India -based audio series platform Pocket FM Audio is to have Netflix. That is, the company intends to match its audio series with hundreds of episodes of its users’ taste. To work, it needs to be rapidly released – something that is now changing to help with AI.
LightSpeed-supported startup is giving an AI tool set to its authors who can do things like suggesting a better ending or making the story more attractive. Hope the equipment will speed up the story-writing process.
Pocket FM already uses some AI devices such as Elevenlabs to generate voice for the audio chain. It also tested AI tools for internal writing and adaptation assistance.
Pocket FM founder Rohan Nayak said it is rolling AI tools for all authors, so it will take them less time to finish their episodes.
Writing equipment, dubbed Co -pilotCan be used to help any writer make a story.
Copilot can turn narrative-based writing for a specific section into dialogue-based writing. It can also do “beat analysis” to shape writing to make it more attractive to a particular style audio chain. In addition to the tool, there are basic chatbot-style writing characteristics such as “shorten,” “expansion,” and ability to generate lessons through an indication.
To make Copilot, the company examined thousands of hours of data points to understand that the users are more attached to a specific style with a particular story.
Based on this, it added suggestions features designed to increase conflict between characters and recommend endings for an episode, so that it could be more exciting. AI may also suggest tags for background effects that can be used when producing audio.
The equipment can automatically generate BIOS of the characters, their relationships, and briefly submit the plot points of various episodes, allowing the creators to refer to these details back when writing.
Copilot also has a review tool, which checks for plot points, grammar, and gives qualitative response through comments on an episode.
Under the hood, the pocket FM is training small models to maintain reference to a story for character arcs and relationships, as well as with narrative stability. In addition, using signs from users, the startup is naked to add more plays to the story.
International expansion and localization schemes
With the advent of the AI tool, Pocket FM launched adaptation tools for various markets, which not only translate the text from one language to another, but also change the names and phrases that are more favorable for the culture of that region.
The company first launched this device as a part of the Copilot suit in Germany earlier this year. To change stories from other areas allegedly Struggle to attach users in European country last year,
Nayak said that the company saw great results from this test, with a steady increase in monthly in-app revenue, which crossed $ 700,000 in June.
“When we started expanding in new areas, it took us 12–18 months to get meaningful in that market. You must have at least 1,000 hours of content to get users and score the market. Now we can do so in less than three months,” the hero said.
The tool increased the author productivity for the German market by 50% in terms of show output. Also, the tool helped the company to create more error-free drafts of those shows, resulting in high user retention for the audio series.
In the US, the series created with the help of these new AI tools is now contributing 10% platte. In addition, these shows have produced $ 7 million in revenue in the last 12 months, while the cost of production has reduced by 2-3 times.
Build technology to score material production
As a result of the adoption of different -different AI features, the pocket FM has been able to score the content quickly. The startup stated that it launches close to 1,000 pilots per month. And just some of them become hits in the sheer amount of material.
But the audio show is just a part. The company is already working on equipment to change stories in comic strips. Pure tons platform. In addition, the hero said that the video is a possible format that the company can also see. Startups, which have collected more than $ 196 million in funding in the round, are also using a micro drama app.
By next year, Pocket FM wants to release its own unique big language model (LLM), which will be based on the data collected from its show and include various tools such as writing aid, adoption, dramatic and reference retention. The company said that when it switchs to its own LLM, it would not need to train a ton small model for different features.
Possible decline of AI
Adopting AI has had side effects.
Pocket FM has already kept people who were employees or contractors across Multiple Round over the past 12 months. There have also been reports of watching writers Low returns For a long time. And the company is facing California trial on employment and wage issues,
“Like the industries led by most materials, we work with diverse networks of writers, voice artists and production partners based on a project, stitching resources in each market. AI has the least impact on our main creative community; instead, it has opened new avenues to expand and expand the production,” a company rape said in response.
There are also questions around quality. The company measures quality by the retention number of a show.
The base argument is that new AI equipment also serves as a room for singles for single creators, so they will be able to produce more materials at a rapid rate. In addition, depending on the numbers, the author can edit the story quickly with the help of AI. However, these devices can very well induce the “AI slope”-or in low quality, AI-generated material-platform and affect the user’s recommendations, making it difficult for them to find good stories.
Pocket FM argues that despite helping AI, stories with a solid structure will get popularity.
The company stated that each piece of material is reviewed by its AI-managed moderation structure to ensure quality and originality. It also claims its AI moderation check for things such as repetition, copyright issues, content health and other quality measures before approve the audio to live the audio. Each show is obtained and the same push, and user engagement eventually determines the ranking of a show.
Another concern is that the author may be highly dependent on the AI over time.
In Germany, AI is writing more content than humans per show for selective title. With the plans of pocket FM to roll more AI devices, the amount of AI-written material may increase. And with this, the hope of churning more shows may also increase. Until user adoption also does not grow rapidly, average returns can fall.
The company did not address the questions of Techchchan about direct returns, but said that its AI equipment could speed up a writer’s work and help them edit an episode based on the number and audience’s response. This is that they can make targeted improvements, rather than a complete rewriting.
A spokesperson said in a statement, “In this way, rapid material construction does not necessarily dilute quality or relevance;
We are always looking to develop, and techcrunch and by providing some insight into our coverage and events in our perspective and response, you can help us! to fill This survey To tell us how we are doing and get a chance to win the award in return!
