AI Visibility vs Personalisation: What SEOs Must Know

James Dooley speaks with Benjamin Tanenbaum about personalization in large language models such as ChatGPT and Gemini. Benjamin explains that different answers appear because token variability and query fan out create distribution shifts, while logged in memory and location data further refine results. He clarifies that true personalization is often limited on free plans, but location and prior context still influence outputs because queries are expanded with user signals. They discuss share of voice tracking in AI visibility, arguing that optimization increases citation probability even with volatility. The episode matters because AI search is shifting rankings toward probabilistic exposure rather than fixed positions.