By Steve Merrill | June 10, 2026
Google just gave Merchant Center merchants two new tools that actually matter. Most are ignoring them. That's a mistake.
Google AI Mode is now the default search environment for all users. Not a test. Not a percentage rollout. The default. Every search starts in AI Mode. Every shopping query runs through Gemini before a shopper ever sees a product listing.
If your feed isn't ready for that, your products aren't showing up.
What Google Just Launched (And Why It Changes Your Feed Strategy)
Two new features launched in Google Merchant Center in June 2026: AI Performance Insights and Conversational Attributes.
They're not cosmetic updates. They're your window into how Gemini is actually reading your product data, and what it's doing with it.
AI Performance Insights shows you exactly which products are appearing in AI Mode responses and which ones are being filtered out. Conversational Attributes shows you the natural-language descriptions Gemini is generating about your products when shoppers ask questions in AI Mode.
Together, they expose what was previously invisible: the gap between your feed and what Google's Shopping Graph needs to surface your products in AI-generated results.
How to Read AI Performance Insights
Open Google Merchant Center. Go to Performance > AI Performance Insights.
You'll see an AI visibility score at the top. Below it: a product-level breakdown of which items are earning impressions in AI Mode and which are flagged as ineligible. The Diagnostics tab has a new filter called "AI Eligibility." That's the one to focus on.
Every product flagged there has a specific reason. Missing GTIN. Weak title. Thin description. Incomplete category attributes. These aren't soft recommendations. They're hard blockers. Products with these flags don't appear in Google's Shopping Graph the same way clean products do.
Fix the top issues first. The volume of blocked products is less important than fixing the ones with the highest impression potential, typically your bestsellers and highest-margin SKUs.
How to Use Conversational Attributes
Go to Products > Conversational Attributes.
This is where Gemini is translating your feed data into natural language for shoppers. When someone asks "What's a good waterproof running jacket under $150?", Gemini is pulling from Conversational Attributes to construct its answer.
If your product descriptions are thin, Gemini's answer about your product will be thin too.
Review each attribute set. Flag descriptions that are vague, technically inaccurate, or missing the kind of detail a shopper would actually ask about, material, sizing, use case, key benefit. You can provide supplemental data directly in Merchant Center to guide how Gemini describes your products.
According to the Google Merchant Center product data specification, description quality is a primary factor in Shopping ad ranking. With AI Mode now the default, that same quality signal directly affects AI-generated recommendation placement.
The 4 Feed Fixes That Move the Needle Most
I've been running feed audits against the AI eligibility criteria Google published for AI Mode. Four issues come up repeatedly.
1. Missing or incorrect GTINs. A missing GTIN is effectively a disqualifier for AI Mode ads. Google uses GTINs to match your products across the Shopping Graph, verify pricing competitiveness, and pull authoritative product data. No GTIN, no match, limited visibility. Check every product in your feed. For Shopify merchants using the Google & YouTube app, missing barcodes in product settings are the most common source of GTIN gaps.
2. Weak product titles. Your title is the first thing Gemini reads. "Blue Jacket" fails. "Men's Waterproof Trail Running Jacket, Lightweight, Packable, Blue, Size S-XXL" works. Include the key modifier that makes the product findable in a conversational query. Think about how a shopper describes the product out loud, not how you'd name a SKU.
3. Thin product descriptions. Descriptions under 150 words with no use-case context consistently underperform in AI Mode placement. Gemini needs material, audience, key benefit, and at least one differentiating detail. Write descriptions the way a knowledgeable sales associate would describe the product to a customer. That's exactly the input Gemini is working from.
4. Missing category-specific attributes. Google's product taxonomy has category-level required attributes that go beyond the basics. Apparel needs color, size, gender, and age group. Home goods need material and item dimensions. Electronics need brand, model, and compatibility. Check your feed against the required and recommended attributes for your product category in the Merchant Center Help Center. Gaps in these attributes lower your AI eligibility score directly.
What's Coming: AI Mode Ads Are Getting More Aggressive
At Google Marketing Live 2026, Google announced four new Gemini-powered ad formats that appear directly inside AI Mode responses. These formats, including Google Direct Offers, place sponsored product recommendations inside the conversational result itself. Not below it. Inside it.
Google AI Max is also in broad testing for search campaigns, applying AI to match, bidding, and creative in real time across AI Mode inventory.
The brands that will win those placements are the ones with clean, complete, AI-eligible feeds right now. Feed quality determines eligibility. Eligibility determines whether your products appear in Direct Offers. The infrastructure work you do today is what puts you in those slots when they fully roll out.
The Audit Checklist
Before you touch anything else, run through these five checks in Merchant Center:
- Open AI Performance Insights and note your AI visibility score
- Filter Diagnostics by AI Eligibility, count how many products are blocked
- Review Conversational Attributes for your top 20 products
- Check GTIN coverage across your full catalog
- Pull the category-specific required attributes for your primary product type and audit for gaps
Do this once a week until your eligibility rate is above 90%. It moves faster than you'd expect once you systematize it.
FAQ
What is Google Merchant Center AI Performance Insights?
AI Performance Insights is a reporting feature inside Google Merchant Center that shows how your products are performing specifically within Google AI Mode search results. It tracks AI visibility scores, identifies which products are being surfaced in AI-generated shopping responses, and flags feed issues blocking AI Mode eligibility.
What are Conversational Attributes in Google Merchant Center?
Conversational Attributes is a feature that lets Gemini describe your products to shoppers in natural language during AI Mode conversations. Google generates these descriptions from your feed data. Thin or inaccurate product data produces poor descriptions that hurt conversion and placement.
Does Google AI Mode affect Shopify stores specifically?
Yes. Any Shopify store running Google Shopping ads or syncing products via the Google & YouTube app is affected. Google AI Mode is the default search experience for all users, meaning AI-generated results are what shoppers see first. Feed quality determines whether your products appear in those results.
What feed issues most commonly block AI Mode placement?
The most common blockers are: missing or incorrect GTINs, weak product titles that lack key descriptors, thin descriptions without material or use-case context, and missing category-specific attributes like color, size, and material. Google's Merchant Center Diagnostics tab surfaces these under "AI Eligibility" issues.
What are Google Direct Offers in AI Mode?
Google Direct has are sponsored product recommendations that appear inside AI Mode's conversational search results. They surface when a shopper's query matches a product in your catalog. Feed quality and AI eligibility are the primary factors determining whether your products earn those placements.
Is Your Store Ready for AI Shopping?
Feed fixes are step one. But there's more to AI visibility than a clean Merchant Center account. Most Shopify stores have gaps across structured data, product content, and site architecture that filter them out of AI-generated results entirely.
We built an audit process specifically for this. It maps exactly where your store stands against what Google AI Mode, ChatGPT Shopping, and Perplexity need to recommend your products.
