Google Merchant Center AI Insights: How to Win Conversational Shopping

Google just gave retailers a native window into one of the most consequential — and previously invisible — layers of modern commerce: how your products perform when shoppers talk to AI instead of type keywords. The new AI shopping insights in Merchant Center, announced on May 29, 2026, don’t just add another reporting tab. They signal that managing your product catalog is now an SEO and content discipline, not an operational logistics task. If you run any product commerce on Google — paid, organic, or both — this reshapes how you should think about your entire catalog strategy.

What Happened

On May 29, 2026, Google announced a set of AI performance insight features rolling out to Google Merchant Center across five markets: the U.S., Canada, Australia, India, and New Zealand, according to Martech.org. The rollout is phased across those regions “in the coming months” following the announcement — meaning the tools are live or imminent depending on your market.

The new capabilities are organized into four distinct reporting areas that collectively map how your products surface (or fail to surface) inside AI-powered shopping experiences including Google Search, Gemini, and AI Overviews.

Share of voice insights deliver comparative visibility data against competing brands within AI-powered shopping experiences. For the first time, retailers get a native view of whether they are appearing more or less frequently than direct competitors when a shopper has a buying conversation with Google’s AI. Previously, this kind of competitive intelligence required third-party tools or manual spot-checking of AI-generated search results. Having it inside Merchant Center means it is updated, scaled, and tied directly to your catalog data.

Shopping funnel performance tracks how products move from initial discovery through to purchase across Google’s AI surfaces. This is a longitudinal measure, not just a click-through snapshot. It shows where in the AI-mediated shopping journey users are engaging with your products and where they are dropping off or selecting a competitor’s offering instead. This data is particularly valuable because it isolates the AI surfaces from traditional Shopping ad performance — allowing you to see the AI commerce layer independently for the first time.

Product term insights surface the actual conversational queries shoppers are using when they find (or don’t find) your products through AI. This is fundamentally different from the keyword data you are accustomed to seeing in Search Console or Google Ads. These are extended, natural-language phrases that shoppers use in dialogue with Gemini and AI Overviews — things like “durable backpack for a college student who commutes by bike in the rain” rather than “waterproof backpack.” If your product descriptions do not map to the language patterns showing up in these queries, you are not a candidate for those AI-generated recommendations regardless of your organic ranking or bid level.

Product attribute gap identification flags specific missing structured data fields — color, material, style, size range, compatibility specifications, and category-specific attributes — that are preventing your products from surfacing in conversational search results. The tool does not just tell you that you have gaps; it identifies the specific attribute types causing eligibility problems at the SKU level.

The underlying logic, as Martech.org explains, is that “AI shopping systems need complete, well-organized product data to match products with natural language searches.” The message from Google is operationally direct: if your product feed has gaps, AI will route shoppers to a competitor whose data is complete. The article frames the resulting platform evolution in terms that practitioners need to internalize: retailers now need to treat product feeds “more like SEO content” as Merchant Center becomes an AI commerce optimization platform rather than simply a feed management tool.

This is a significant reframing. Merchant Center has existed in some form since 2010, and for most of that time, getting your feed into the system — with correct pricing, availability, and basic categorization — was table stakes. The new AI insights raise the bar considerably: completeness, richness, and conversational relevance of your product data now determine whether you participate in AI-mediated commerce at all. Getting the product into the feed is no longer sufficient. The product’s data has to be rich enough for Gemini to match it to the natural-language conversations shoppers are having, and complete enough structurally for the AI system to treat it as a qualified candidate for recommendation.

Why This Matters

The weight of this update becomes clear when you understand what the commercial search landscape actually looks like right now. An analysis of 500,000 prompts by Peec AI, reported by Search Engine Journal, found that AI Overviews appeared in 87% of commercial and buying-intent queries. At the decision stage — when a shopper is actively comparing products or ready to purchase — that rate climbs to 88.5%. For longer, more detailed queries running 11 to 15 words, which correlate strongly with high-intent shopping research, AI Overviews appeared approximately 89% of the time.

These numbers look radically different from the 20–42% AI Overview appearance rates cited in most broad industry studies. The discrepancy is methodological: broad studies include all query types, including navigational searches where users are looking for a specific brand or website. When you isolate the queries that actually matter for product commerce — the comparison searches, the “best X for Y” queries, the research-phase conversations — AI Overviews are essentially ubiquitous. The Merchant Center AI insights operate inside this environment. Product discovery is no longer happening primarily through keyword-matched organic listings or traditional Shopping ads. It is happening inside AI-generated responses where Gemini decides which products to surface based on data quality, structural completeness, and relevance to natural language queries.

The implications land differently depending on your organizational context.

For in-house ecommerce teams: Your catalog management function is now a revenue-critical discipline. What was previously treated as a “feed hygiene” task — often delegated to a junior analyst or a feed management vendor on autopilot — now directly determines whether your products appear in the search layer where high-intent shoppers are making purchasing decisions. The head of ecommerce needs to understand this reframe and staff accordingly. A product data quality problem is no longer a back-office logistics issue; it is a direct revenue leak at the top of your acquisition funnel.

For performance marketing agencies managing retail accounts: The product attribute gap identification feature creates an entirely new audit deliverable. You can now pull a structured report showing exactly which missing data fields are preventing a client’s products from surfacing in conversational search results. That is a tangible, sourced diagnosis of a revenue problem — the kind of deliverable that justifies retainer scope expansion and demonstrates expertise that generic agencies cannot easily replicate. The share of voice data creates an equally valuable competitive intelligence deliverable that previously required expensive third-party tooling to approximate.

For solopreneurs and lean DTC brands: The share of voice feature provides a category of competitive intelligence that previously required expensive third-party tools or manual monitoring. You can now see whether you are consistently losing AI search share to a specific competitor and trace that loss to a diagnosable product data issue. For a founder managing 200 SKUs, this is actionable in a way that broad competitive analytics rarely are.

For brand marketing teams: The product term insights data reframes how you should approach product description copy and catalog content strategy. Shoppers are asking AI things like “comfortable running shoes for flat feet under $150 that work for standing all day at a retail job.” If your product descriptions do not include language that maps to those conversational patterns, you are structurally invisible to the most detailed and highest-intent queries in your category. This is a content strategy problem that sits upstream of everything in your paid and organic search program — and it requires collaboration between marketing, content, and catalog teams that most brands have not yet built.

The core assumption this overturns: organic ranking and paid placement are no longer sufficient signals of search health. A product can be ranking number one in organic results and winning impressions in Shopping ads while being completely absent from the AI-generated responses where actual purchase decisions are forming. Retailers operating without native visibility into that AI layer have been flying blind in the most commercially significant part of search. The new Merchant Center AI insights change that.

The Data

The following table maps each of the four new Merchant Center AI insight types to its measurement dimension, the specific gap it exposes, and the operational action it should drive:

AI Overview saturation in commercial search — based on Peec AI’s analysis of 500,000 prompts, as reported by Search Engine Journal:

The gap between the 20–42% figures that circulate in most industry discussions and the 87% reality for commercial queries is the central miscalibration in how most retail marketing teams currently think about AI search exposure. If you have been benchmarking your AI Overview presence against broad, undifferentiated search datasets, you have been operating with a fundamentally misleading baseline. For product commerce specifically, AI Overviews are the dominant result format in the queries that matter most.

AI Mode scale — from Search Engine Journal: Google’s AI Mode has exceeded one billion monthly users, and query volume has been more than doubling every quarter since launch. That trajectory means AI-mediated shopping is not a future consideration — it is the current, primary channel for high-intent product discovery at scale.

Merchant Center rollout geography — from Martech.org:

Real-World Use Cases

Use Case 1: DTC Apparel Brand Running Quarterly Feed Audits

Scenario: A mid-size direct-to-consumer apparel brand sells 2,400 SKUs across five categories — activewear, outerwear, footwear, basics, and accessories. Their Google Shopping campaigns are delivering positive ROAS, but organic traffic has plateaued despite maintaining stable keyword rankings. The disconnect between ranking stability and traffic growth is a classic symptom of AI-mediated traffic capture by competitors with richer product data.

Implementation: After the AI insights rollout, the ecommerce manager pulls the product attribute gap report in Merchant Center. The report surfaces 340 SKUs missing material composition data and 580 SKUs with no size range context in their descriptions. The team prioritizes the top 40 highest-revenue SKUs with gaps and rewrites product descriptions using language pulled directly from the product term insights report — incorporating conversational phrases like “breathable fabric for hot weather gym sessions” and “runs true to size for athletic builds, not fashion fits.” All structured attribute fields — material, care instructions, fit type, and intended activity — are completed in the feed within 48 hours and the feed is resubmitted. The team sets a 30-day review cadence on the product term insights data to catch any new conversational query patterns emerging in the category.

Expected Outcome: Products with completed attribute data become eligible to surface in AI Overviews for high-intent queries where they were previously absent due to data gaps. At 87% AI Overview presence in commercial buying-intent queries per the Peec AI data reported by Search Engine Journal, closing attribute gaps on high-revenue SKUs provides direct access to the most commercially saturated search result format. The expectation is a measurable improvement in AI-surface traffic within 30–60 days of feed resubmission.

Use Case 2: Marketing Agency Building an AI Commerce Audit Service

Scenario: A performance marketing agency managing 12 retail clients has started noticing traffic patterns that do not explain well in Google Analytics — clients are reporting revenue shifts that do not correlate with changes in paid spend or organic rankings. The agency suspects AI-mediated search is the missing variable and wants to productize an AI commerce audit offering before competitors can build a comparable service.

Implementation: The agency designates a senior performance analyst to run monthly Merchant Center AI insight reports across all eligible client accounts. Using the share of voice data, they build a competitive positioning dashboard showing each client’s AI search visibility relative to their top three identified competitors, segmented by product category. The product term insights data becomes the foundation for a “conversational copy audit” — a structured comparison between the queries shoppers are actually using and what the client’s product descriptions actually say. Attribute gaps are quantified and ranked by estimated revenue impact. The agency packages the output as a monthly deliverable with a 90-day rolling action plan for catalog enrichment and copywriting updates.

Expected Outcome: The agency gains a differentiated retainer offering that addresses the revenue problem clients are experiencing — declining traffic attribution despite stable rankings — and connects it to a diagnosable, solvable product data quality problem. The new deliverable creates natural upsell pathways into feed management, catalog copywriting, and ongoing AI visibility reporting, expanding average client retainer value without proportional increases in overhead.

Use Case 3: Specialty Retailer Optimizing for Conversational Query Patterns

Scenario: A specialty home goods retailer sells 800 SKUs. Their paid search campaigns are profitable, but they have no visibility into how their products perform in AI-generated shopping responses. After accessing the product term insights report, they discover that shoppers are finding competitors through queries like “ceramic cookware that’s safe for all stovetops including induction and dishwasher safe” — while the retailer’s product descriptions only include the phrase “induction-compatible” without conversational context or supporting structured attributes.

Implementation: The retailer’s content team uses the product term insights data as a direct brief for a catalog copywriting sprint on all 60 SKUs in the cookware category. Descriptions are rewritten to include natural-language phrasing that mirrors the conversational query patterns shown in the Merchant Center report. Simultaneously, all cookware SKUs are updated to include complete structured compatibility attributes — stovetop type, oven-safe temperature limit, dishwasher safety status, and induction compatibility — in the appropriate Merchant Center attribute fields. The entire sprint takes one week for a two-person team.

Expected Outcome: With complete structured attributes and copy that mirrors how shoppers research cookware through AI, the category SKUs become eligible candidates for AI-generated shopping recommendations in queries where they were previously invisible. At 87% AI Overview saturation in commercial queries, closing this gap moves products from non-participation to active competition in the primary product discovery format.

Use Case 4: Enterprise Retailer Using Share of Voice for Competitive Intelligence

Scenario: A large sporting goods retailer with 15,000 SKUs competes primarily against two dominant online retailers and several specialty category brands. Their existing competitive intelligence program focuses on price monitoring and paid search auction data. The new share of voice feature in Merchant Center provides a strategically distinct angle: which competitor is winning AI-mediated shopping conversations in each product category, and by what margin.

Implementation: The retailer’s search marketing lead builds a monthly share of voice review process segmented by major product category — footwear, apparel, equipment, and accessories. Categories where a specific competitor consistently leads in AI visibility get flagged for immediate diagnosis using the attribute gap identification tool. The team cross-references share of voice deficits with revenue data to prioritize which categories warrant immediate catalog enrichment versus which can wait for the next quarterly cycle. A simple scoring model is created: share of voice gap multiplied by category revenue equals enrichment priority score. The top ten categories by priority score receive dedicated sprint resources each quarter.

Expected Outcome: The retailer builds an ongoing competitive intelligence workflow focused on AI visibility that complements — rather than replaces — their existing paid search and SEO competitive programs. By connecting share of voice losses directly to specific attribute gaps and query language mismatches, the team can justify headcount and budget for catalog management by tying it to competitive positioning metrics rather than framing it as operational maintenance overhead.

Use Case 5: Multi-Brand Distributor Rationalizing Catalog Coverage

Scenario: A distributor managing products from 30 brands in Merchant Center has significant long-tail catalog sprawl — thousands of SKUs with thin descriptions, inconsistent attributes across brand data submissions, and no conversational search strategy. The attribute gap identification feature reveals that 60% of the catalog has at least one critical missing attribute field. Fixing everything at once is not operationally feasible.

Implementation: Rather than attempting to fix 60% of the catalog simultaneously, the distributor uses the shopping funnel performance data to identify which product categories are experiencing the highest funnel abandonment rates in AI-mediated shopping sessions. Categories with high funnel abandonment and high attribute gap rates receive immediate enrichment investment. Categories with high attribute gaps but low AI funnel activity are deprioritized to the next cycle. The distributor also uses the brand-level view to identify which supplier brands provide complete, high-quality product data natively versus which require proactive outreach and structured data templates. Brands with chronically incomplete data get added to the next quarterly business review agenda.

Expected Outcome: A phased catalog enrichment plan grounded in revenue impact rather than uniform maintenance effort. The attribute gap data also provides a concrete, data-backed basis for holding suppliers accountable to product data quality standards — transforming an internal operational problem into a vendor performance conversation with real metrics behind it.

The Bigger Picture

The Merchant Center AI insights announcement is one node in a much larger infrastructure shift that Google has been executing throughout 2025 and into 2026. At Google I/O in May 2026, the company unveiled several features that compound the significance of the Merchant Center update, per Search Engine Journal.

Universal Cart allows consumers to add products across Google surfaces into a single cart and complete checkout without leaving Google’s ecosystem — a fundamental shift in where commerce closes. Agentic booking combines real-time pricing and availability data to enable direct transaction completion through preferred providers inside AI conversations. Information agents monitor products and listings passively in the background, surfacing recommendations without requiring the user to initiate a traditional search session. Each of these features routes product discovery and purchase consideration through Google’s AI layer, and each depends on the quality and completeness of retailer data in Merchant Center to function correctly.

The I/O announcements also surfaced a measurement problem that the Merchant Center AI insights begin — but do not fully — address. As Search Engine Journal reported: “merchants still own the transaction, but not the purchase intent or product discovery.” Retailers now face three structural attribution gaps in AI-mediated commerce: they cannot see the selection criteria Google’s AI uses to include or exclude products in generated responses; they have no native metrics for whether an agent considered and then rejected their product during a shopping session; and they cannot separately attribute agent-initiated transactions from organic traffic in standard analytics.

The new Merchant Center AI insights do not close all three of these gaps, but they represent Google’s first substantive move toward making the AI commerce layer legible to the retailers who depend on it. Share of voice and funnel performance are imperfect proxies for the selection and attribution gaps — but they are the first proxies retailers have had at all, delivered natively inside the platform.

At the advertising level, the direction is equally clear. Google unveiled two new ad formats designed for AI commerce environments at Google Marketing Live, per Search Engine Journal. Conversational Discovery Ads respond to detailed, exploratory prompts within AI Mode — using Gemini to generate tailored creative and surface product features tied to the context of the conversation, rather than relying on keyword match type. Highlighted Answers embed ads directly within AI-generated recommendation lists, placing sponsored products inside the response itself. Both formats are clearly labeled as sponsored. No confirmed launch date has been provided for either format, but their existence signals where Google’s commerce advertising architecture is heading: away from keyword-matched placement, toward context-matched recommendations inside AI conversations.

The through-line across all of these developments is consistent: Google is building an AI-mediated commerce stack in which product data quality, conversational relevance, and structured attribute completeness determine who participates. The Merchant Center AI insights are the diagnostics layer for that stack. This follows a familiar Google pattern — Smart Shopping in 2020, Performance Max in 2022 — where a mandatory platform evolution is wrapped in new visibility tools that reward early adoption. The retailers who engage with the diagnostics and act on them will compound their AI search advantage over time. Those who treat Merchant Center as a logistics tool rather than a content platform will face compounding visibility losses as AI-mediated commerce scales.

What Smart Marketers Should Do Now

1. Audit your top 100 revenue SKUs for attribute completeness before the Merchant Center AI insights roll out in your market.

Do not wait for the new reports to identify your attribute gaps — a proactive audit using your current feed data will surface the same problems weeks or months before the new tools arrive. Pull your Merchant Center feed and check your highest-revenue products systematically: do they have material composition, size and dimension data, color variants, compatibility specifications, intended use context, and any category-specific attributes required by Merchant Center’s product taxonomy? Any gaps you close before the AI insights launch give you a cleaner baseline and a head start on competitors who wait for the report to confirm what a proactive team could have fixed already. The rollout covers U.S., Canada, Australia, India, and New Zealand “in the coming months,” per Martech.org — the window for getting ahead of this is shorter than it looks.

2. Rewrite your highest-revenue product descriptions to match conversational search language, not legacy keyword patterns.

The product term insights feature will eventually tell you exactly which natural-language phrases shoppers are using in AI conversations to find or miss your products. You do not need to wait for that data to start shifting how your catalog copy is written. Look at your product descriptions and ask honestly: do they sound like something a person would say in a conversation with an AI assistant, or do they sound like a keyword list formatted as a sentence? “Lightweight trail running shoe designed for overpronators — breathable mesh upper, cushioned midsole, reinforced toe box for technical terrain” provides far more conversational surface area than “trail running shoes breathable lightweight.” Start rewriting your top-revenue and top-traffic SKUs immediately using the conversational copy principles, and use the Merchant Center data to validate and refine when it becomes available.

3. Build the share of voice report into a formal monthly competitive intelligence workflow the moment it becomes available.

When the share of voice feature lands in your Merchant Center account, it needs to become a standing monthly input to marketing strategy — not a dashboard you check when you remember. Build a formal review cycle: which product categories are you losing AI share in, which competitors are taking that share, and what does the attribute gap data suggest is driving it? The goal is a closed feedback loop: share of voice deficit leads to attribute gap diagnosis, which drives feed enrichment or copy update, which is re-measured at 30 days. This is the same discipline search managers built around impression share in paid search, now applied to the AI visibility layer. Teams that formalize this cycle will compound their AI search presence; teams that treat it as passive reporting will gradually cede the AI commerce layer to more disciplined competitors.

4. Separate your AI commerce optimization workflow from your traditional shopping campaign workflow — they require different inputs, metrics, and team ownership.

Your existing Google Shopping optimization process — bid adjustments, negative keywords, product group segmentation, ROAS targets — optimizes for clicks in traditional Shopping ad placements. The Merchant Center AI insights optimize for a structurally different outcome: appearing in AI-generated shopping responses that may involve no traditional ad auction at all. These require different inputs (product data completeness, conversational copy), different metrics (share of voice, AI funnel performance), and different team ownership. Assigning both to your paid search manager without acknowledging the distinction leads to one predictable outcome: AI optimization work gets deprioritized in favor of more immediately measurable paid search work, and your AI commerce presence quietly deteriorates. Create separate OKRs, reporting cadences, and team ownership for each layer.

5. Train your catalog management and content teams on conversational search now — before the audit data creates unfamiliar, urgent deadlines.

The action items generated by product attribute gap reports and product term insights are going to land on catalog managers, content writers, and creative teams — not on search managers or data analysts. The gap between “here’s the Merchant Center data showing we need to rewrite 400 product descriptions to incorporate conversational query patterns” and “the content team understands what that means and how to do it well” can cost weeks of execution time. Start that education process now. Explain what AI Overviews are, why product data quality drives AI visibility, what conversational query language looks like compared to keyword language, and what a well-written AI-optimized product description looks like compared to a poorly written one. Teams that understand the underlying logic before the data arrives will execute quickly and make good decisions. Teams encountering this framework for the first time under deadline pressure will make expensive mistakes.

What to Watch Next

Product term insights data granularity — The strategic value of this feature depends entirely on how specific and actionable the conversational query data is when it surfaces in Merchant Center. If Google aggregates query data heavily before surfacing it, the insights may be too broad to drive specific SKU-level copy decisions. Watch for reporting from early adopters in U.S. and Australian markets on whether the query data is actionable at the individual product level or only useful at the category level. This will determine how quickly brands can operationalize the feature.

Integration of AI visibility metrics with automated bidding — The logical next evolution after introducing AI share of voice metrics is connecting them to Smart Bidding. If Google introduces the ability to optimize toward AI visibility targets the same way Smart Bidding currently optimizes toward ROAS or CPA, it would fundamentally restructure how retail media budgets are allocated across AI and traditional surfaces. Watch for signals in Q3 and Q4 2026 Merchant Center beta announcements or at the next Google Marketing Live.

Universal Cart merchant data access terms — Google’s Universal Cart allows consumers to check out across Google surfaces without leaving the ecosystem, according to Search Engine Journal. What transaction and attribution data retailers receive from Universal Cart purchases remains undefined. Retailers participating in Universal Cart pilot programs should review data ownership terms carefully and track any policy updates through Q3 2026 that clarify what signals flow back to Merchant Center versus what is retained inside Google’s ecosystem.

AI Mode ad format performance benchmarks — Google’s Conversational Discovery Ads and Highlighted Answers formats have no confirmed launch dates, per Search Engine Journal. When performance benchmarks emerge from early testing — likely Q3 or Q4 2026 if the formats launch on schedule — the click-through, engagement, and conversion data from conversational ad placements will tell retailers a great deal about how much purchase intent is actually being activated inside AI Mode versus traditional Shopping.

EU regulatory shaping of AI commerce tools — AI Overviews operate at a 76% appearance rate in EU markets and have not launched in France at all, per Search Engine Journal. Retailers operating in European markets will encounter the Merchant Center AI insights in a substantially different regulatory and competitive environment. Pressure from the Digital Markets Act and related EU frameworks may constrain feature availability or rollout timelines in those geographies, creating a two-speed market that multi-national retailers will need to account for explicitly.

Bottom Line

Google’s new AI shopping insights in Merchant Center are the first native visibility tools retailers have had for understanding how their products perform inside AI-generated shopping responses — the format now dominating 87% of commercial search queries. The four features — share of voice, shopping funnel performance, product term insights, and attribute gap identification — collectively reframe Merchant Center from a product logistics tool into an AI commerce optimization platform, and they make the previously opaque AI product discovery layer legible for the first time. The urgency behind this is not manufactured: with Google AI Mode exceeding one billion monthly users and query volume doubling every quarter, AI-mediated product discovery is already the primary commercial search layer, not a future state. Retailers who treat this announcement as a concrete call to action on catalog quality, conversational copy, and competitive AI visibility monitoring will build a structural advantage that compounds as the feature set matures and the AI commerce layer deepens. The product feed is the new landing page — and it is time to give it the same strategic investment you bring to your highest-performing paid search creative.