A practical measurement routine for checking whether products are discoverable across Merchant Center AI insights, Google Search, ChatGPT shopping, structured data, and first-party analytics.
There is no single dashboard that proves your products are visible in every AI shopping experience. Google, ChatGPT, Gemini, and other assistants use different indexes, product feeds, public pages, merchant metadata, locations, and personalization signals.
The useful question is narrower: can your team build a repeatable routine that shows whether products are discoverable, whether product data is readable, and whether traffic changes after catalog or page improvements? This guide gives ecommerce teams a practical way to check AI shopping visibility without pretending that any one metric is complete.
Start with 20 to 50 products that matter commercially: best sellers, high-margin items, new launches, and products with clear search intent. Pair each product with natural-language queries shoppers might ask, such as use case, budget, material, size, comparison, and location questions.
Record the market, language, device, date, product URL, product title, availability, price, and the exact prompt or query. Without this baseline, AI visibility checks become screenshots with no trend value.
Google announced AI performance insights in Merchant Center on May 27, 2026. The report is designed to show how products are discovered on AI Mode, AI Overviews in Search, and Gemini, including share of voice, shopping funnel performance, product term insights, and product attribute completeness. Google says rollout is coming first to the United States, Canada, Australia, India, and New Zealand.
Use this as the most direct Google-side signal when your account has access. Treat it as directional: it helps you identify product terms and missing attributes, but it does not promise future placement or cover non-Google assistants.
Google Search Central says appearances in AI Overviews and AI Mode are included in overall Search Console Web performance data. That is useful, but it is not the same as a separate AI Mode channel report.
Use Search Console to monitor query growth, product-page impressions, landing pages, and click changes around the products you are testing. Filter for long, comparison-heavy, question-like queries, but label the result as a signal, not proof. Pair Search Console with analytics so you can see whether product-page visits, engagement, and conversions move after feed or page changes.
AI shopping systems need readable product facts. Google recommends keeping structured data aligned with visible page text, keeping Merchant Center information up to date, and using Search Console to diagnose technical issues. Google also notes that both Product structured data and Merchant Center feeds can help it understand and verify product data.
Check a sample of product pages for Product structured data, visible price and availability, image quality, identifiers, variant attributes, and conflicts between the page and feed. Merchant Center product data rules warn that inaccurate, missing, or conflicting data can cause disapprovals, limited eligibility, incorrect displays, or products not showing.
Manual checks are still useful because ChatGPT shopping and Google AI experiences may show details that dashboards do not. Use clean, repeatable prompts. Run each prompt in the target market and language. Log whether your product appears directly, whether a competitor appears, whether only category advice appears, and what product facts the assistant repeats.
OpenAI says ChatGPT shopping may use structured metadata, third-party content, merchant and product metadata, public product information, and other retail sources. It also says not all available products will necessarily be shown. That means manual checks are evidence of a session, not a universal ranking result.
A useful log has five columns: product, query, surface, result, and next action. Next actions might be missing attribute cleanup, product-page copy improvement, image improvement, feed correction, Search Console inspection, or no action because the result is already stable.
Review the log monthly and after important catalog changes. Avoid chasing one-off AI answers. Look for repeated patterns: products missing attributes, pages not indexed, prices stale, variants unclear, or assistants describing competitors more clearly than your own products.
Foundax should be used for the parts merchants can control: clean product data, Product structured data on PDPs, strict Google Merchant Center preflight and sync, Search Console verification and sitemap submission, and first-party analytics with GA4 as a supplemental diagnostics layer. Foundax does not promise inclusion in Google AI Mode or ChatGPT shopping, and it does not expose a single cross-surface AI visibility score. The value is making the controlled signals easier to maintain and audit.
No. Google says AI Overviews and AI Mode are counted in overall Search Console Web performance data. Use it for directional monitoring, not as an separate AI Mode channel report.
Start with Merchant Center AI insights if your account has access. Then check Merchant Center diagnostics, Search Console, structured data, manual AI shopping prompts, and first-party analytics.
No. AI shopping results can vary by time, location, product availability, user context, and prompt wording. Track repeated patterns instead of one result.
Use repeatable shopping prompts, target markets, and product terms. Log whether your product, a competitor, or only category advice appears. Compare the product facts ChatGPT repeats with the facts on your page and feed.
Track Product structured data, Merchant Center preflight issues, feed sync status, Search Console sitemap and landing-page signals, product-page engagement, and conversion changes.
Recheck after feed, price, availability, template, image, or localization changes. For stable catalogs, a monthly review is usually enough.