ThinkAnalytics is to introduce a step change in its suite of recommendation, search, and discovery services.
Debuting at IBC2025 later this week, ThinkMetadataAI uses agentic AI to fully automate the enrichment of metadata for entire content catalogues, maintaining high quality throughout.
ThinkMetadataAI is the first market offering to eliminate manual tagging entirely while supporting numerous languages and complementing all existing metadata provider solutions. During IBC, ThinkAnalytics will demonstrate how automated content tagging – including live TV – significantly enhances search, discovery, recommendations, and personalisation.
ThinkMetadataAI is now integrated into ThinkMediaAI, ThinkAnalytics’ unified platform for content monetisation, contextual advertising, content curation, and bundling, and is available both as part of the platform and as a standalone product for customers.
Customers are also selecting ThinkMediaAI for its ability to personalise experiences for anonymous users, including those on FAST services who do not require login credentials. The platform can deliver tailored recommendations after just one second of viewing, utilising factors such as time of day, location, device type, and more.
“Since we launched ThinkMediaAI earlier this year, the reaction from customers has been exceptional as they achieve dramatic operational efficiencies that cut costs while continuing to increase engagement with viewers,” said Peter Docherty, CTO and co-founder, ThinkAnalytics.
“Our years of experience in AI and discovery has given us a clear lead in understanding the DNA of content. Combining this with our use of agentic AI breaks new ground and delivers rich metadata DNA unlike anything we have seen previously, ushering in a new era of viewing based on deeper personalization,” Docherty added.
To confirm the quality control of its agentic AI metadata, ThinkAnalytics compared results against manually tagged content. The analysis revealed that ThinkMetadataAI identified additional elements for tagging that had not been surfaced by the manual tagging.
