Arash Pendari, Founder and Creative Director at Vionlabs, has offered his insights on how streaming services will benefit from AI in 2020 and beyond.
Vionlabs is a media and entertainment technology company that was founded to solve the issue of online video viewers spending too much screen time looking for something they want to watch. It does this through the development of a highly innovative AI, that combines multi-layered deep content analysis and the consumers watch-history to provide an advanced content discovery platform.
“Competition in the direct-to-consumer video market dramatically intensified in 2019, with Apple, Disney and UK broadcasters all launching services to compete with the likes of Netflix and get a slice of the streaming pie. The streaming wars are expected to rumble on in 2020 but while the battle-lines have been drawn around content, we expect to see a much greater focus on content discovery as streaming providers start to realize that consumers are splitting their time across multiple services and have less patience for sifting through libraries to find relevant programming,” said Pendari
Pendari’s top 3 predictions on the role of AI and machine learning on content discovery, and how it will impact operators and enhance and differentiate consumer experiences, are:
1) A picture is worth a thousand words: Traditional approaches to content discovery rely on the use of metadata, which broadly labels content with keywords – from the actors to the genre – to provide similar recommendations. However, metadata on its own provides simplistic and often inaccurate recommendations. In 2020, we’ll see AI and machine learning prove that a picture is worth a thousand words when it comes to content discovery by analysing an entire video asset including, the length, pace, soundtrack, colour palette, energy and emotions in a film or TV show. This provides a more in-depth, and accurate analysis of a viewer’s content preferences and viewing habits, leading to recommendations that are more relevant and increased viewer engagement.
2) Understanding content to understand the viewer: Current content discovery systems typically focus on the viewer and don’t try to understand the content he or she is watching. For example, they will focus on the types of TV shows and movies a viewer is watching and use that to inform recommendations. However, that is not enough to offer truly relevant suggestions. AI and machine learning help content service providers identify nuances in programming that are missed by metadata and match them with watching habits, such as when a viewer prefers to watch them, whether they like to binge watch certain types of movies and TV shows and what they like to watch afterwards. By taking a programming-centric approach, streaming providers can better understand their viewers and provide more intuitive and personalised services.
3) Bringing the lean-back experience back: The linear TV market may be falling behind the non-linear world in many ways, but it still champions the lean-back viewing experience where consumers can turn on the TV and not have to spend lots of time and energy looking for things to watch. In comparison, the non-linear world requires viewers to lean in and actively engage with a service. We expect the video streaming industry to follow the lead of popular music streaming services, removing the burden of content discovery and recreating the lean-back experience of linear TV, by using AI and machine learning to create curated channels and playlists specific to individual users.