Driving AI product discovery with Google conversational attributes

Google says AI Mode has surpassed 1 billion monthly users. That scale means AI-driven search is now a major product discovery channel in Google conversational commerce.

But discovery in AI Mode works differently. People search with longer, more detailed queries instead of simple keywords. Standard feeds, built for keyword matching, are not suitable for this evolved form of search. That's why Google introduced conversational attributes in the Merchant Center. These fields help AI systems understand product context that standard feed data cannot provide.

In this article, we explain what Google conversational attributes are and why they matter now. We also look at how retailers can add rich product context at scale without creating manual work.

Key takeaways

• Google's new attributes help customers discover information through detailed, natural-language searches, supporting the shift toward conversational commerce.

• The six fields include item group title, variant option, and popularity rank. Together, they capture product-specific nuances that standard feed data may miss.

• These attributes support agentic commerce while enhancing traditional search experiences. They give Google more context about product relationships, variations, and demand.

• Retailers can add the fields to an existing feed. This does not affect product approvals or disrupt current Google Shopping campaigns.

• Channable gives you a smarter way to generate, review, and publish product content at scale, supporting easy discovery across AI-powered shopping platforms.

What Google conversational attributes are and why they exist

Google conversational attributes are optional product data fields in Merchant Center that describe how a product is used, who it is for, and common questions about it. They help Google's AI experiences, like AI Mode and Gemini, match products to natural-language queries.

Shoppers are moving away from rigid keyword searches. Instead, they are turning to natural, conversational searches in Gemini and AI Mode in Search.

— Amy Bateson, Senior Product Marketing Manager at Channable

For example, you can add a short Q&A entry such as "Is this stroller suitable for newborns?", product highlights like "designed for wide feet" or "best for daily 5k runs on concrete," and links to manuals or sizing guides.

These fields give Google's AI the extra context that standard Shopping attributes can't provide on their own, so your products can be matched to detailed, full-sentence questions.

The 6 Google conversational attributes explained

As Google leans further into conversational and agentic shopping, it needs product data that reflects how people talk and ask questions about what they want to buy.

Standard titles and descriptions are not enough for that.

Google's new conversational attributes allow you to convey deep, descriptive information about your products and variants.

— Amy Bateson, Senior Product Marketing Manager at Channable

To close this gap, Google has added six conversational attributes in Merchant Center that map to the main ways shoppers describe and compare products.

These six attributes cover:

1. Questions
2. Deeper documentation
3. Related items
4. Group variants
5. Different versions of the same product
6. Popularity

Let's break down each attribute in detail, with examples.

1. question_and_answer: Answer what shoppers actually ask

Google's question_and_answer attribute lets you pre-load FAQs directly into your product feed in the exact format AI needs for conversational shopping.

You add plain-language question-answer pairs like:

Q1: Does this have a headphone jack?

A1: This version doesn't have a headphone jack.

This way Google's AI can respond instantly when shoppers ask the same thing in AI Mode or Gemini.

Under Google's conversational attributes specifications, you can submit up to 30 Q&A pairs per product, with up to 1,000 characters for each question and each answer, which is usually enough to cover sizing, compatibility, care, and basic "is this right for me?" concerns.

Google recommends basing these on what people ask in reviews, support tickets, and on-site FAQs. It explicitly warns against duplicating information that already lives in description, product_highlight, or product_detail fields.

2. document_link: Give AI a richer product source to read

The document_link attribute lets you add links to product PDFs, such as user manuals, assembly guides, spec sheets, or sizing charts. It sits inside your Google Merchant Center attributes as part of the new conversational attributes group.

This field is primarily intended for AI-driven experiences like AI Mode in search. Each URL must point to a publicly accessible PDF that Google can crawl, and that contains product-specific information shoppers use to judge fit, safety, or compatibility.

3. related_product: Help AI make smart connections

The related_product attribute tells Google exactly how one product in your catalog relates to another. You use it to mark accessories, required parts, substitutes, and often bought with items. That way, conversational agents on AI Mode can support complete conversational shopping journeys.

Under the hood, this conversational attributes Merchant Center field has three specific sub-attributes:

• Relationship_type
• Identifier_type
• Identifier

You submit all three for each related item, and you can repeat related_product multiple times if a product has several accessories or substitutes.

related_product is one of the most useful Google Merchant Center attributes for conversational commerce optimization. It teaches Google's AI which products belong together in a bundle, which add-ons are required, and which alternatives make sense to suggest.

4. item_group_title: Group variants so AI can navigate them

item_group_title gives the whole variant family a single, shared product name in your conversational attributes setup.

Google recommends using it together with item_group_id, so AI-driven surfaces like AI Mode can first understand "this is one product with variants" and then use the variant titles and variant_option values to drill into the exact choice.

item_group_title is the family name (example: "Organic cotton men's T-shirt") and the individual title is the specific SKU (example: "Organic cotton men's T-shirt, navy, size L").

That structure makes conversational agents' jobs easier. They can present the group once, keep the Google Merchant Center attributes for each variant clean, and help shoppers navigate color, size, or spec differences.

5. variant_option: Describe the specific choice clearly

variant_option describes what makes one variant different from another in a format that works for both conversational AI and traditional search experiences. It sits alongside item_group_title and item_group_id as one of the newer data attributes in Merchant Center for handling complex item groups in a more conversational way.

You use variant_option with item_group_title and item_group_id to spell out the specific attribute name and value for each variant, like "memory: 16 GB" vs. "memory: 32 GB," or "blade type: fine" vs. "blade type: coarse." That extra structure helps AI tools such as Gemini understand which single variant the shopper is asking about.

Google treats variant_option as optional, but only for products that actually come in variants. If a product belongs to a variant family, you use variant_option to define the property that identifies that exact choice, using the required sub-attributes name and value.

6. popularity_rank: Signal what's trending

popularity_rank tells Google how popular a product is within your own catalog. You assign a number between 0 and 100 (with up to one decimal place, no percent sign) that reflects how well this product is selling compared with your other SKUs, where a higher value means stronger performance.

Google calls this a completely optional conversational attribute. Google also suggests using it to highlight your latest bestsellers and give shoppers a clearer signal of what is trending.

When you update popularity_rank regularly from your sales data, you give AI-driven experiences a simple, trustworthy way to prioritize high-demand products. You can do this without changing bids, prices, or product descriptions.

Why populating these attributes manually doesn't scale

Populating six conversational attributes across a full catalog means creating and maintaining new fields for Q&A, documentation, relationships, variants, and popularity for every SKU. At a few thousand SKUs, doing that manually leads to inconsistent data.

The math problem: 6 fields x thousands of SKUs

You now have six new fields that Google expects you to maintain for every eligible SKU. Even at one minute per field, that is six minutes per product. On a 10,000-SKU catalog, you are already at 60,000 minutes of work, roughly 1,000 hours, before you touch updates, new launches, or seasonal changes.

Most mid-to-large retailers run much larger catalogs. Many also sell into multiple markets. That means tens or hundreds of thousands of potential entries to keep in sync with real product detail, stock, and performance data.

What happens when merchants skip Google conversational attributes

Skipping Google conversational attributes doesn't break your Merchant Center setup, but it limits how well you can compete in AI shopping:

• You lose conversational visibility. AI models pull directly from fields like question_and_answer and related_product when answering natural-language queries. If those fields are empty, your products are less likely to be recommended for intent-led questions.

• You get no warning that anything is wrong. These attributes are completely optional in the Merchant Center. Your products stay approved, your Google Ads campaigns keep running, and there are no error flags. You just miss out on AI-driven discovery without any explicit signal.

• You give up control of the AI narrative. Without document_link and pre-written Q&A, the AI engine falls back to whatever it can crawl from your site and other sources. That means you have less control over how a virtual agent summarizes your product or compares you to alternatives.

• You miss AI performance insight. Merchant Center's AI-focused reporting features can only work with the data you provide. If conversational attributes are blank, you lose the chance to see how AI-driven shopping is discovering and using your catalog.

How to set up Google conversational attributes with Channable

With a product feed management tool like Channable, conversational attributes become part of your standard Google Merchant Center workflow.

Here's how to use conversational attributes to configure all the fields, generate content efficiently, and choose the right feed setup for your catalog using Channable.

Channable's Google feed template already supports all 6 attributes

In Channable, conversational attributes live in the same place as the rest of your Google Shopping data: the Mapping step of your Google feed.

Once you've created or opened a Google Shopping feed under Channels, go to Feeds, open your Google Shopping feed, click on Mapping, and you will see Google's channel fields ready to be matched to your internal project fields.

To use conversational attributes, you simply create or reuse internal fields in your project (for example "qa_materials" or "manual_url") and then select those fields in the internal field dropdown next to the relevant Google attribute in the Mapping tab.

Channable will send those values through to Merchant Center with every feed run.

Use AI Text Generation to populate them at scale

Once your conversational attributes are mapped in the Google feed, you can use AI Text Generation to create conversational attributes, including question_and_answer content, richer variant details, or other supporting text fields in bulk, based on the product data you already import.

You connect your Gemini or OpenAI account, choose which internal fields the AI should read from, and pick a separate field where the generated text should be saved so you never overwrite your original data. Then you filter the items you want to work on, refine your prompt with a handful of previews, and generate texts for the full selection once you are happy with the output.

You don't have to write a single line of data manually. You can use our built-in AI text generation solution to automatically create optimized conversational data for your entire catalog.

— Amy Bateson, Senior Product Marketing Manager at Channable

Every generated value can be reviewed, edited, or regenerated before you approve it, and only approved values will be used in your feed mapping. That way, you can scale conversational attributes across your catalog while controlling token costs, quality, and tone, turning AI into a structured part of your feed workflow.

Primary feed or supplemental feed: which to use

Google lets you send conversational attributes through either your primary feed or a supplemental data source. And it notes "supplemental" as the recommended option for most setups.

These attributes are optional, so adding them will not change the approval status of your existing products or disrupt your current campaigns.

In Channable, the workflow is the same in both cases. You generate and approve conversational content in Optimize, store it in dedicated project fields, and then map those fields into whichever feed structure you choose.

If your priority is simplicity and a single source of truth, we recommend that you add conversational attributes to your primary Google Shopping feed so everything lives in one managed flow.

If you want a low-risk testing path or need separation from your core product data, use a supplemental feed instead and layer the conversational fields on top of your existing primary feed.

Use Channable to surface your products in AI-powered search

As Google moves further into conversational and agentic shopping, retailers need product data that can answer detailed questions. Conversational attributes give Google the product-specific context it needs to match detailed questions and improve your AI visibility across AI-powered shopping experiences such as Gemini and AI Mode.

Channable helps your brand adapt to AI-powered shopping without rebuilding or disrupting your existing Google Shopping setup.

With Channable, you can add all six conversational attributes directly to your Google feed setup, using either a primary or supplemental feed. Channable's AI Text Generation then uses your existing product data to create the required conversational content at scale, which you can review before publishing.

Channable optimizes your catalog for discovery in Gemini and AI Mode without manually updating thousands of products or disrupting existing approvals and campaigns.

Book a demo