The New Discovery Stack: AI, Metadata and Audience Intelligence

Insights from NAGRAVISION's latest industry webinar featuring One Hungary, Liberty Global and Media Press Group

In this blog, Laura Rognoni explores the key insights from NAGRAVISION's latest industry webinar, Achieving Leadership in AI-Driven Content Discovery, presented in partnership with Adwanted, and examines why discovery is rapidly evolving from a supporting feature into a strategic capability.

Discovery has become one of the most important battlegrounds in television. As audiences navigate an ever-expanding mix of linear channels, streaming services and on-demand content, the challenge is no longer access. It is helping viewers quickly find something they genuinely want to watch.

This was the central theme of NAGRAVISION's recent webinar, hosted in partnership with Adwanted and moderated by The Media Leader Editor, John Moulding. We were delighted to be joined by Ben Gidley, Senior Director Product Management for OpenTV ENTera at NAGRAVISION, Ren ta F l p- rvai, Head of TV & Content Services at One Hungary, Chris Van Der Linden, Director of AI Roadmap at Liberty Global, and Rafal Fagas, CTO at Media Press Group. Bringing together industry experts, the discussion explored how AI, enriched metadata and audience intelligence are reshaping the future of content discovery.

Discovery at the Heart of the Customer Experience For years, operators competed by expanding content libraries and adding new services. Today, content abundance has shifted the focus towards navigation. Viewers increasingly expect TV platforms to behave less like traditional electronic programme guides and more like intelligent assistants. Search is becoming conversational, discovery is becoming contextual, and personalization is moving beyond static recommendations. As discussed during the webinar, the next generation of discovery is not simply about matching keywords to content. It is about understanding viewer intent.

Perhaps most importantly, people are no longer searching in neat metadata categories. They search emotionally. A viewer may be looking for something easy to watch after work , or a documentary I can dip in and out of over the weekend . Traditional metadata structures were never really designed to support these kinds of interactions. This creates both a challenge and an opportunity for operators. The platforms that can best understand viewer intent and reduce friction between audiences and content will increasingly define engagement, loyalty and long-term value.

Chris Van Der Linden highlighted this evolution through Liberty Global's AI-powered Super Search deployment with Sunrise & Vodafone Ziggo, where generative AI is used to better understand what viewers are looking for before returning results. Rather than processing queries literally, the technology interprets context and intent to deliver more relevant recommendations. Sports offer a particularly compelling example. Fans are often looking for far more than a live match, they may want interviews, player statistics, highlights, league information or related content. AI creates new opportunities to connect audiences with these experiences faster and more intuitively.

Personalization Beyond the Profile One of the clearest messages from the discussion was that TV personalization remains fundamentally different from personalization on smartphones or laptops. Televisions are shared devices, often used by multiple members of a household throughout the day. Understanding who is watching, and what they may want to watch in a particular moment, requires a more sophisticated approach. This is where platforms such as OpenTV ENTera are helping operators unlock greater value from the behavioural data they already possess.

As Ben Gidley noted during the discussion, the average viewer watches around 112 hours of television each month and may visit the operator home screen roughly 200 times1. Every interaction creates valuable signals, from viewing habits and package choices to browsing behaviour, session times and content engagement. The real opportunity lies in using those signals intelligently. Rather than relying solely on collaborative filtering, NAGRAVISION combines recommendation techniques with enriched metadata and behavioural data to create more relevant experiences. Continuous testing and optimisation help operators measure the impact and demonstrate a clear return on investment.

For operators such as One Hungary, the goal is to create increasingly fluid user experiences that adapt to different audiences and viewing contexts. A casual viewer may require a simple and highly curated home screen, while more engaged viewers may benefit from deeper personalization and richer content exploration. The industry is moving steadily towards interfaces that adapt dynamically rather than relying on static user profiles alone.

Unlocking Content Intelligence with AI Metadata was once largely limited to genres, cast information and programme descriptions. Today, advances in AI are enabling metadata enrichment at a much deeper level, incorporating emotional tone, contextual relevance, visual understanding, chapterisation and enhanced searchability. As Rafal Fagas explained, the challenge is that much of the industry's source metadata remains relatively traditional. Scaling richer content understanding across vast catalogues requires a combination of automation, AI and editorial oversight.

The discussion also highlighted an important reality: while AI can dramatically accelerate metadata creation and enrichment, trust remains essential. Human expertise, source validation and quality control continue to play a critical role in ensuring accuracy and consistency.

The webinar audience was asked what level of engagement uplift operators could realistically expect from AI-powered content discovery. Most respondents select

LINK:	https://nagra.vision/newsroom/blog/the-new-discovery-stack-ai-metadata...
	See more stories from nagra

Most recent headlines