How to Get Your Products Into ChatGPT and Google AI Mode Shopping Answers.

To get your products into ChatGPT and Google AI Mode shopping answers, give both systems clean product data from two directions. Send a product feed (Google Merchant Center for AI Mode; Shopify Catalog or OpenAI's merchant feed for ChatGPT). Keep product pages crawlable, with structured data that matches the page. Then earn mentions in the reviews they cite.

None of that is a trick. It is the product data discipline good ecommerce SEO already needs, pointed at a new reader: a model building an answer.

This guide covers what AI shopping looks like right now, where the assistants get their product information, what each merchant program asks for and how to check whether your products appear. These surfaces change monthly, so every platform claim below is dated to when we checked it: 4 November 2026.

1. What AI shopping is right now

ChatGPT. When a question has shopping intent, ChatGPT shows product options with images, details and links to merchants. OpenAI's Shopping with ChatGPT Search help article says products are chosen on relevance to the request, using context such as price and reviews. It also says product results "are not ads, nor influenced by any OpenAI partnerships".

There is one catch for Australian stores. OpenAI's merchants page says shopping is currently live for users in the US and will expand to more regions over time. OpenAI's product feed specification says the standard upload "currently targets the US". In our own test on 4 November 2026, an Australian shopping prompt in ChatGPT still returned product cards with A$ prices and links to Australian retailers, built from their public product pages. So Australian stores can appear today without a feed. An Australian store can also prepare for the feed program now. Getting into the feed program is a separate question, which we come back to below.

Google AI Mode. Google launched AI Mode in Australia in English on 8 October 2025. Google described it as drawing on "shopping data for billions of products" alongside its other sources. In November 2025 Google described AI Mode shopping as answering product questions with visuals, prices, reviews and inventory details. That post puts the Shopping Graph behind it at more than 50 billion product listings.

Checkout inside the assistant. Both companies are building ways to buy without leaving the chat. OpenAI's Agentic Commerce key concepts says Instant Checkout in ChatGPT is available to approved partners. Google's Universal Commerce Protocol help article covers a checkout button on eligible listings in AI Mode and Gemini. It applies to products eligible in the United States, Canada and Australia. The feature "is available for select merchants at this time". For most Australian stores today, the realistic goal is to be recommended and clicked through, not to sell inside the chat.

Perplexity. Perplexity's Shop like a Pro help article says its in-app purchasing is available to Pro subscribers in the United States. It says product listings are chosen by its algorithm on relevance and ratings. That article was last updated well over six months before we checked it, so treat it as the floor of what Perplexity does, not the ceiling.

2. Where AI assistants get product information

Every assistant builds its shopping answer from some mix of the same five inputs:

1. Merchant feeds. Structured files of products, prices and stock that merchants send to the platform.
2. Retailer product pages. The page itself, read by the platform's crawler, including its structured data.
3. Reviews. OpenAI's help article says ChatGPT's review summaries are model-generated from public review sites.
4. Publishers and listicles. Best-of lists, buying guides and comparison articles that the answer cites as sources.
5. The platform's own index. Google's Shopping Graph; for ChatGPT, structured metadata from first-party and third-party providers.

Here is how that maps to each surface, as checked on 4 November 2026:

No single input is enough. A perfect feed with no third-party coverage gives the answer nothing to cite. Great reviews with broken product markup give it a reason to mention you but no clean price or link to show.

3. Merchant programs and feeds

Google Merchant Center

For Google, the feed is Merchant Center. Products in it can appear in free listings at no cost across Google Search, the Shopping tab, Images, Lens, Maps, YouTube and Gemini. The required attributes for every product are ID, title, link, image link, price, description and availability. Condition, brand, GTIN or MPN are required in some cases. Australia is one of the countries where shipping settings are mandatory. Google also asks for return policy information.

AI Mode is not named as a separate free listings surface on that page. Google's AI features guidance does tell site owners to keep Merchant Center information up to date as part of doing well in AI features. Google's own announcement ties AI Mode shopping to the Shopping Graph. Merchant Center is the one input to that graph you control directly.

Two things to get right in the feed:

• Attributes shoppers ask about. Google's May 2026 announcement of AI performance insights includes an "attribute completeness" score for specifications such as colour, style and material. That tells you which fields Google expects AI surfaces to lean on. Fill them.
• Feed matches page. Price and availability in the feed should match the product page at all times. Keep them in sync so no surface ever shows a price or stock level you do not honour.

OpenAI (ChatGPT)

ChatGPT has two routes in. If your store runs on Shopify or Etsy, OpenAI says your catalogue is already integrated through the platform and no application is needed. Everyone else applies through the merchants page, which says applicants go onto a waitlist. It accepts feeds over SFTP, APIs and common feed providers.

OpenAI's product feed specification is worth reading before you are accepted. Nine fields are required for discovery: item ID, title, description, URL, brand, seller name, image URL, availability and price. Recommended fields include GTIN, MPN, sale price, condition, product category, extra images, shipping, returns and review aggregates. The spec says to omit a GTIN rather than invent one. It also says search eligibility "does not guarantee display".

Additional markets need OpenAI to confirm the integration and its allowed countries and currencies. An Australian store not on Shopify or Etsy should apply and build the feed, without expecting an immediate result.

Perplexity

Perplexity's help article says merchants who give it more product detail (availability, reviews and specs) are more likely to be recommended. Retailers can apply through a form. We have not been able to confirm its current eligibility for Australian stores.

4. On-site signals

A feed tells a platform what you sell. Your product pages are what its crawler reads and what the answer links to. Four things matter most.

Crawler access. OpenAI's crawler documentation says OAI-SearchBot is used to surface websites in ChatGPT's search features. Sites that block it will not appear in ChatGPT search answers. Check your robots.txt allows it. For Google, a page must be indexed and eligible for a snippet to appear in AI Overviews or AI Mode, per the AI features guidance. Some CDN and security settings block AI crawlers by default, so check rather than assume.

Product structured data. Product markup with a nested Offer (price, currency and availability) is what Google reads for merchant listing experiences. Only pages where a shopper can buy the product are eligible. Google is explicit that AI features need no special schema. The markup you already need for Search is the markup that counts. It must match the visible page. We cover how to plan this across a catalogue in our guide to schema mapping.

Specs in text, not images. Shoppers ask assistants narrow questions: "waterproof hiking boots with a wide toe box under $200". An assistant can only match that if the width, waterproofing and price are written on the page as text. Specs locked in an image or a PDF size chart are invisible to it. Google's AI features guidance makes the same point: keep important content in text form.

Comparison tables and FAQs. AI Mode runs several related searches behind one question, a technique Google calls query fan-out. A product page or buying guide that answers the follow-up questions ("how does it compare to X", "is it worth it for beginners") gives the assistant more to match. We explain the mechanics in our note on query fan-out.

5. Off-site signals

This is the easiest part to skip. It is often the part that decides the answer.

When an assistant recommends products, it usually cites review sites, publishers and best-of lists. If your product is in none of them, the assistant has nothing to quote about you, however good your feed is.

What to do about it:

• Find the sources it cites now. Run your category prompts (section 6) and record every URL the answers cite. That list is your outreach list.
• Get reviewed where shoppers read reviews. Review platforms, Google reviews and category publications. Send product; never script the review.
• Be the maker, not just a reseller. OpenAI's help article says merchant ranking considers whether the merchant is the maker or primary seller. If you make the product, say so clearly on the page and in the feed's brand and seller fields.
• Earn links from real publishers. Digital PR that builds authority for Search also builds the coverage AI answers draw on.

6. How to check whether your products appear

You cannot fix what you have not measured. The method we use:

1. Write ten prompts a real shopper would type. Mix the shapes: "best X for Y", "X vs Z", "X under a set budget", "X for a specific problem", "where to buy X in Australia" and "gift ideas for someone who likes X". The prompt test generator below writes them for you.
2. Run each prompt in ChatGPT, Perplexity and Google AI Mode. Use a logged-out or fresh session where you can, because memory and history change the answer.
3. Record three things per answer. Does your brand or product appear? Which competitors appear? Which sources are cited?
4. Repeat on a schedule. Answers vary between runs, so a single test is a snapshot, not a result. Monthly is enough to see a trend.

That table is your baseline for measuring every later change. Our guide to AI visibility tracking covers how to do this at scale. Once Merchant Center's AI performance insights reach Australia, they should add Google's own share of voice data on top.

Build your AI shopping test prompts

7. What nobody knows yet

Be wary of anyone selling a ranking formula for AI shopping. Here is what is not confirmed:

• Ranking factors. OpenAI lists factors it considers (availability, price, quality, maker or primary seller). It does not publish weights. Google says there is no special optimisation for AI features. Neither platform has published a ranking formula for AI shopping answers.
• How much feeds count against crawled pages. OpenAI's merchants page says feeds give "greater control" than crawling alone. It does not say how the two are weighted.
• Australian timing. ChatGPT's feed program is US-first, though Australian queries already show product cards built from public pages. In-chat checkout is limited to select merchants on both platforms. Dates for wider Australian rollout have not been published.
• Answer stability. The same prompt can return different products on different days. Treat any single screenshot, including your own, with caution.

What we do know is that every platform asks for the same things: accurate product data, crawlable pages, honest reviews and trusted coverage. That work holds its value however the details change.

How we approach it for stores

We treat AI shopping visibility as part of the product data work: product template and markup first, then the feed, then third-party coverage. For 2XU we fixed templates before building anything new, which made it possible to add 800+ indexed pages in six months. Over the same period organic revenue rose 50% and average order value rose 53%. That was search work, not AI shopping work. It is the same foundation. You can read the 2XU case study for the detail.

If you want to know whether AI assistants recommend your products today, the next step is an audit. Request an audit and we test your category prompts across ChatGPT, Perplexity and Google AI Mode alongside the full crawl. Our ecommerce SEO audit guide shows what else the audit covers. The audit is the first stage of the Ecommerce SEO add-on, scoped on catalogue size. For the broader AI search work behind it, see our AI SEO agency page or our explainer on what GEO is.

Common questions

Can I pay to appear in ChatGPT shopping results?

Not in the product results. OpenAI's help article says they are selected independently and are not ads. Its feed specification carries a separate ads eligibility flag, because ads are handled apart from product results.

Do I need a product feed if my site is already crawled?

It is not the only route but it gives you control. OpenAI's merchants page says feeds give greater control over how products appear. On Google, Merchant Center is how your price, stock and shipping reach its shopping surfaces.

Is ChatGPT shopping available in Australia?

As at 4 November 2026, OpenAI's merchants page says shopping is live in the US and will expand to more regions. Its feed upload targets the US by default. Shopify and Etsy stores are already integrated through their platforms. Other Australian stores can apply and prepare their feed now. In our test on 4 November 2026, an Australian query already returned product cards with A$ prices from Australian retailers.

Does Google AI Mode use Merchant Center data?

Google describes AI Mode shopping as built on its Shopping Graph. Its AI features guidance tells site owners to keep Merchant Center information up to date. Merchant Center is the direct way to put your product data into Google's shopping systems.

Do I need special schema for AI shopping?

No. Google says there is no special structured data needed for AI features. Standard Product markup with an Offer is what matters, kept consistent with the visible page and with your feed.