Methodology
Overview
To understand how AI models like ChatGPT and Google AI Mode respond to product-related queries, we analyzed 2,500 prompts across five major industries:
Business & Professional Services
Consumer Electronics
Digital Technology & Software
Fashion & Apparel
Finance
Prompt design
We created three types of questions to reflect realistic consumer behavior in AI search. We avoided branded queries to keep the results unbiased and made sure prompts were evenly spread across different product subcategories.
01
General product research
Best running shoes
02
Feature-based research
Best waterproof running shoes
03
Specific use cases
Best waterproof running shoes for rugged terrain
How we collected the data
We ran the prompts through ChatGPT and Google AI Mode, both from a U.S. desktop setup, using Semrush Enterprise's AI Optimization. Responses were collected weekly and refreshed regularly to reduce noise from day-to-day AI variability.
What we measured
Mentions
When and how often brands appeared in the AI answers
Sources
Which websites were cited (e.g., brand sites, review platforms)
Mention Position
Whether a brand showed up towards the start or end of a list
How we ranked results
Each chart or table clearly states the metric used (e.g., SOV, mentions, source count). In many cases, we combined ChatGPT and Google results using a weighted model: 80% ChatGPT + 20% Google AI Mode, reflecting expected usage trends.
Managing variability
Because AI responses can change from day to day, we collected data over longer timeframes and averaged the results to ensure consistency.
Brand mentions in AI responses
Cross-model patterns and differences
Sector-specific visibility trends
Co-occurrence patterns amongst top competitors
From this, we also calculated
Share of voice (SOV)
A weighted score showing how dominant a brand is based on frequency, position, and citations
Diversity Scores
How many different brands or sources appeared across the prompt, higher scores = more competitive variety
Download the full AI Visibility Index study
Explore a deeper analysis and get actionable strategies on how to win in AI search.
Powered by Semrush Enterprise AIO
