Tivo Looking for Metadata to Feed AI Engines

We met up with Tivo and had a wide ranging chat about the impact of IP technologies. However, we also got an update on Tivo’s developments which continue to revolve around ‘amplifying value’ and exploiting its depth of information on TV content that it has from its long-standing provision of TV guides. Consumers, these days, want any content wherever they are and that is a challenge. There is also the question of how you should best apply the intelligence that is starting to be available.

These days, the Tivo front end can allow access to content that is from OTT channels and services such as Google and YouTube. Tivo works closely with Android TV and is one of two special developers and it has APIs for Android, Linux, Roku and Apple TV.

Increasingly, Tivo finds that it is needing longer form descriptions, as AI technology works better when there is more data. However, there is often little metadata available for library or historic content, so getting the search functions to work as well with this content is a continuing challenge. The company has standard methods for checking the quality of incoming metadata and processing it.

It has support now for search in German, Italian and Portugese and is very much driven by customer demand and interest when developing new language support. The company has also worked with Com Hem on a personalised content discovery project (TV hub) based around a non-DVR Android STB and the Android Operator Tier. The box has support for YouTube, Netflix, SVT Play (Swedish) and TV4 Play (a Swedish commercial channel). These four cover 75% of OTT consumption in Sweden.