Are we about to witness an AI-led revolution in how consumers interact with TV platforms and services, and how content is found and presented? We address this question and consider the role for large language models in search and recommendation – and the guardrails needed when using LLMs. Is there an opportunity for TV providers to become the ultimate source of TV-related knowledge? This panel investigates the next steps in content personalisation. We want to know whether the algorithm, rather than content, is the magic ingredient that keeps users on video sharing and social platforms, and whether TV can improve its own ability to understand people and content and match them together. What else can we do to steal attention from the likes of Meta and YouTube?