Profound: answer engines tailor brand picks to demographics
Profound tested ChatGPT, Gemini, and Claude with personas built from combinations of income, age, gender, and occupation, asking each about credit cards, clothing, and furniture across 71,147 responses collected August 11-24. Per the company, only about 1 in 4 recommended brands matched between different personas, versus roughly 2 in 5 when the same persona repeated the question, and brand mentions rose from 5.8 for low-income personas to 7.5 for high-income ones. On Gemini clothing queries, high-income personas drew 24 percentage points more brand-owned citations and 17 points fewer earned-media citations than low-income personas, Profound said.
Why it matters: It suggests AEO tracking needs to segment by audience persona as well as by platform, since the same answer engine can favor different brands, price points, and source types depending on who appears to be asking.
The record: ChatGPTGeminiAnthropicProfound
Glossary: Answer engine
Posted to the wire September 10, 2026. Edited by Joe Balewski.