Skip to content
Skip to content
SEO & GEO

Google has made AI search visibility a product-data question

Google's AI performance insights report gives merchants a clearer view of discovery inside AI Mode and AI Overviews. The useful shift is not another score to chase. It is a reminder that AI search visibility depends on product evidence, query intent and the limits of the report itself.

Summarise with
Set as preferred source on Google
15 min read
Quick answerGoogle's Merchant Center AI performance report shows share of voice, frequency and shopping stages for organic AI traffic in five English-language markets. It is a scoped diagnostic, not a universal AI ranking score.
SECTION 01

AI performance insights change what visibility means

The Google Merchant Center help page describes AI performance insights as a report for conversational shopping queries with shopping intent on AI Mode and AI Overviews. It is designed to show how brands are discovered across Google's generative AI ecosystem, then break that discovery into useful shopping stages. For an SEO team, the important development is that AI search visibility is being described as an evidence problem rather than a mysterious impression. The official Merchant Center guidance is the source for the current scope and metrics.

That does not mean Google has given every site a universal answer to the question, 'How visible am I in AI?' The report is tied to Merchant Center, product categories, eligible markets and English-language queries. It is a window, not the whole house. A useful AI search visibility programme should therefore keep three layers apart: what Google reports, what a customer can reproduce in a real query and what the business can prove about the product itself.

The first layer is measurement. Google lists share of voice, competitors' average share, frequency and products showing among its metrics. The second is interpretation. A product can appear for a high-frequency query while being the wrong answer for a buying stage. The third is evidence. If a feed omits material, size, compatibility or delivery information, the model has less reliable material to use. This is why an AI visibility tracking tool should never become a substitute for inspecting the page and feed that the tool is measuring.

The change is strategically useful because it turns a vague executive request into a set of questions. Which product category is being discovered? For what kind of conversational query? At which stage? Against which competitor set? With how many products showing? A credible AI search visibility review can answer those questions without pretending to know every answer generated for every person.

AI search visibility illustration with product cards and conversational search bubbles
AI search visibility is a window into discovery, not a universal ranking score.

shopping stages

3

Discovery, Evaluation and Ready to buy are named in the scorecard

eligible markets

5

Google lists Australia, Canada, India, New Zealand and the United States

traffic scope

1 organic AI

Paid Ads traffic is excluded from the current report

SECTION 02

What merchants can see, and what they still cannot

Google's report has a concrete operating path: sign in to Merchant Center, open Analytics, choose Products and select the AI performance tab. The metrics then provide a view of how often a brand or product is surfaced relative to a defined competitor set, how frequent a search type or attribute is, and how many products appear for the top terms. That is enough to make AI search visibility actionable for a feed owner. It is not enough to judge the complete quality of an answer produced outside the report.

Share of voice is a relative measure. Google says it is based on a brand or product's AI impressions divided by the total impressions across the defined competitor set and the merchant's own results for related queries. Relative measures are useful when the set and time period stay stable. They become misleading when the category, competitor list or date range changes without a note. An AI visibility checker should show those boundaries beside the percentage rather than presenting it as a fixed market share.

Frequency is also easy to misread. Google describes it as a representation of how popular specific search types, search terms, search intent or attributes have been. A frequent query may be broad and early-stage. A less frequent query may be close to a purchase. The sensible reporting habit is to read frequency beside shopping stage, products showing and the actual query framing. That is more informative than asking a best AI visibility tool to sort every opportunity by one number.

The report also excludes paid Ads traffic and focuses on free listings. That boundary is valuable. It keeps an organic AI discovery diagnostic separate from paid media performance. It also means a merchant should not use the report to explain every commercial impression or every campaign result. Google's product data specification and free listings guidance belong in the same review because the product evidence is part of the visibility question.

For a Malta or EEA team, the availability restriction is not a small footnote. If the report is only available for English-language queries in five named markets, a local strategist cannot infer that a missing result in another market means the product is invisible. The Search Console generative AI performance guidance is a separate surface and should not be silently combined with Merchant Center figures.

The report's declared scope
Five countries, three shopping stages and no paid Ads traffic in the current Merchant Center AI performance report. This is scope, not performance.eligible countries named: 5shopping stages: 3paid traffic included: 064205eligible countriesnamed3shopping stages0paid trafficincluded
Five countries, three shopping stages and no paid Ads traffic in the current Merchant Center AI performance report. This is scope, not performance.
ItemValue
eligible countries named5
shopping stages3
paid traffic included0
Five countries, three shopping stages and no paid Ads traffic in the current Merchant Center AI performance report. This is scope, not performance.
SECTION 03

The feed is the evidence layer

It is tempting to treat conversational search as a copywriting challenge. The more dependable starting point is the product feed. A customer can ask an AI surface for a waterproof coat for a particular temperature, a phone with a specific camera capability or a chair that fits a measured space. If the Merchant Center record does not contain the relevant facts, the system has less material to match against the request. Improving AI search visibility therefore starts with completeness and consistency, not with sprinkling the word AI across a description.

The report's references to search intent, terms and attributes are a useful prompt to inspect the relationship between the query and the product evidence. Treat the relationship as a hypothesis. A product appearing for a term does not prove that the description caused the appearance, and a product missing from a term does not identify the missing field on its own. An AI visibility tracking tool can point to a pattern. A human still has to open the feed, product page, availability rule and category mapping to explain it.

Structured data helps, but it is not a guarantee. Google's structured data documentation explains how markup can make page information machine-readable. Google's Search Essentials provide a broader set of technical and spam boundaries. A good AI search visibility workflow checks whether the page, feed and structured data tell the same story, then records where an inference begins.

There is also a customer trust layer. Product attributes should be accurate, material limitations should not be hidden and availability should not be implied if it is conditional. A model may generate a neat answer from incomplete information, but that neatness is not proof. The product owner, merchandising team and legal reviewer need a shared escalation route for claims that affect safety, suitability, price or a person's ability to make an informed purchase.

This is where an AI visibility checker can become genuinely useful. Use it to find a product family with low showing rates, a query stage where competitors dominate or an attribute that appears frequently in customer language but rarely in the feed. Then make one controlled improvement and record the date, changed fields, markets and result. The workflow is closer to technical SEO and merchandising than to a leaderboard.

Choose the product family

Start with a category where the business can name a commercial question and an owner.

Inspect the query stage

Separate discovery, evaluation and ready-to-buy patterns before deciding what good visibility means.

Audit the evidence

Compare the feed, product page, structured data, availability and terms for the attributes customers need.

Change one layer

Improve one material field or page section, record the date and keep the competitor set stable.

Read the result carefully

Compare the report with a human query check and label any causal explanation as an inference.

SECTION 04

Choose tooling without building a visibility theatre

New reporting creates a new category of software promise. Every best AI visibility tool page will need a clear answer to four questions: which surfaces does it measure, which markets and languages does it cover, how does it sample queries, and can a human reproduce its result? These are not procurement details. They determine whether a number describes a stable comparison or a story assembled from an unknown sample.

Google's own report has declared boundaries that third-party tools may not share. It is currently tied to Merchant Center, English-language queries and five countries, with organic AI traffic only. A tool that claims wider coverage may be helpful, but its figures should not be merged with Google's without a method note. An AI visibility tracking tool can be used for discovery, while Google's report can serve as one platform-specific check. The numbers answer different questions until proven otherwise.

Practitioners are already looking for that distinction. A fresh X post from @aleyda on 13 September 2026 summarised that Google Merchant Center AI Performance now shows AI search intent and terms. The post is a discovery signal, not primary evidence. The help page remains the source to use for the metric definition, traffic scope and country list. That is the right relationship between social discovery and citation.

For a small team, a spreadsheet can be a better AI visibility checker than a costly dashboard if the question is narrow. Record the product, query, market, date, surface, answer, cited source and confidence. Do not turn a single answer into a trend. If the work becomes repetitive, then assess tooling against the audit trail it can preserve. A best AI visibility tool is the one that helps the team make a better product decision, not the one with the loudest score.

Internal links should follow the same evidence logic. folkfox's SEO practitioner survey reading explains why expert belief is not Google's ranking formula. The regional search experience brief keeps eligibility and local context visible. The GEO service analysis shows why a generative engine optimisation claim needs evidence rather than a slogan. Together they make a more honest map than any single score.

@aleyda
Google Merchant Center AI Performance Now Shows AI Search Intent, Terms & Attribution
13 September 2026View on X
SECTION 05

AI search visibility needs a report with boundaries

Google's AI performance insights report is a meaningful move because it gives merchants a named place to inspect conversational discovery. It makes share of voice, frequency, products showing and shopping stages available as an operating conversation. It also gives the organisation a reason to improve product evidence. Those are useful gains for AI search visibility, even before a team has a perfect answer to attribution.

The caveats are equally meaningful. The current help page limits availability to English-language queries in Australia, Canada, India, New Zealand and the United States. It says the report is limited to organic AI traffic and excludes paid Ads traffic. It defines relative metrics against a competitor set and a selected time period. A responsible AI visibility tracking tool or internal report should put those facts on the first page, not in a footnote after a green arrow.

The practical plan is simple. Start with a product family, choose one stage, inspect the attributes and compare a stable period. Use an AI visibility checker to identify where the evidence may be thin, then verify the page and feed. If you buy a best AI visibility tool, demand its sample and reproducibility notes. If you use Google AI Overviews as a phrase in a brief, define whether you mean the surface, the report or a specific query. Shared language prevents expensive confusion.

The wider SEO lesson is familiar but newly visible. Search visibility has always depended on whether a system can find, interpret and trust the information a business publishes. Conversational interfaces make the interpretation step more obvious. That does not turn SEO into prompt theatre. It makes product governance, structured evidence and useful pages more important. AI search visibility is a technical and editorial job with a commercial question at the end.

Use the report as a lantern. Let it show where to look, then walk back to the source of the product claim. If the feed, page, terms and customer experience agree, visibility has a chance to become useful discovery. If they disagree, a stronger score would only hide the problem. The best AI work leaves the evidence legible to a merchant, a marketer and a person deciding whether to buy. The folkfox den keeps a scent trail through the brush and thicket, and the vixen checks each pawprint before the team changes course. AI performance insights are a scoped signal, AI performance insights need a stable period, and AI performance insights should end in a product decision. AI performance insights also need an owner, a date and a comparison rule. AI search optimisation should improve the evidence a customer can check, not merely the appearance of a score. The Google AI features guidance, Search Engine Roundtable and Search Engine Land can extend the research. folkfox's AI consultancy and content marketing service pages keep the delivery route close to the evidence.

AI performance insights should be treated as a measurement rehearsal. AI performance insights need a named market. AI performance insights need a stable period. AI performance insights need a product owner. AI performance insights need a source link. AI performance insights need a caveat. AI performance insights need a decision at the end. That is the difference between AI search optimisation and visibility theatre.

A visibility number is useful when it points back to evidence a customer can check.
folkfox editorial reading
Questions

Frequently asked questions

What is Google's AI performance insights report?

It is a Merchant Center report showing how brands and products are discovered for conversational shopping queries on AI Mode and AI Overviews.

What metrics does the report include?

Google lists share of voice, competitors' average share, frequency and products showing. It also groups queries into Discovery, Evaluation and Ready to buy shopping stages.

Does the report include paid Google Ads traffic?

No. Google's current guidance says the report is limited to organic AI traffic, such as free listings, and does not include paid Ads traffic.

Where is AI performance insights available?

The current help page lists English-language queries for Merchant Center accounts in Australia, Canada, India, New Zealand and the United States.

Is an AI visibility score the same as a ranking?

No. The report is a scoped diagnostic based on queries, products, a competitor set and a time period. It is not a universal ranking score for every AI answer.

What should a merchant improve first?

Start with the product evidence: feed attributes, page copy, structured data, availability, terms and the information a customer needs at each shopping stage.

Keep reading

Read more on this topic

Need AI search visibility that a human can audit?

folkfox helps teams connect search, product evidence and clear measurement boundaries so visibility becomes a useful commercial signal.

Want folkfox in your Google results and AI answers? Set folkfox as a preferred source.

The den

Where to?

Pricing

Choose a section. Enter opens it, Escape continues reading.

Cookie preferences

folkfox uses data the way we use strategy: only when it earns its place.