India-based audio platform Pocket FM has greater than 200,000 hours of content material on the service. Nevertheless, the corporate’s CEO, Rohan Nayak, believes that the platform nonetheless has room to develop when it comes to creating unique content material and increasing its library to a number of genres and sub-genres. The quickest means to do this is to faucet into AI instruments that will assist in audio manufacturing, writing methods, and adapting these tales for various geographies.
“I nonetheless really feel that our content material catalog shouldn’t be adequate for our customers. There are such a lot of genres and sub-genres that we don’t have in our library. I don’t suppose we now have a depth of content material that’s within the mature leisure class,” Nayak instructed TechCrunch over a name.
The corporate already has an excellent partnership with ElevenLabs to transform writings into audio sequence. This has resulted in 5 occasions faster manufacturing and 30 occasions decrease value as in comparison with a professionally generated audio sequence.
“We now have already examined how these AI variations carry out in several markets, and we now have seen encouraging outcomes. We’re nonetheless refining our fashions for errors, however we really feel the expertise is nice sufficient to make use of within the manufacturing of reveals,” Nayak mentioned.
One of many AI instruments Pocket FM is making an attempt out caters to the adoption of tales in several areas. The corporate mentioned it has educated in-house fashions that don’t have a look at mere translation however deal with cultural nuances when changing tales of 1 area to a different.
He added that it’s difficult to resolve the hallucination of the fashions within the context of tales that span a whole bunch of episodes. Pocket FM mentioned it needed to deal with the limitation of context home windows of open-source fashions and in addition construct maps of relations between completely different entities in a narrative to take care of character consistency.
One other software for writers the corporate is testing works as a artistic assistant to them, serving to them with various storylines or giving them concepts about plots. The corporate additionally plans to infuse some insights from its historic knowledge within the software to point to writers what works on the platform.
Nayak talked about that whereas this software is in its early levels, the corporate needs to provide the ability of a author’s room to a solo author, who may be placing out an episode of a day. He famous that the author’s room offers you the power to brainstorm with out taking out artistic controls, and that’s the core concept behind this writing assistant.
Moreover, the corporate is investing in making a blockbuster engine that may feed on insights from the platform about what reveals have gotten a success.
The top aim of Pocket FM is to scale its catalog, provided that it’s producing some content material by itself and creating reveals by way of its community of writers. However to scale and turn out to be extra widespread, it has to create hit reveals.
“Blockbusters is what drives any content material platform. Whereas we now have high of the funnel with user-generated content material, it’s nonetheless a tough drawback to identify blockbusters.”
Pocket FM has seen encouraging outcomes with its AI deployment. It has greater than 40,000 sequence on the platform with assist from AI when it comes to voice creation. Plus, the corporate has generated $3 million in income from them. General, the platform clocked in $127 million in monetary 12 months 2024.
Probably the most tough problem for the corporate is to strike the proper steadiness of AI serving to creatives and churning out content material rapidly. There’s at all times a danger of individuals utilizing AI to hurry up content material manufacturing and compromise high quality. In consequence, the platform turns into crammed with mediocre content material, and it turns into tough for algorithms to spotlight good reveals.
Puneet Sharma, an India-based author and lyricist, identified that in a world the place a lot work is formulaic, artists could have an onus to show their genuine work.
Sharma added that AI instruments may help writers generate concepts and study completely different kinds. However which means within the technique of utilizing these instruments, the method of studying by way of failure may very well be misplaced.
Nayak mentioned that a few of its writers and creatives already use AI instruments. The thought for the corporate is to supply instruments with the context of tales and the platform.
Pocket FM has raised $197 million in funding throughout a number of rounds with backers together with Lightspeed Ventures, Tencent, and Instances Web. The corporate competes on a number of fronts with different content material gamers, corresponding to Audible, Omidyar Community-backed Pratilipi and Google-backed Kuku FM.
