ChatGPT Shopping: How to Prepare Your Google Merchant Center Feed

Introduction

A user types into ChatGPT: "find me a waterproof hiking jacket under $130, ideally lightweight." Within seconds, they get a selection of products with a photo, price, brand and purchase link — without ever opening Google.

This scenario is no longer hypothetical. ChatGPT has gradually integrated shopping features directly into conversation: product recommendations, comparisons, and for some merchants, a purchase flow that completes without ever leaving the chat interface. For an e-commerce business, this opens a new product discovery channel — but one that plays by its own rules.

The question many merchants are asking too late is a simple one: how do you get your products to show up in these responses? The answer largely comes down to an asset you may already own: your Google Merchant Center feed. Its quality, completeness and compliance no longer serve Google Shopping alone — they also shape your visibility across the new surfaces of AI-assisted, conversational shopping.

This guide explains what ChatGPT Shopping actually is, where the product data it relies on comes from, why GMC compliance is the foundation, and how to prepare your feed for this new channel — without losing sight of the fundamentals that remain, regardless, those of GEO. For a broader view of optimizing your product pages for AI engines, see our complete e-commerce GEO guide.

What ChatGPT Shopping actually is
Where the product data used by ChatGPT comes from
Why GMC feed compliance is the prerequisite
ChatGPT Shopping vs traditional Google Shopping SEO
GMC feed optimization checklist for ChatGPT Shopping
FAQ

What ChatGPT Shopping Actually Is {#what-is-chatgpt-shopping}

ChatGPT Shopping refers to the set of features that let a user search for, compare and — in some cases — buy products directly within a conversation with ChatGPT, rather than going through a traditional search engine or a marketplace.

In practice, the user experience unfolds in three stages:

The conversational query: the user describes a need in natural language, with constraints ("under $130," "size 9," "for running") rather than typing isolated keywords.
The enriched recommendation: ChatGPT presents a selection of products as visual cards — image, title, price, brand, sometimes rating and reviews — along with a brief explanation of why each product fits the request.
The transition to purchase: depending on the merchant and the partnerships in place, the user may click through to the merchant's product page to complete the order, or, for some integrated merchants, complete the purchase without leaving the chat interface.

This last capability — a purchase flow entirely inside the chat — relies on payment and ordering protocols designed to let AI agents securely transmit an order to a merchant. OpenAI has announced this type of partnership with several e-commerce platforms, but the exact scope and availability of this feature vary by market, by industry, and by the technical integrations already in place at each merchant. For most e-commerce businesses at this stage, the main challenge remains showing up in product recommendations — direct in-chat purchase concerns a smaller number of integrated merchants.

What fundamentally changes compared to a traditional Google search:

The response is synthesized, not a list of links to browse
The user expresses a complete intent rather than isolated keywords
The number of products shown is deliberately limited (a handful of relevant options rather than dozens of results)
The trust placed in the data source is an explicit selection criterion, since the AI has to "choose" which products to surface without the user clicking through to compare on their own

For a merchant, this means that being visible in ChatGPT Shopping isn't a question of ranking position — it's a question of being judged reliable and relevant enough to make it into a narrow shortlist.

Where the Product Data Used by ChatGPT Comes From {#origin-of-product-data}

This is the most important question — and the one where caution is warranted, since the exact mechanisms are evolving and not all of them are fully documented publicly. Here is what can be stated with a reasonable level of confidence, distinguishing what is established from what remains to be confirmed.

Product feeds supplied by merchants

According to publicly available information, OpenAI has set up mechanisms that let merchants directly supply structured product data feeds — in a format that largely mirrors the same principles already used across e-commerce: product ID, title, description, price, availability, image, brand, and in many cases a universal product identifier such as a GTIN. This choice is not incidental: these are largely the same attributes that Google Merchant Center has required for years.

In practice, this means an already clean, complete and compliant GMC feed forms a directly reusable base of work — if not for an identical technical submission, then at least for the data-structuring logic to apply.

Partnerships with e-commerce platforms

OpenAI has announced partnerships with several major players in e-commerce infrastructure (online sales platforms, payment solutions). These partnerships are intended to make it easier for merchants already using these platforms to expose their product catalog to ChatGPT without heavy additional development. A merchant on a partner e-commerce platform would thus have a more direct path to exposure in ChatGPT Shopping than a merchant on a proprietary, non-integrated stack.

Public web content

As with other conversational engines, it is reasonable to assume that ChatGPT also draws, at least in part, on publicly indexed data — product pages, Schema.org structured data present on those pages, public customer reviews. This source is less directly controllable by the merchant than a voluntarily submitted feed, but it remains influenceable by working on the quality of markup and product page content — exactly as with GEO more broadly.

What remains uncertain

It would be unwise to state with certainty the exact weighting between these sources, the frequency at which data is refreshed, or the precise algorithmic selection criteria ChatGPT uses to decide which products to recommend. These mechanisms are evolving quickly and are not fully documented publicly. What is established, however, is the general principle: the more structured, complete and consistent your product data is across multiple channels — merchant feed, page structured data, information shown to the user — the more you multiply your chances of being accurately represented, regardless of the precise technical mechanism used to select you.

This is exactly why a well-maintained Google Merchant Center feed is so valuable: it is no longer just an entry ticket for Google Shopping — it becomes a reusable product database for the entire AI-assisted shopping ecosystem.

Why GMC Feed Compliance Is the Prerequisite {#gmc-compliance-prerequisite}

Some merchants might be tempted to treat ChatGPT Shopping as a separate project requiring a dedicated technical integration before even worrying about the compliance of their existing Google Merchant Center account. That's a prioritization mistake.

A compliant feed is a reusable feed

The attributes that make a feed compliant with Google Merchant Center requirements — a descriptive, non-promotional title, a detailed and honest description, a valid GTIN, up-to-date availability, a price consistent between the feed and the page, compliant images, correct categorization — are exactly the same attributes that make a feed usable by any third-party system that comes to rely on it, ChatGPT Shopping included.

Conversely, a GMC feed with active disapprovals, missing attributes or price inconsistencies reveals an underlying problem in how product data is managed — a problem that, by construction, affects any other surface that consumes that same data.

Data consistency is a universal trust signal

One principle shows up consistently across every system that synthesizes answers from product data, whether it's Google AI Mode, Gemini or ChatGPT: inconsistencies between sources (a different price between the page and the feed, "in stock" availability when the product is actually sold out, a misleading title) reduce the confidence placed in that data. An AI synthesizing a response for a user — with the risk of recommending an unavailable or poorly described product — has a structural incentive to favor sources whose reliability is best established.

A Google Merchant Center account with a clean compliance history, few disapprovals and up-to-date data sends a reliability signal that extends well beyond Google Shopping alone.

Rich attributes become more valuable, not less

In a conversational search, the user often expresses precise constraints ("waterproof," "under $130," "size 9," "compatible with..."). For a product to be selected in response to this kind of query, the attributes that allow it to be filtered and finely described — color, size, material, product type, detailed characteristics — need to be filled in the feed. A feed that sticks to the bare minimum (title, price, link, image) mechanically limits the cases where the product can match a complex query.

The priority is the same as for GEO

This point connects directly to what we cover in our e-commerce GEO guide: GMC feed compliance and richness form the common foundation for every AI-engine visibility strategy, ChatGPT Shopping included. Investing in an advanced technical integration before fixing active violations on your GMC account is building on unstable foundations — a problem we break down attribute by attribute in our guide to mandatory GMC feed attributes.

Key point: you don't "configure" your visibility in ChatGPT Shopping independently of your GMC compliance. You improve the quality of your product data once, at the source, and that quality propagates to every channel that consumes it — current and future.

ChatGPT Shopping vs Traditional Google Shopping SEO {#chatgpt-shopping-vs-traditional-seo}

It's worth clarifying how this new surface differs from traditional Shopping SEO, to avoid applying the wrong reflexes.

Dimension	Traditional Google Shopping SEO	ChatGPT Shopping
Result format	List of products, ad grid, free-form visual comparison	Narrow selection, often 3 to 5 products, with a rationale
User's role	Browses, filters and compares on their own	Delegates part of the pre-selection to the AI
Main lever	Title, price, bidding (for paid), feed relevance	Completeness and consistency of product data, source trust
Merchant control	Bids, budget, precise targeting via Google Ads	Little to no direct advertising lever at this stage
Performance measurement	Impressions, CTR, position, ROAS	Hard to measure directly — no mature equivalent to GMC/Ads reporting yet
Update cycle	Relatively frequent recrawl and re-evaluation	Depends on ChatGPT's own refresh mechanisms, less documented
Dependence on a GMC account	Mandatory for Shopping and Performance Max	Not strictly mandatory depending on the mechanism in place, but a high-quality structured feed — often the same one built for GMC — remains the decisive asset

What stays the same

Traditional Shopping SEO and ChatGPT Shopping share a common foundation: the intrinsic quality of product data. A feed with precise titles, honest and detailed descriptions, up-to-date prices and compliant images performs better on both fronts. At this stage, there is no "ChatGPT-specific" strategy that would run counter to a solid GMC strategy — there is a high-quality product data strategy, which then plays out across each channel, including AI-driven campaigns like Performance Max, which relies on the same feed-completeness requirements.

What changes

The main shift in posture concerns measurability and control. With Google Ads and Google Shopping, you have direct levers — bids, budgets, exclusions — and detailed performance reports. With ChatGPT Shopping, at this stage, merchant control is more indirect: you optimize the quality of your data upstream, without necessarily having the same dashboards to precisely measure your visibility. This makes it all the more important to treat feed quality as a permanent asset rather than a one-off campaign you could adjust in real time based on performance metrics.

GMC Feed Optimization Checklist for ChatGPT Shopping {#optimization-checklist}

This checklist covers the fundamentals of GMC compliance, with emphasis on the points that matter most for being accurately represented in a conversational search.

Account compliance and reliability

No active disapprovals in Google Merchant Center diagnostics
Account history with no recent suspension
Return policy filled in and up to date in GMC
Complete and verifiable company information (address, contact, legal notices)
Legal pages (T&Cs, privacy policy) accessible from the site

Product data consistency

Price identical between the GMC feed, the product page and any Schema.org structured data
Availability synchronized in real time (or at minimum every 24h) with actual stock
Product title identical or consistent across the feed, the page and structured data
No mismatch between the images in the feed and those shown on the page

Attribute richness and completeness

Valid GTIN filled in for every branded product
brand filled in consistently
product_type and Google Product Taxonomy category correctly mapped
Complete variant attributes: color, size, material as relevant to the category
Detailed description (at least 300-500 words) that answers the concrete questions a buyer would ask about the product
At least 3 images per product, including one usage-context image

Content and trust signals

Customer reviews structured in Schema.org with average rating and a sufficient review count
Product FAQ on the page for common questions (compatibility, care, sizing)
Product Schema.org structured data present and consistent with the GMC feed
Description free of excessive promotional language or all-caps text (the same rules that apply to GMC likely apply to any data source consumed by an AI)

Ongoing monitoring

Monthly check of GMC approval rate (target above 95%)
Process in place to update the feed as soon as a price or stock level changes
Watch for new feed features and requirements published by conversational shopping platforms

A feed that checks all of these boxes isn't just better prepared for ChatGPT Shopping — it's better prepared for the entire range of AI-assisted shopping channels that will continue to emerge, as well as for Google Shopping and Performance Max themselves. This is the basic logic of GEO applied concretely to your product feed: see our complete GEO guide to go further on structured data and conversational content.

FAQ {#faq-en}

Is ChatGPT Shopping available in France? The exact availability and scope of shopping features in ChatGPT vary by market and change regularly. Rather than relying on a fixed statement about geographic availability, the best approach for a merchant is to prepare their feed now — the quality of your product data never loses value, whatever the rollout pace of these features in your market.

Do I need a Google Merchant Center account to appear in ChatGPT Shopping? No, a GMC account is not a mandatory technical requirement for ChatGPT Shopping, which can rely on its own feed submission mechanisms or on partnerships with e-commerce platforms. That said, if you already have a compliant, well-maintained GMC account, you already have a structured, high-quality product database — largely reusable for any other channel, ChatGPT Shopping included.

What's the difference with traditional Google Shopping SEO? Traditional Shopping SEO optimizes your position in a list of results that the user browses themselves, with direct levers like Google Ads bidding. ChatGPT Shopping presents a narrow, AI-synthesized selection of products, where the trust placed in your data and its completeness matter more than traditional advertising levers. Both approaches, however, share the same foundation: accurate, complete and consistent product data.

Do I need to build a separate feed for ChatGPT in addition to my GMC feed? It depends on the integration mechanism available for your e-commerce platform at the time you're reading this — some merchants go through partnerships built into their platform, others through a dedicated feed submission. Either way, the data structure you've already built for GMC is the best starting point: the principles of completeness and consistency carry over, even if the exact technical format may vary.

Do customer reviews impact my visibility in ChatGPT Shopping? It's reasonable to think so, given that structured reviews are a trust signal widely used by systems that synthesize answers from product data, as is already the case for Google AI Mode and Gemini. A product with verified reviews, a solid average rating and a sufficient review volume has a structurally better chance of being perceived as reliable, regardless of which conversational engine is used.

Is direct in-chat purchase already available to all merchants? No. According to available information, this capability currently applies to a smaller number of merchants integrated through partnerships or dedicated payment protocols, and its rollout varies by industry and market. For the vast majority of merchants, the immediate priority remains appearing in product recommendations rather than handling a full transaction inside the chat.

How can I tell if my products are already being cited by ChatGPT? The simplest method remains manual testing: ask ChatGPT shopping-style queries matching your catalog and see whether your brand or products show up. To our knowledge, there is not yet a standardized measurement tool equivalent to Google Merchant Center's performance reports for this specific channel.

Prepare Your Feed Before Optimizing the Channel

ChatGPT Shopping, like other conversational shopping surfaces, rests on a simple but demanding principle: structured, complete and trustworthy product data. That's exactly what Google Merchant Center has required for years — which is why a compliant GMC account isn't an isolated box to check, but the single most cost-effective investment you can make before turning to an emerging channel.

Merchants who wait for a "perfect" technical integration with every new AI platform, without first fixing the active violations on their GMC account, are needlessly delaying their visibility across all of these channels — present and future.

The first step remains the same as for any AI-visibility strategy: a complete audit of your feed's compliance. MyGoogle analyzes your product page in 30 seconds and identifies the violations to fix first.

Launch the free MyGoogle audit — check your Google Merchant Center feed's compliance and prepare your catalog for ChatGPT Shopping and the next wave of conversational shopping channels.