The fast-growing field of AI-assisted indexing and search is explored in this three-presentation session, plus one supporting paper. The first presentation, from RAI, covers unlocking documentary archives through retrieval-augmented generation for automatic retrieval, reuse, and narrative recomposition. The second, from Gracenote, USA, quantifies the benefits of large language models grounded in verified data for entertainment search and discovery. The third, from FlowState AI Inc, USA, presents a multi-agent video question-answering system that searches indexed video moments, localises candidate evidence, and inspects a short clip only when visual verification is needed.

Presented Papers:
'Unlocking documentary archives with multimodal RAG: Automatic retrieval, reuse and narrative recomposition' - Centro Ricerche Innovazione Technologica e Sperimentazione (CRITS), RAI - Radiotelevisione, Italiana
'Quantifying the efficacy of grounded LLMS in entertainment search and discovery' - Gracenote, USA
'Search first, inspect visually when needed: A multi agent architecture for semantic video archival retrieval' - FlowState AI Inc, USA

Supporting Papers:
'Beyond metadata: A unified multimodal framework for automated broadcast orchestration' - Big Blue Marble, ORS Comm GmbH & Co KG, Austria & Big Blue Marble Sp z.o.o, Poland