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.Where to Listen to This EpisodeAI Visibility vs Personalisation is available across all major podcast platforms. Choose your preferred platform below.Listen to AI Visibility vs Personalisation on TransistorListen to AI Visibility vs Personalisation on Pocket CastsListen to AI Visibility vs Personalisation on Amazon MusicListen to AI Visibility vs Personalisation on AudibleListen to AI Visibility vs Personalisation on CastroListen to AI Visibility vs Personalisation on CastboxListen to AI Visibility vs Personalisation on Podcast AddictListen to AI Visibility vs Personalisation on DeezerListen to AI Visibility vs Personalisation on RephonicListen to AI Visibility vs Personalisation on MetacastListen to AI Visibility vs Personalisation on Player FMListen to AI Visibility vs Personalisation on SpotifyListen to AI Visibility vs Personalisation on Listen NotesListen to AI Visibility vs Personalisation on Podcast GuruListen to AI Visibility vs Personalisation on PodchaserListen to AI Visibility vs Personalisation on PodverseListen to AI Visibility vs Personalisation on Podlink