It intends to use the technology to deliver a more personalised experience for audiences, based on consumption habits and context, as well as to offer advertising partners more targeted options.
Mr Saraceni said digital audio companies should draw on AI and machine learning to gain a “deep understanding of your users and what they’re interested in and passionate about”.
SourseAI CEO Matt Jones said the AI and machine learning platform could let broadcasters offer Netflix-style program suggestions.
“That’s really hard to do just on data and analytics alone,” he said.
“The power of AI and machine learning is that it’s much faster than we could ever be coming out with a theory and testing it and then taking the results into that next phase of personalisation.”
SourseAI CEO Matt Jones said the AI and machine learning platform pulls in audience data, listening behaviour, programming data such as show descriptions and talent and provides tools to forecast audience numbers.
“And, more importantly, for the listener, it personalises the content recommendations. Just like Netflix does, it learns your interests and tastes,” he said.
“When a user comes into the LiSTNR app, they’re going to get a personalised feed of content that considers all of their behaviours: What content do they like to listen to? What time of the day? What did [they] listen to recently, or listen to frequently?”
Mr Jones said SourseAI also drills down into podcasts to find topics and provide audiences with recommendations.
User engagement
“If you’re listening to a podcast and they’re talking about the NRL, we can make suggestions about other podcasts that might also be talking about the NRL,” he explained.
For SCA, it is about keeping the LiSTNR audience moving around the platform, discovering more content and potentially being served ads specific to their interests and who they are.
“Advertisers want to be able to target users at scale. And what Sourse is showing us is that the way that people listen to audio changes across the day. You can’t segment users like you normally would,” Mr Saraceni said.
“The way we refer to it with Sourse is dynamic cohort. The notion that your segment can shift and change would be so hard to do without AI and machine learning.
“We think that’s going to unlock a whole bunch of commercial value as well. To be able to get people who we know are passionate about different things at different times and give them access to advertisers who want to matchmake with them as well.”
For SCA , the investment comes after it invested in Melbourne start-up Sonnant, which uses AI and machine learning for content discovery for the spoken word, in March.
Mr Saraceni said the two companies complement each other to create better recommendations for audiences.
“Sonnant can deeply understand what our content is about by having an AI listen to it, and then pull out the topics from it. So the topics that Sonnant pulls out become an ingredient for the recommendation of Sourse,” he explained.
“So we’ve got Sonnant understanding what’s beyond just a title and description, what’s inside our spoken word content. And then we’ve got Sourse going, ‘thank you for that information, I will now use that in a variety of different ways to help match that content with the users and what we know about them’.”
For SourseAI, the investment round will help it double its existing team of five in the first half of next year and help expand its target markets from media and telco companies to include utilities.
SourseAI is working with six brands, including SCA, and is in the proof of concept stage with about six more brands, including some major retailers.
“We’ll be doubling the customer base every half year for the next little while,” Mr Jones said.