e4m webinar: Consumer data can build AI to enhance personalization & UX: Experts

Industry experts were speaking at e4m webinar powered by Myelin Foundry on 'Transforming Viewer Experience on OTT using Artificial Intelligence'

e4m by exchange4media Staff
Updated: May 16, 2020 9:19 AM
e4m webinar powered by Myelin Foundry: Transforming Viewer Experience on OTT using Artificial Intelligence

“There are around 35-40 large over-the-top (OTT) companies operating in India, and there are around 50 million-odd people who are currently using one or other OTT platforms,” said Prashant Rao. He was giving his opening remarks as the session chair for e4m webinar powered by Myelin Foundry on the topic 'Transforming Viewer Experience on OTT using Artificial Intelligence'.

The virtual discussion was brought together by exchange4media and Myelin Foundry. The panel for the webinar comprised of Manish Verma, Head of Technology, SonyLIV; Lokesh Chauhan, Chief Technology Officer, Eros Now; Mike Carr, Former Chief Science Officer, BT; Dr Gopichand Katragadda, Founder and CEO, Myelin Foundry; and Shahabuddin Sheikh, Chief Technology Officer, Alt Balaji.

Rao further observed that on an average a person spends 3.5 hours on an OTT platform on daily basis, and leaves behind a trail of data about who is watching what, for how much time and during which time of the day.

With this observation, he opened the discussion and asked Katragadda, “How OTT players can use the said data to enhance the consumer experience?”

Answering the question, Katragadda said, “There are three journeys on OTT platforms in which artificial intelligence (AI) plays a big role. The journey of a story; the journey of a user; and the journey of pixel and audio bite. The consumer data can be used to create an AI that can assist players in all three journeys.”

He further elaborated that AI can be used for content generation, to modernizing it. It is also used for plot generation and could also assist the user in finding the right content. It can also be used to improve and fix the pixels. “The user data can help in building the AI to improve the efficiency and quality of OTT platforms”, he said.

Adding on to Katragadda's statement, Carr said, “One thing we don’t talk enough is the difference between real-time data like sports and live discussions verse movie”.

He said that it is easy to achieve optimum quality images for movies than for live events, and in order to improve the image quality of live events, we have to work on AI that knows what good picture looks like. “If it can learn what a good picture looks like, and apply it to real-time content, user experience can be improved,” he said.

Talking the discussion further, Rao asked Shaikh- what kind of insights go into recommending content to users?

Answering, Shaikh said, “You have to identify user habits and their preferences, and accordingly program your content. AI plays a major role in monitoring the consumer’s behaviour, which helps in recommending new content to the user.”

He added, “Fresh content is something the user hasn’t watched, and recommendations change depending upon what time of day the user comes on OTT platforms and what screen they are using”.

Taking Shaikh’s statement further, Verma observed that recommendation is something important for engagement on the OTT platform. “AI plays an important part, right from the processing of the content. It helps in enriching the Metadata, and if we have enriched metadata, we will be able to give better recommendations,” he said.

He added that there are limitations to how much Metadata can be entered the system manually. AI comes in handy in generating Metadata and putting markers on it in the correct places based on key moments in the process.

He further said, “AI can be useful not just for generating recommendations, but for personalization as well. AI can stop consumers from churning.”

He explained that AI can be used to monitor consumers’ consumption patterns and recommend not just the content, but subscription plans as well. It can be used to motivate users to upgrade from free walk to paid content.

“There is a complete funnel that we manage. At each level, we need to make sure we are capturing the right kind of data to convert the consumer behaviour into a habit,” he added.

Chauhan added to Verma's statement and said that the angle of personalization that comes with OTT is what makes different from normal TV. “On OTT, the user can select what they want to consume on a large screen with family and what they want to on a smaller screen, alone. The content that is recommended can be identified based on location, time of the day, day of the week, recent event”, he said.

He further added that the engagement matrix is the key factor for a subscription business. “If we don’t have engagement, and not able to add newer content, and resurface the older ones; if we are not able to identify the taste of the user based upon when it makes sense for them, we won’t be able to engage them for longer. The price of engagement is a constant drive,” he concluded.

For more updates, be socially connected with us on
WhatsApp, Instagram, LinkedIn, Twitter, Facebook & Youtube