An attribute query is a search that names the specific, granular properties an answer must satisfy, instead of a broad topic keyword. "Waterproof jacket under a set price", "family-friendly hotel with a pool", and "vegan restaurant with outdoor seating" are attribute queries; "jackets", "hotels", and "restaurants" are keyword queries.
Attribute queries became common as AI answer engines changed how people search. Google reports that the average Google AI Mode query is about triple the length of a classic search and that follow-up queries grew more than 40% per month (mid-2026, platform-reported), so people state full requirements in natural language and refine them turn by turn. The query fan-out then splits one request into attribute-level sub-questions, and the engine assembles its answer from pages that satisfy each attribute.
The practical consequence: a page is only citable for an attribute query if it names the same attribute in extractable form. Facts stated in text, tables, and structured data can be matched and lifted; attributes buried in images or implied by vague copy ("great for families") cannot. This is why exposing concrete attributes, rather than adjectives, is central to getting cited in conversational search.