Google Merchant Center Conversational Attributes: Optimize Product Data for AI Mode
What Conversational Attributes are and why standard feed fields fall short
On May 21, 2026, at Google Marketing Live, Google added a new category of product fields to Merchant Center called Conversational Attributes. The name is accurate: these fields are built for conversational queries, not database lookups.
Standard feed fields — title, description, price, GTIN — are organized around product specs. They work well when users type “men’s black puffer jacket size M under $200.” They break down when users ask “I’m camping at minus 5 degrees for three nights and need something light enough to fit in my pack with a budget around $300.” That query has a scenario, temperature constraints, a practical requirement (compressible), and a price ceiling. There is no feed field for any of that.
Google AI Mode now appears in 50 billion monthly searches. In those results, users are not clicking blue links — the AI generates a response and cites products directly. Which products get cited depends heavily on whether the feed contains information the AI can parse and match to a natural language query. Standard title and description optimization gets you into keyword-based rankings. It does not get you into AI Mode responses.
Conversational Attributes fill that gap. Google has not published the complete field specification, but from the Google Marketing Live announcement and early access in Merchant Center, the core fields cover use cases, material feel, size context, suitability by skill or audience, and environmental conditions like temperature range or weather resistance.
The new fields: what they are and how to write them
Here is a breakdown of the main Conversational Attributes fields and how to approach each one:
| Field | What it tells AI | Writing approach |
|---|---|---|
| Use case | Which activities or scenarios the product fits | Specific activities, not category names |
| Material feel | Texture, sensation, wearing experience | Sensory words, not material percentages |
| Suitability | Who it is for — skill level, age, body type | Comparative descriptions work better |
| Conditions | Temperature range, weather, terrain | Specific numbers and ranges |
| Size context | Weight or size relative to familiar objects | Everyday comparisons |
Use case — “Suitable for outdoor activities” is too broad to be useful. “Designed for 3 to 7 day backcountry camping trips; compression sack fits in a side pocket of a 40L pack” gives the AI something to work with when a user asks about camping gear. Be specific about the activity and the physical constraints.
Material feel — A spec sheet that says “90% duck down, 10% feathers” does not help the AI answer “is this sleeping bag one of those crinkly ones?” Rewrite it as: “Outer shell uses a silent matte nylon that does not rustle in a tent; inner lining has a fleece-like softness against the skin.” That language maps directly to how shoppers describe what they want.
Suitability — “Suitable for beginners” is vague. “Appropriate for adults hiking their first multi-day trail; no technical climbing required; total pack weight under 1.2 kg including the jacket” — that is the kind of description AI retrieves when a user asks “is there something good for someone who has never done a long hike before?”
Conditions — Give numbers. “Suitable for winter” is worse than “rated comfort temperature -5°C, lower limit -12°C; designed for late autumn through early spring in temperate climates.” When a user asks for gear that handles minus five degrees, the AI needs a number to match against.
These fields can be edited individually in the Merchant Center feed management interface, or uploaded in bulk via feed file. If you sync through a Shopify or WooCommerce plugin, check whether it supports the new fields — plugin updates supporting Conversational Attributes are expected before Q4 2026.
AI Performance Insights: tracking how your products perform in AI search
Google launched AI Performance Insights alongside the new fields — a new analytics section in Merchant Center that separates AI surface data from traditional channel data.
Find it at: Merchant Center admin, Analytics, AI Performance. You will see two sets of data side by side: AI surfaces (impressions and clicks from AI Mode and Gemini App) and Traditional (standard shopping ads and search results).
Three numbers worth watching:
AI Impression Share — How often your product appears in AI Mode queries where it should be eligible. A product with strong traditional impressions but near-zero AI impressions almost always has missing or low-quality Conversational Attributes.
Query Type Breakdown — What proportion of the queries that triggered impressions were conversational rather than keyword-style. This tells you how much AI search is already affecting a given category. Outdoor gear, home appliances, and baby products currently show higher conversational query rates.
Attribute Coverage Score — Google scores each product’s Conversational Attributes completeness. Products scoring below 60 have significantly lower chances of appearing in AI Mode recommendations. The score updates within a few days of feed changes.
This section is rolling out to accounts globally through 2026. If you do not see the AI Performance tab yet, it should appear by Q3.
Practical approach: sort by AI Impression Share, then find products with strong traditional performance but weak AI performance. Those are your priority items — the audience signal is already there, the data to serve AI results is not.
Which product categories benefit most from Conversational Attributes
Not all products gain equally here. Based on how AI Mode query patterns differ from traditional search, four categories stand out:
Scenario-dependent products — outdoor gear, fitness equipment, baby products, kitchen tools. Buyers describe a context rather than a spec. “A pan that works well for Korean-style cooking” will pull a product with a well-written use case field; “28cm nonstick pan” will not.
Products where sensory perception matters — bedding, apparel, skincare. Thread count and material composition do not answer “is it soft?” Material feel fields do. These categories see higher than average AI query rates.
Products with a skill or experience curve — cameras, instruments, sports equipment. Shoppers frequently ask about beginner vs. advanced suitability, and that question has no home in a standard feed. The suitability field directly addresses it.
Gift and occasion purchases — holiday gifts, anniversary items, event-specific products. AI shopping recommendations for gift queries rely heavily on occasion-fit fields. If your product makes sense as a gift for a specific occasion, say so explicitly in the field.
Lower priority: highly standardized items like specific electronic component models, industrial parts, or generic consumables where users search by model number. AI search adds limited incremental value for those queries.
If you have a large catalog, use the Attribute Coverage Score to filter — sort low-scoring products by revenue and work from the top down.
The full optimization path with Universal Cart
Conversational Attributes are one layer of a broader change Google announced at Google Marketing Live. Universal Cart now works across Google Search, Gemini App, YouTube, and Gmail. A user who adds something to their cart in Gemini App will see it when they open YouTube. A product discovered in AI Mode can be purchased without ever visiting your website.
Here is the full optimization sequence:
Step 1: Foundation data — Make sure your standard Merchant Center fields are accurate and complete: GTIN, variants, pricing, real-time inventory, shipping, and return policy. Conversational Attributes build on top of this layer. Gaps in the foundation will drag down your overall Attribute Coverage Score regardless of how well you write the new fields.
Step 2: Fill in Conversational Attributes — Start with the high-AI-potential products identified in the previous section. For each product, write the use case, material feel, suitability, conditions, and size context fields. Test each field against a real user query: if someone asked “is there something good for X situation?”, does your field answer that directly?
Step 3: Enable Universal Cart integration — In Merchant Center, go to Programs and confirm that Buy on Google and Universal Cart are both active. If you are on Shopify, check the Checkout settings in your Google Sales Channel and verify that cart sync is working.
Step 4: Monitor with AI Performance Insights — Check AI Impression Share every two weeks. Products with newly completed fields typically take 2 to 4 weeks to show data movement. Once Attribute Coverage Score reaches 80 or above, AI impression improvements tend to be more consistent.
Step 5: Test the actual output — Open Google AI Mode and search the way your target buyer would. Check whether your products appear in the generated recommendations, and whether the AI’s description of your product matches what you intended. If the description is off, the field content is probably too vague or not matching the query pattern.
Universal Cart means the same Merchant Center data now affects purchase paths across Google Search, Gemini App, YouTube, and Gmail. A product that appears in AI Mode can be purchased without ever visiting your website.
Related Articles
AI-Generated Product FAQ Pages: Capture Featured Snippets and AI Overviews
Product pages with FAQ content are 61.9 percent more likely to appear in AI Overviews. This guide shows how to use AI to mine real questions from reviews, support tickets, and search data, generate FAQ pages at scale, add FAQPage schema markup, and get picked up by both search engines and AI engines.
AI Returns Management: Automating Ecommerce Reverse Logistics End to End
Online return rates run 25 to 30 percent, and reverse logistics eats more margin than most sellers realize. This guide covers how AI automates the full returns chain from RMA approval to restocking, compares Loop Returns, ReturnGO, AfterShip, and Optoro, and lays out fraud detection and exchange-first strategies.