The video industry has spent the better part of a decade trying to solve the discovery problem with ever more elaborate rows of content tiles. One of the latest fixes being floated involves replacing the grid entirely—with a conversation. Executives from Gracenote and Nordic telecom operator Telia, speaking at Connected TV World Summit 2026, took to the stage to discuss how large language models (LLMs) could reshape how viewers navigate sprawling streaming catalogues. Rather than forcing users to scroll endlessly through menus, AI-driven interfaces could and should allow them to ask natural language questions and receive contextual recommendations. In theory, that could dramatically reduce the friction that has long plagued TV interfaces. But in practice, even the speakers were quick…