Google has added a new set of AI performance reports to Merchant Center. The reports, announced at Google Marketing Live on 20 May and now confirmed as rolling out, give retailers visibility into how their products appear across AI-powered shopping surfaces for the first time. Previously, retailers could see how their products performed in traditional Shopping results. They had no equivalent view for how they appeared in AI Overviews, AI Mode, and conversational shopping queries.
The gap mattered because the number of shoppers using those surfaces is no longer marginal. AI Mode has surpassed 1 billion monthly users, queries are more than doubling every quarter, and AI Overviews are surfacing across 2.5 billion monthly users. Product discovery on Google is increasingly happening through conversational interfaces rather than traditional product listing pages. Until now, retailers had no data on whether their products were appearing in those conversations at all.
The new reports address that directly. Here is what each one does.
The Four New Report Types
Share Of Voice Insights
Compares a retailer’s product visibility against similar brands across AI shopping surfaces. This is the most commercially significant of the four reports. It answers the question that every e-commerce brand has been unable to answer since AI Mode launched: Are we appearing as often as our competitors, and where are we losing ground?
The Share of Voice framing is familiar from search impression share, but applied here to AI-generated recommendation and discovery surfaces rather than traditional paid or organic listings. Where traditional impression share is bid-dependent, AI share of voice is more directly linked to product data quality, which changes the nature of the optimisation lever.
Shopping Funnel Performance
Tracks how products perform from initial discovery through to purchase across the full AI shopping journey. The report shows where products are surfacing in the funnel, whether they are appearing in exploratory discovery responses, in comparison and shortlisting responses, or only at the high-intent bottom of the funnel, and where they drop out.
This is directly relevant to AI Mode’s funnel behaviour. Google’s own research shows that AI Mode queries are, on average, three times longer than traditional searches, which means the search journey is more considered, and a product that appears early in exploratory queries has a longer window to influence the eventual purchase decision than one that only surfaces when the shopper has already narrowed their options.
Product Term Insights
Shows the conversational queries shoppers are using to find products in AI search. The natural language phrases, questions, and comparison requests that AI Mode is processing when it surfaces (or fails to surface) a retailer's products. This is the AI equivalent of the search terms report in Google Ads, applied to product discovery rather than paid click data.
The practical value is feed optimisation. If the conversational queries shoppers are using to find products in a category don’t match the language used in product titles, descriptions, and attributes, there is a data gap that the new Conversational Attributes, also announced at GML, are designed to address. Product term insights is the diagnostic; conversational attributes are the fix.
Product Attribute Gaps
Flags missing structured product details, such as colour, material, style, and other attributes, that AI shopping systems need to match products with natural language queries. Where the product term insights report shows what shoppers are searching for, the attribute gaps report shows what the product record is failing to provide.
This is the most immediately actionable of the four reports for most retailers. A product with incomplete attribute data cannot be matched to queries that include those attributes. If a shopper asks for “navy blue linen trousers under £80” and the product record doesn’t include colour or material, the product is invisible to that query regardless of bid, budget, or campaign structure.
Merchant Center started as a product feed management tool. It is becoming the optimisation surface for AI commerce.
What This Tells You About Merchant Center’s Direction
The four reports are not a standalone feature update.
They are part of a consistent direction Google has been signalling across GML 2026, the Conversational Attributes spec update, and the AI Performance Insights announcement in the EMEA session: Merchant Center is being repositioned from a product data submission tool into an active AI commerce optimisation platform.
The analogy Google used in the EMEA GML session in Dublin was direct: Merchant Center is becoming the central nervous system for how a retailer’s products show up across Google’s AI surfaces. That framing makes the new reports meaningful beyond their immediate utility. They are data inputs to a feedback loop. Retailers can now see where they are invisible in AI search, diagnose why, and fix the underlying product data. Then measure whether the fix worked.
That loop did not exist before. Retailers optimising for AI search visibility were making changes to product data without any direct feedback on whether those changes improved their appearance in conversational results. The new reports close that gap, partially, at least in this initial release.
The Connection To Conversational Attributes
The product term insights and attribute gaps reports are most useful when read alongside the six new Conversational Attributes Google added to the Merchant Center spec immediately after GML, such as question_and_answer, document_link, related_product, item_group_title, variant_option, and popularity_rank. The reports diagnose the gaps. The conversational attributes are how you fill them. Running the attribute gaps report without having reviewed the conversational attributes spec is doing half the work.
What It Does Not Yet Tell You
The reports show performance on AI surfaces. They do not yet explain the weighting behind AI surfacing decisions, which signals matter most, how product data quality is scored, or how a retailer’s AI share of voice is calculated relative to competitors. That methodology is not published.
The share of voice data may be the most consequential of the four reports in the long run, but it also carries the most important caveat: it reflects Google's measurement of visibility, using Google's methodology, reported inside Google’s platform. It is a useful directional signal. It is not independent verification. Retailers should treat it the way they treat impression share in Google Ads, as a useful internal benchmark, not as a neutral third-party measurement.
What To Do Now
The rollout is in the US, Canada, Australia, India, and New Zealand in the coming months. UK availability is not yet confirmed.
For retailers in those markets, the immediate action is to run the attribute gaps report first. It is the most directly actionable output of the four. It tells you specifically what product data is missing, and missing product data is the most common reason products fail to appear in conversational search queries.
Pair it with the product term insights report to understand the query language shoppers are actually using, then use the new Conversational Attributes, particularly question_and_answer, to close the gaps the reports surface.
The Honest Read
These reports give retailers something genuinely new: direct visibility into AI search performance rather than inference from downstream metrics.
That is useful regardless of where you land on how much of your current traffic is already AI-mediated. The direction of travel is clear. AI surfaces are becoming a larger share of product discovery on Google, and the retailers who build product data optimisation habits now will have a structural advantage when that share grows further. The reports make that optimisation possible in a way it wasn't before. That is the straightforward case for using them.
Four new AI performance report types in Merchant Center. Share of voice against similar brands. Funnel performance across the AI shopping journey. Conversational query data. Attribute gap diagnostics. Rolling out in the US, Canada, Australia, India, and New Zealand in the coming months. UK availability not yet confirmed.
The reports are the diagnostic layer. The Conversational Attributes announced at GML are the fix layer. Used together, they are the closest thing to an AI search optimisation workflow for product data that Google has published to date.
Need a fresh perspective? Let’s talk.
At 360 OM, we specialise in helping businesses take their marketing efforts to the next level. Our team stays on top of industry trends, uses data-informed decisions to maximise your ROI, and provides full transparency through comprehensive reports.
.png)
