In November 2024, together with SE Ranking’s research team, we kicked off a 16‑month study to see how AI-generated content performs in organic search. We launched 20 websites in multiple niches and monitored their results over time. But rankings alone weren’t enough. We also wanted to understand how AI systems find, interpret, and reference information. That led us to broaden the initiative into a larger series of experiments focused on AI search and LLM visibility.
For this next stage, we invented a new fictional brand in a real, competitive niche to observe how fast AI systems would recognize it and whether it could be cited alongside—or even ahead of—established industry authorities and government sites. After the first month of data, several clear trends emerged.
How we ran the experiment
We built a fictional brand and distributed content about it across:
- A brand-new website created and registered specifically for this experiment.
- 11 additional domains that were more than a year old, with existing histories and rankings.
Across these properties, we experimented with seven content types:
- In-depth guides
- “Alternatives” listicles
- “Best of” listicles
- Review articles
- Comparison (“vs”) pages
- How-to/tutorial content
- Clickbait-style posts
We began publishing in March 2026 and tracked how five AI systems reacted: ChatGPT, Google’s AI Overviews, Google’s AI Mode, Perplexity, and Gemini. Overall, we monitored 825 prompts spanning various query types and use cases, which produced 15,835 AI responses in the first month alone.
For every prompt, we evaluated three aspects: whether our brand (or any of our sites) appeared in…