Google Merchant Center conversational attributes: what retailers need to know

Google says these attributes can help customers discover product information across AI-driven surfaces such as AI Mode while also enhancing traditional search experiences. As of October 2026, Google documents:

• question_and_answer
• document_link
• related_product
• item_group_title
• variant_option
• popularity_rank

They complement the normal Merchant Center specification rather than replacing it.

Google says the question_and_answer attribute is primarily intended for conversational experiences such as AI Mode in Search. That makes it highly relevant to retailers whose shoppers ask questions a conventional product title cannot answer. For example:

• Is this suitable for beginners?
• Will this fit under an airline seat?
• Is the fabric likely to crease in a suitcase?
• Does this part work with model X?
• Can it be used outside in winter?

The answer must be supportable. This is not an excuse to auto-generate marketing claims at scale. Google also says you do not need to duplicate information already supplied in other relevant attributes such as description, product highlight or product detail.

Start with actual customer language. Useful sources include:

• on-site search;
• customer-service questions;
• product reviews;
• returns reasons;
• the prompts shoppers give AI assistants;
• marketplace Q&A;
• sales-team and store-associate feedback.

Then map those questions back to factual product information. The objective is coverage of the decisions shoppers repeatedly need help making; a long Q&A field that answers none of them is wasted.

SO-ME, the consumer shopping platform Sighted powers, sees the same gap from the shopper's side. It receives requests in natural language, and matching them to the product information in the feeds it gets from ASOS, REVOLVE, Vinted, M&S, New Look, Debenhams and EGO is difficult. A question_and_answer attribute filled from facts the retailer can stand behind puts that information where a conversational system can read it.

This is exactly where enrichment has to be more than an AI writing a nicer description. You need to know:

• which questions matter;
• which products can answer them;
• which facts are missing;
• which facts can be verified;
• which Merchant Center field or site field should carry them;
• whether adding them changes AI shopping visibility or recommendation.

Google states that the question_and_answer attribute can be used in all countries. Availability and serving of particular AI shopping features can still vary by country and surface, so keep data-specification availability and feature availability apart.

Google's Universal Commerce Protocol (UCP) supports agentic checkout and related commerce actions on AI Mode and Gemini for participating merchants. Google's Merchant Center help says the checkout UCP enables is available to select merchants in the US, Canada and Australia, and Google has said the UK comes later; it gives no date for the UK.

Do not confuse this with OpenAI's Agentic Commerce Protocol (ACP). They are different standards.

1. Check the Merchant Center account is healthy and the core product attributes complete.
2. Group product variants correctly.
3. Identify the shopper questions that matter from real data: on-site search, customer-service questions, reviews and returns reasons.
4. Validate each answer against source facts.
5. Add the relevant conversational attributes without duplicating the fields you already fill.
6. Update the product pages where the same information also helps shoppers.
7. Track AI Mode, Search Console and Merchant Center visibility as feature access allows; the measurement guide sets out what each report shows.

Knowing which questions matter, which facts answer them and which field should carry them is a product-intelligence problem, and it is the one Sighted is built to solve: enriching retailer catalogues at scale for conversational discovery, the same technology that powers SO-ME, with the approved attributes carried in the Merchant Center feed it builds. The free AI visibility audit shows which facts your product pages leave unstated, which is where the list of questions to answer starts.

Sighted maps real shopper questions to missing Merchant Center and product-page data, then measures whether the change improves recommendation.