Artificial Intelligence1 min read

How Social Department built culture-aware video discovery

By · Published by Everything Blog

In short

Built on Amazon Web Services infrastructure, the tool translates nuanced editorial concepts into automated pipelines that surface shareable video moments. The technical architecture relies on TwelveLabs video reasoning models, Anthropic's Claude, and Amazon OpenSearch Serverless, with integration executed by TrackIt. To maintain quality and safety, the platform verifies prospective quotes against official transcripts and applies Amazon Bedrock Guardrails to screen incoming queries. Human operators retain full final authority by reviewing and approving every generated post before publication.

Key points

  • Subjective editorial criteria can be successfully translated into automated discovery pip…: Subjective editorial criteria can be successfully translated into automated discovery pipelines for entertainment video catalogs.
  • Responsible AI frameworks ensure automated tools propose candidate content while humans r…: Responsible AI frameworks ensure automated tools propose candidate content while humans retain final publishing approval.
  • Rigorous source verification checks generated quotes directly against episode transcripts…: Rigorous source verification checks generated quotes directly against episode transcripts to prevent factual errors.

Sources checked by our analysis pipeline: 1

Artificial Intelligence

How Social Department built culture-aware video discovery

This post is co-written by Mariah Miller and Sirajus Salekin from AWS and Nicholas Chase from Social Department.

Entertainment marketing teams manage tens of thousands of video assets across studios, networks, and streaming applications. Most campaigns begin with the same video discovery problem: finding the one moment in the footage that will resonate with an audience. When a brief lands, the clock starts, and what they need is almost never something a database can describe.

Semantic search, on its own, has a ceiling: It answers the question you can articulate. Ask it for a character holding a coffee mug and it will find every coffee mug in a thousand episodes. The queries that actually drive entertainment campaigns look nothing like that. They are questions of judgment: which moment in a show will travel, which line is quotable, and which single frame will work as a meme.

Those questions can’t be typed into a search box. Answering them requires two things the search index doesn’t have: an editorial definition of what “quotable”…

Artificial Intelligence