# Profound: answer engines tailor brand picks to demographics

Published: 2026-09-10T16:04:51.704Z · Source: Profound research blog (https://www.tryprofound.com/blog)
Source date: 2026-09-10
Entities: chatgpt, gemini, anthropic, profound

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.

## What this answers

**Does ChatGPT give different answers based on who's asking?**

Yes — per Profound's analysis of 71,147 responses from ChatGPT, Gemini, and Claude, personas built from different income, age, gender, and occupation combinations received notably different brand recommendations and citations for the same question.

**Do AI answer engines cite different sources for different incomes?**

On Gemini clothing queries, Profound found high-income personas drew 24 percentage points more brand-owned citations and 17 points fewer earned-media citations than low-income personas.


Canonical: https://anythingengineoptimization.com/item/2026-09-10-profound-answer-engines-tailor-brand-picks-to-demographics/
From Anything Engine Optimization (AEO Wire) — https://anythingengineoptimization.com/ · Standards: https://anythingengineoptimization.com/standards/