TL;DR
- Google Merchant Center's AI Performance Insights report launched in July 2026 and began expanding to more regions earlier in September 2026.
- Brodie Clark, who spotted the three new sections and posted a screenshot on X, said the report now covers AI Search intent, AI Search terms and AI attributes.
- AI Search intent shows how products match customer AI searches; AI Search terms recommends popular terms for product descriptions; AI attributes recommends missing product attributes.
- Google also posted numerous changes to the help documentation for the report, and Search Engine Roundtable reported the expansion on 9 September 2026.
- The source did not say which regions have it, whether Thailand is included, what share of accounts see it, or any performance figures.
Google has added three sections to the Merchant Center AI Performance Insights report: AI Search intent, AI Search terms and AI attributes. Search Engine Roundtable reported the expansion on 9 September 2026, crediting Brodie Clark, who posted a screenshot of the updated report on X.
What changed in the report
Barry Schwartz wrote at Search Engine Roundtable on 9 September 2026 that the AI Performance Insights report inside Google Merchant Center has gained new metrics, and published a side-by-side comparison of the older version of the report against the one Brodie Clark captured. Clark's own description of the change was short: "The report has now expanded to include sections for the AI performance reporting related to AI Search intent, AI Search terms and AI attributes."
Google also posted numerous changes to the help documentation covering the report, which Search Engine Roundtable noted alongside the screenshot. A documentation update is the part worth registering here, because it is the closest thing to a first-party signal in the story. The screenshot shows one account. The help document describes what the feature is meant to do for every account that has it.
The timeline behind this report
AI Performance Insights is not new in itself. Search Engine Roundtable reported that the report launched in July 2026 and began expanding to more regions earlier in September 2026, and the three sections spotted on 9 September 2026 are an addition on top of that, not the arrival of AI reporting in Merchant Center. Anyone describing this as Google's first attempt at telling retailers how their products perform in AI surfaces would be skipping the July launch entirely.
The sequence matters for a practical reason. A report that launched two months ago and is still expanding in both regions and sections is a report still under construction. Data seen inside it now is early data, and the shape of the interface a merchant sees today is not a guarantee of what the interface will look like next month.
What each new section reports
AI Search intent
AI Search intent lets a merchant see how their products match customer AI searches, in Brodie Clark's description of the section. That is a matching view rather than a ranking view: it addresses whether the catalogue lines up with the sort of thing shoppers are asking AI surfaces, which is a question most merchants have never had a Google-side answer to.
AI Search terms
The AI Search terms report recommends improving visibility by including popular terms in the product descriptions of products that show up in AI searches, again per Clark's summary. The recommendation is aimed at descriptions specifically, and it applies to products already appearing in AI searches rather than to the whole catalogue.
AI attributes
AI attributes recommends increasing visibility by adding missing attributes to products that show up in AI searches. This is the section that maps most directly onto work a feed manager already understands, because a missing attribute is a concrete, fixable gap in a product record rather than a judgement call about wording.
The table below sets the three sections against what a merchant can actually do with each one. The middle column is Clark's description as reported; the right column is the practical reading of it.
| Section | What it reports, per Brodie Clark | What a merchant can do with it |
|---|---|---|
| AI Search intent | How your products match customer AI searches | Judge whether the catalogue covers the demand, and spot ranges that match nothing |
| AI Search terms | Recommends including popular terms in the descriptions of products showing up in AI searches | Rewrite descriptions for the products that already surface, starting with the highest-value ones |
| AI attributes | Recommends adding missing attributes to products showing up in AI searches | Fill the named gaps in the feed, which is the cheapest and most mechanical of the three tasks |
Why attributes keep turning up as the lever
All three sections point at the product record rather than at the website, and that is the underlying mechanism worth understanding. A product feed is a structured description of an item, and an AI surface answering a shopping question needs the specifics: size, material, compatibility, colour, condition, whether it fits the thing the shopper already owns. A retailer who leaves those fields blank and describes them in a paragraph of marketing prose has given a human a readable page and given a machine nothing to match on.
That explains why a section called AI attributes exists at all. Google is not asking for more content in the general sense. It is asking for the specific fields that let a product be matched against a specific question, which is a much narrower request and a much easier one to complete.
How the AI Search terms recommendation differs from keyword stuffing
The AI Search terms section recommends adding popular terms to product descriptions, and that phrasing will read to some teams as an invitation to pack keywords in. Treating it that way would be a misreading of the mechanic. A description is matched against what a shopper asked, so the value comes from the description containing the words that name the product accurately, in the vocabulary buyers use rather than the vocabulary the merchandising team uses internally.
The honest test is whether a term added to a description is still true of the product. A term that describes the item and happens to be popular is a fix. A term stapled on because the report listed it is a description that misrepresents the product, and a shopper who arrives on that basis leaves again. The report recommends terms; it does not remove the merchant's obligation to describe the product correctly.
The limits of a platform reporting on its own AI visibility
A report like this is useful and partial at the same time, and the distinction is worth holding onto. Google can report on what Google's own AI surfaces do with a catalogue. It cannot report on what any other assistant does with it, so a merchant reading strong numbers here still has no picture of visibility inside ChatGPT, Perplexity or any other answer engine. The AI-visibility picture stays fragmented across as many dashboards as there are platforms.
There is also the ordinary caution that applies to any recommendation surfaced by the platform selling the placement. Recommendations describe what the system would like the data to look like. That is genuinely useful information, and it is not the same thing as a measured result. For the same reason, the sensible way to use the three sections is as a work queue for feed quality rather than as a scoreboard, which is also how the rest of an ecommerce marketing reporting stack tends to earn its keep.
What Search Engine Roundtable did not report
Several obvious questions have no answer in this story, and inventing one would be worse than leaving the gap visible.
- Which regions the report has expanded to. The reporting says expansion to more regions began earlier in September 2026 and names none of them.
- Whether Thailand is included. Nothing in the report addresses Thailand either way.
- What share of Merchant Center accounts can see the three new sections. The story rests on one screenshot from one account plus the help documentation.
- Any completion date for the rollout.
- Any performance figures. No numbers from the report itself were published, so there is no benchmark to compare an account against.
What this means for Thai marketers
The source did not say whether this is live in Thailand, so the first practical step for a Thai retailer is to open the AI Performance Insights report and see whether the three sections are present in the account. That takes a minute and answers the question for that account, which is more reliable than any general claim about market availability.
The wider relevance is not conditional on the rollout. Feed quality is already the input that decides how a product is represented across Shopping surfaces, and a Thai retailer running Shopping campaigns through Google Ads is feeding the same product data into all of them. Attributes filled in now pay off whether or not the AI attributes section has appeared yet, and the same catalogue discipline is what any generative engine optimisation work for ecommerce has to start from.
One local wrinkle is worth naming as analysis rather than as reporting: Thai retailers often maintain product data in Thai, English or both, and the vocabulary a shopper uses in each language is different. A recommendation engine surfacing popular terms will surface them for the language and market it is measuring, so a merchant selling into more than one market should read the recommendations per feed rather than applying one set of terms across the catalogue.
What to check in your Merchant Center account
None of this requires a strategy meeting. It is feed work with a report now pointing at where the gaps are.
- Open the AI Performance Insights report and confirm whether AI Search intent, AI Search terms and AI attributes are present for the account.
- Start with the AI attributes recommendations, since a missing field is unambiguous and filling it is a mechanical change.
- Compare the AI Search terms suggestions against the words the product descriptions currently use, and treat any term that is not true of the product as a rejection rather than a task.
- Look at the AI Search intent view for ranges that match nothing, which is a merchandising signal as much as a data one.
- Check the help documentation, which Google updated alongside the change, before drawing conclusions from a screenshot circulating on social platforms.
- Record the state of the account today so that a change next month is visible as a change, given the report is still expanding.
Frequently asked questions
Is the expanded report available in Thailand?
The source did not say. Search Engine Roundtable reported on 9 September 2026 that the report began expanding to more regions earlier that month without naming any region, and made no statement about Thailand.
Do I have to do anything to turn these sections on?
Nothing in the reporting describes an opt-in, a setting or an application process for the three sections. The report was described as having expanded, so the practical answer is to open AI Performance Insights and see what is there.
Is this the first AI reporting in Merchant Center?
No. Search Engine Roundtable reported that AI Performance Insights launched in July 2026, so the three sections spotted on 9 September 2026 are an expansion of an existing report rather than the first AI reporting Merchant Center has offered retailers.
Will adding the recommended search terms guarantee more AI visibility?
Nothing in the reporting promises a result. The AI Search terms section was described as recommending popular terms for product descriptions to improve visibility, which is a recommendation and not a stated outcome, and no performance figures from the report were published.
What numbers does the report actually show?
Search Engine Roundtable published no figures from the report. The story documents which sections exist and what each is for, based on a screenshot and Google's updated help documentation, so any specific metric attributed to it would not be coming from this source.
Where this leaves retailers
The change is small in itself and points at something that is not small. Google is now telling merchants, inside their own Merchant Center account, which terms and which product attributes are missing for items appearing in AI searches. That is the platform describing the shape of the data it wants, which is the most direct guidance a feed owner ever gets, even when it arrives without a single published number attached.
Search Engine Roundtable's report on the change, published on 9 September 2026, is available at Search Engine Roundtable. If your Merchant Center account has the sections and nobody has looked at the recommendations yet, that backlog is the fastest feed work available right now, and Relevant Audience takes on that kind of catalogue review.
