised bowls for large breeds”
Feature-based: “Non-slip stainless steel dog bowl,” “slow feeder BPA-free”
Problem-solution: “Dog bowl that keeps ants out,” “dog bowl that doesn’t slide on tile”
Think of these buckets as templates.
Test prompts in each category and ask yourself:
Does your product show up? If so, are the details accurate?
If not, who does — and why? (Are their reviews fresher, their authority stronger, or their copy closer to buyer language?)
Repeatedly run these checks to gather more data.
You’ll learn which prompts lead to product mentions, how your LLM visibility changes, and how buyers talk about your brand.
Rather automate this process?
Tools like Semrush’s AI SEO Toolkit let you:
Track which prompts surface your products
Monitor brand sentiment
Compare visibility in different platforms
Beyond ChatGPT Shopping: Your AI Visibility Playbook
There isn’t a magic formula for getting ChatGPT to recommend your products.
But the brands that consistently get recommended all have three things in common:
A rock-solid technical foundation
Clear, buyer-focused product copy
Strong trust signals across the web
Get these right, and you’re not just optimizing for ChatGPT Shopping.
You’re setting yourself up to be discovered across EVERY AI platform out there.