Discovering Media Insights with AI

Client's Challenge

• VRT's media library, containing millions of items across various formats (videos, news articles, voice recordings, images), is fragmented across multiple database systems, each with unique metadata storage.
• Experienced archivist teams manage these libraries but lack a unified, business-friendly search and exploration function.
• VRT sought to leverage knowledge graph technology to enable smart content search and exploration for archivists, journalists, and program creators.

Our Solution

• We developed a knowledge graph integrating various media libraries, connecting different metadata types and content.
• Utilized natural language processing to extract and relate important named entities (persons, organizations, locations) from content, enriching them with data from Wikidata. For example, a search for "Calatrava" could now include the Luik train station, even if initially absent in the data.
• Recommended content and related topics using the knowledge graph's structure and context, enhancing search relevance and discovery.

Business Impact

• Revolutionized data exploration with a visual representation of connected content, enabling easy and intuitive discovery of new and related topics.
• Reduced dependency on archive experts, allowing business users to explore media autonomously and freeing up experts to focus on data governance.
• Empowered VRT to provide smarter, more efficient content search and exploration, enhancing productivity and the user experience for archivists, journalists, and program creators.