Key results
The company
Mediahuis
mediahuis.comEuropean media group publishing newspapers, magazines, and digital news brands.
Result highlights
- Research time cut from 30 mins to seconds
The story
A European media group operating across five countries publishes over 30 independent news brands, managing a vast archive containing millions of historical articles.
Journalists struggled to efficiently access historical context, with background research for a single story taking up to 30 minutes. This latency hindered continuous reporting and made it difficult for staff to leverage the archive for fact-checking or developing new story angles.
The organization built a custom tool that combines Elasticsearch for document retrieval with Vertex AI and Gemini models for content analysis. The system processes natural language queries to verify facts, generate summaries, and suggest editorial angles based exclusively on trusted internal data. This solution was developed and deployed within six months, integrating directly into journalist workflows to replace manual archival research.
Scope & timeline
- AI search tool operational in 6 months
Quotes
“Our methods for accessing our archives clearly could be more efficient: Researching and finding background information for a story could take up to 30 minutes. When you're reporting news continuously, this situation is no longer tenable.”
Implementation partner
ML6
ml6.euHelped Mediahuis implement AI solutions like NewsSearch using Google Cloud and Gemini models.