Google has added a new class of product feed data to Merchant Center called conversational attributes, designed specifically to help AI-driven surfaces such as AI Mode, Google’s conversational search experience, Gemini, and Business Agent respond accurately when shoppers ask natural-language questions about products.
The announcement came as part of Google Marketing Live on May 26, with Barry Schwartz at Search Engine Roundtable noting the accompanying help documentation that Google linked from its official announcement. The six initial attributes are: question_and_answer, document_link, related_product, item_group_title, variant_option, and popularity_rank. All are optional and described as a supplement to the core Merchant Center product data specification.
This is a structural break from how product feed optimization has worked for the past decade. Traditional Merchant Center optimization meant aligning title, description, and product_type fields to keyword queries, then tuning GTIN and price accuracy to win the auction. The goal was to match a token in the user’s search string. Conversational attributes operate on a different premise: they give an LLM enough context to synthesize an accurate answer to a question the user has not yet typed but is likely to ask.
The question_and_answer attribute makes that logic explicit. A retailer can submit pre-formed questions and answers about a product, the kind a customer might pose to a store associate. An AI agent retrieving that product will have structured answers ready, rather than inferring them from a marketing description written for a keyword-dense environment. This parallels how structured data evolved for rich results: early adopters who added Product schema before Google made it standard got price and review stars in blue-link results well ahead of competitors who waited.
The popularity_rank attribute is worth watching separately. Product visibility in AI-generated answers likely reflects some blend of relevance signals and popularity heuristics. Submitting an explicit rank gives Google a stated signal, though Google has not disclosed how heavily it weights merchant-supplied ranks against its own derived signals. The attribution is one-sided: Google’s documentation describes the benefit to merchants, not the mechanism by which the attribute influences ranking.
The related_product attribute has implications for catalog architecture. If an AI agent is assembling a multi-item recommendation (a complete outfit, a camera bundle, compatible accessories), it can follow declared relationships rather than infer them from category proximity. Merchants who build explicit relationship maps in their feeds are giving AI Mode a cleaner retrieval path than those who leave product relationships implicit in the category tree.
The document_link attribute points to an external document about the product, which could be a spec sheet, a user manual, or a comparison guide. This is GEO, generative engine optimization, applied to commerce: the feed becomes a retrieval index, and the merchant controls what documents an LLM is pointed toward when it needs depth on a specific product.
Google has not said how quickly these attributes will influence ranking or whether any of the six will become standard rather than optional. The announcement includes no independent measurement of how conversational attributes affect product visibility in AI Mode. Merchants should treat this as a beta surface, not a confirmed ranking lever, while the behavior of AI Mode commerce solidifies.
The historical parallel to structured data is instructive here. When Product schema became a strong signal for Shopping results, sites that had already instrumented their product detail pages with full schema markup were six to twelve months ahead of those who implemented reactively. The window for building that lead in conversational attributes may be shorter; AI Mode adoption is moving faster than rich results did in 2012.
Ecommerce SEOs should audit which products in their catalog have the highest conversational demand, typically considered-purchase categories where shoppers ask detailed comparison questions, and build question_and_answer entries for those SKUs this week before the attribute gets factored into AI Mode product retrieval at scale.
Reported by Search Engine Roundtable (Barry Schwartz), published May 26, 2026, citing the Google Marketing Live announcement and associated Merchant Center help documentation.
