Is AI Showing the Right Price for Your Products? A 30-Minute Check
More shoppers now ask an AI assistant what to buy before they visit a store. The answer may name your product, quote a price, describe a feature and link to a seller. When that answer uses old or incomplete data, a customer can arrive expecting a different price, find the size they wanted is gone, or be sent to a reseller or competitor instead.
You can't edit an AI answer. What you can do is keep the product data behind it accurate, and check what it says on a schedule. That's what this guide covers.
What you should take away
Once a month, you'll spend 30 minutes asking AI shopping tools the questions your customers ask. Write down every wrong price, stock status, spec or link. Then fix the source: your product page, your product feed and your structured data, so all three say the same thing.
Know which AI channels can reach an Australian store
Coverage changes by provider and country, so it's worth checking before you plan around a channel:
- Google AI Mode and AI Overviews: Google Merchant Center now reports how products appear for shopping questions in AI Mode and AI Overviews, and lists Australia among the supported countries. Google points merchants to its standard product listing requirements.
- Shopify agentic storefronts: Shopify can share eligible products with AI channels. Its Google AI Mode and Gemini channel accepts stores selling to customers in Australia, the United States or Canada.
- ChatGPT shopping: Shopify's ChatGPT channel currently requires selling to customers in the United States, and OpenAI's product feed format currently targets the US. ChatGPT search can still read your public pages if its search crawler is allowed.
These rules move quickly. Recheck the provider pages before you rely on one.
Run the 30-minute check
- Pick ten products. Choose best sellers, high-margin items and anything with frequent price or stock changes.
- Write the buying questions. Use plain customer words, for example "best size 5 soccer ball under $60 in Australia" or "waterproof hiking boots for wide feet".
- Ask the tools. Use Google AI Mode, Gemini and ChatGPT in a logged-out or private window so past searches have less influence.
- Record what you see. Note whether your product appears, the price shown, stock shown, key specs and where the link goes.
- Compare with your live page. Mark each answer as correct, wrong, missing or sent elsewhere.
Copy this error log
- Date and tool: for example AI Mode, Gemini or ChatGPT.
- Question asked: the exact words used.
- Product: name and SKU.
- What the answer said: price, stock, specs and link.
- What is true: the same fields from your live product page.
- Error type: missing, wrong price, wrong stock, wrong spec, or linked to another seller.
- Likely source: product page, feed, structured data or marketplace listing.
- Fix and owner: what changed, who did it and when to recheck.
Fix the source, in this order
1. Make price and stock agree everywhere
Google requires the price in your product data to match your product page and checkout, and products that fail its requirements can be disapproved. Check that sale prices, currency and stock status match across your store, your Merchant Center feed and any marketplace listing.
2. Complete the product identifiers
Fill in brand and, where the product has one, the GTIN (the barcode number). Use the manufacturer part number as the fallback. Clear identifiers help shopping systems match your product to your own listing.
3. Check your structured data
Product pages should include product markup with name, image, price and currency, plus availability, brand and GTIN where possible. Google also recommends shipping and return policy details. Your theme or an app may already add it; test a few product pages with Google's Rich Results Test to confirm.
4. Write specs in plain text
Put sizes, materials, compatibility and what's in the box in the description as text, as well as in images. A model repeats the specs it can read.
5. Review crawler settings
If your robots.txt blocks AI search crawlers, those tools lean on older third-party information about your products. OpenAI, for example, says blocking OAI-SearchBot keeps a site out of ChatGPT search answers, while GPTBot is used for training. Decide each setting on purpose.
Limits to keep in mind
AI answers vary between sessions and people, so treat one wrong answer as a signal worth checking. A clean feed improves your odds; recommendations stay the provider's call. Put your effort where you have control: accurate, consistent and complete product data, checked on a schedule.
Takeaway
Think of AI shopping answers as a new shop window. You don't arrange it, but you do supply what goes in it. Ten products, five questions each, once a month, with an error log and a named owner for each fix.
Want help checking your product feed, schema and AI shopping visibility? Contact Maud Digital to review your store data before the busy season.
Source note: This original Maud Digital guide uses Google's AI performance insights help page, product data specification, price mismatch policy and merchant listing structured data guide; Shopify's agentic storefronts, Google AI Mode and Gemini channel and ChatGPT channel pages; and OpenAI's crawler documentation. Provider eligibility was checked on 16 November 2026 and can change.
General information to help you decide, based on public sources at the date shown. For legal, tax or compliance questions about your business, speak with a qualified adviser.