Beware blanket advice on AI search visibility. No two large language models (LLMs) behave the same, neither do two versions of the same LLM. The same probabilistic behavior that makes models so easy to talk to makes their responses highly sensitive to the subject matter, sentiment and syntax of users’ prompts.
Take earned media, for example. These media, which range from news publications to customer reviews, can be some of the hardest for marketers to affect. A marketer could much more readily post on their owned domains (corporate media), pay an affiliate (affiliate media), work with influencers (social media) or collaborate with a retailer (ecommerce media) than get a reporter to take an interest in their brand or impress a meaningful number of reviewers.
It’s enough to make a marketer question how much earned media matters for AI visibility and how much effort it is worth. On the surface, the answer is earned media domains account for about 32% of all domains AI models cite, a meaningful portion, but not a dominant share. Below the surface, however, we found the story is not that simple.
Digging deeper, our research found that earned media’s share of AI-cited domains vary drastically depending on subject matter. Medical topics, for example, averaged an earned media share of 58% on ChatGPT, with certain topics tallying an earned media share as high as 87%. Topics pertaining to supplements, meanwhile, averaged just 24% and reached as low as 7%.
No One-Size-Fits-All
Many marketers use our AI marketing platform to understand how frequently and in what average position their brand shows up in LLM responses, looking specifically at prompts related to the top consumer preferences in their category and how their brand compares to their competitors. These kinds of prompts often look like:
- “Give me a list of [product name] that have the best [consumer preference].”
- “What are the best [industry] brands for [consumer preference/product name]?”
- “Name [industry] brands that stand out for [consumer preference].”
We asked ChatGPT and Google AI Mode more than 108,000 unique prompts that used this structure, repeating each for statistical validity, across a wide range of industries, consumer preferences and product categories and large disparity in earned media shares.
For example, prompts associated with medical industry verticals had a significantly higher average earned media share, 58%, than the dataset as a whole, 32%. Supplements, meanwhile, had a much lower share on ChatGPT, only 24%, but supplements maintained a higher-than-average earned media share on Google AI Mode, 47%.
We found similar disparities - both topic-to-topic and model-to-model - when examining other categories like electric vehicles, vehicle insurance and property insurance.
This exploration clearly shows that knowing “earned media accounts for about 32% of all domains AI models cite” is not enough information to inform a marketer’s decision. It significantly understates earned media’s importance for a marketer of a pharmaceutical brand, for example.
What a marketer needs to know to make informed decisions on things like earned media campaigns are the AI search particularities of their brand and product category - the prompt volumes, consumer preferences and most-cited domains in their segment.
Methodology
At Evertune Research, we track hundreds of brands across 250 categories. For this analysis, we repeatedly sampled more than 108,000 unique product and category-related prompts on ChatGPT and Google AI Mode, running prompts throughout the first quarter of 2026.
Evertune is the AI marketing platform for Generative Engine Optimization (GEO) that helps brands improve visibility in AI search by analyzing responses at scale and delivering actionable insights. Evertune works with leading brands across all verticals, including Finance, Retail and E-Commerce, Automotive, Pharma, Tech, Travel, Food and Beverage, Entertainment, CPG, and B2B. Founded by early team members of The Trade Desk, Evertune has raised $20M in funding from leading adtech and martech investors. Headquartered in New York City, the company has a growing team of more than 40 employees.
Don't forget to share this post
