A Reelgood analysis found that AI assistants like ChatGPT and Claude are inconsistent in accurately identifying streaming availability for popular titles, achieving 43.76% and 50.21% accuracy respectively, compared to Reelgood's 96.89%. The study identified systematic errors such as stale data, bundle confusion, and long-tail service gaps, which threaten user trust as AI is increasingly used for content discovery. These inaccuracies stem from large language models not being designed to track real-time catalog changes.