Datansh helped a digital news platform move from region-based content delivery to a behavior-driven recommendation system that used scalable data processing, AI models, and natural-language content discovery to personalize the reader experience.
~22% higher click-through rate from more relevant, personalized recommendations.
~18% longer reader sessions through behavior-driven matching and better content relevance.
~15% lower bounce or churn behavior by reducing irrelevant content exposure and improving stickiness.
The platform relied on region-based recommendations instead of individual reader preferences. That made discovery generic, forced users to search manually for relevant stories, and limited the publisher's ability to improve engagement, subscriptions, and monetization.
Datansh built an AI-driven recommendation engine that tracked reader behavior in real time, processed interaction data through Databricks, used model-driven personalization to improve recommendations, and added AI-based natural-language query handling for faster content discovery.