Rocksalt Research
What AI engines actually cite, and why.
We run experiments on how ChatGPT, Perplexity, Gemini and other AI engines answer real B2B buyer questions: who gets named, which sources get cited, and what changes the answer. Every write-up shares the method and the data.
- The question
- The method
- The data
- What it means for you
AEO experiments
Latest experiments
Experiment 01 · Coming soon
Our first AEO experiment publishes here soon.
Each write-up will cover the buyer question we tested, how we ran it across AI engines, the raw data, and what it means for B2B marketing teams.
Archive
Earlier research: LinkedIn & Reddit
Our earlier studies of how B2B marketers and buyers talk on LinkedIn and Reddit, two of the communities AI engines cite most.
-
753k Reactions analyzed
Where B2B marketing attention went in 2025
A year of engagement data across 6,956 posts: which narratives B2B marketers actually engaged with, and which were noise.
Read the report -
6,753 LinkedIn posts
How to get more engagement on LinkedIn in 2026
An analysis of posts from 31 B2B marketing influencers: the formats, topics and habits that drive engagement.
Read the report
Put the research to work
See where your brand stands in AI answers.
Rocksalt runs this kind of analysis on the real questions your buyers ask, continuously, and tells you what to publish next.
Book a demo