Your company has produced hours of expert content — webinars, product deep-dives, customer sessions, conference talks — and stores it as a YouTube channel nobody can search by what was said. The knowledge inside is real: your best SE explaining the integration story, the CEO's sharpest articulation of the category, the customer Q&A that answered every objection. Functionally, it's write-only memory with a play button.

The waste compounds through turnover: when the SE leaves, their recorded explanations remain — unfindable. New teammates re-derive answers the archive already holds. The content team plans next quarter's webinar without a searchable record of what the last twelve covered.

library_import_youtube_channel converts the channel into memory: one import, every video, transcripts included. library_search makes the shelves navigable by title, tag and type — the instant, free layer that answers "do we have a session on X?" before anyone re-records one.

From there, the whole library stack applies to your own content: deep_context interrogation for "how did we explain <feature> in past webinars?" (cited to the minute — see the cite-the-exact-minute use case), repurposing via the webinar-to-SEO workflow, case-study extraction from customer sessions. The archive stops depreciating and starts compounding — which is what institutional memory means.