Condé Nast Accelerates Video Discovery with AWS AI Technologies
Publishing giant Condé Nast has developed a new AI-based system in collaboration with AWS to accelerate video search processes.
Condé Nast has implemented a new multimodal video discovery solution using Amazon Bedrock infrastructure to shorten search times in the massive video library within its archives of brands such as Vogue and GQ.
Processes in the Video Library
Condé Nast editorial teams had to manually search through a vast library containing more than 140,000 videos belonging to major brands such as Vogue, GQ, Vanity Fair, and Wired. This situation caused an average of 250 minutes to be spent on each task.
New AI-Powered Solution
To eliminate this problem, the company partnered with the AWS Generative AI Innovation Center to develop an AI-powered multimodal video discovery system built on Amazon Bedrock and Amazon OpenSearch Service.
Working Principle of the System
The developed solution utilizes the TwelveLabs Marengo embedding model, which has the ability to jointly encode visual, audio, and text signals, and reduces content discovery time to approximately 2 minutes.
Achieved Savings and Results
Thanks to the new architecture, content discovery time decreased by 99.2 percent, manual video review effort dropped by more than 90 percent, and approximately 800,000 dollars in annual operational savings were achieved.