Condé Nast released a multimodal video discovery solution using Amazon Bedrock on an unspecified date, saying it significantly improves editorial workflow. It is the company's first major update to its video discovery system since the launch of its initial search tools.
Condé Nast reported a reduction in discovery time from 250 minutes to under 2 minutes per task, measured across its video library of more than 140,000 videos. That compares with the previous manual process of scrubbing through video titles and descriptions.
The solution is built on Amazon Bedrock and Amazon OpenSearch Service and targets editorial teams needing faster access to video content. Availability begins with a production rollout, initially for brands such as Vogue, GQ, Vanity Fair, and Wired.
"The core problem was structural: Existing search tools can’t look inside video content," said a representative from Condé Nast. The team selected the TwelveLabs Marengo embedding model for its native ability to jointly encode visual, audio, and transcript signals.
The announcement follows a partnership with the AWS Generative AI Innovation Center to address operational inefficiencies in video discovery.
Condé Nast did not say how the solution will scale to non-English content, and it raised the question of how to handle video archives with no metadata. The solution’s architecture is now being used for backfill processing of the 140,000-video library.
Source: awsml
