Short videos have become the true virality of modern social media. Instagram, Facebook, YouTube, and TikTok fill feeds with billions of views every day, giving businesses a huge advantage in the global content landscape.
For creators, this means constant pressure to produce more content and find new ways to capture their audience’s attention. In particular, we consider the growing share of artificial intelligence filling platforms and the changing rules of the game in the content industry.
Jay Nio, creator and former head of short video content at MrBeast, believes artificial intelligence can help creators determine what works for them and provide new content ideas. To achieve this goal, he is developing the Palo platform with Shivam Kumar, a former Palantir engineer, and Harry Jones, a team aimed at supporting creators.
Nio, Kumar, and Jones are the key figures behind Palo’s concept, which aims to integrate analytics, ideas, and a community of creators directly into one system.
“I was very obsessed with retention graphs and why viewers stayed and why they left. I had a document with all of that. Then my role became more responsible for editing and ideation.”
– Jay Nio, TechCrunch
What is Paro and how does it work?
According to the developers, Palo consists of three main parts. AI tools, analytics, and community for ideation and planning. Once the creator has connected all their accounts, the tool analyzes the short videos and provides conclusions about what works and what doesn’t.
“Palo has three main parts: AI tools for ideation and planning, analytics, and a community. The company engages with creators and asks them to unite all their accounts. The tool analyzes all the short videos and provides conclusions about what works and what doesn’t.”
– Paro
Shivam Kumar, the startup’s CTO, explained that Palo uses a blend of models to build data trees that provide insights into hooks, audience emotional responses, topics of interest, originality, and potentially related search queries.
“Shivam Kumar, the startup’s CTO, explained that Palo uses a blend of models to build data trees that provide insights into hooks, audience emotional responses, topics of interest, originality, and potentially relevant search queries.”
– Shivam Kumar
Next is the mechanism for drawing conclusions. The raw data is processed using a cocktail of high-level, large-scale language models that enable the creation of memory caches, embeddings, and other structured data that help shape personas according to the creator’s style and preferences.
“An inference engine that takes these raw data and applies a cocktail of upper-level LLMs to hierarchically aggregate these data into memory caches, embeddings that can be retrieved later semantically, and other structured data formats. All of this helps build a creator persona that fits the creator’s tastes and style.”
– Shivam Kumar
The AI-based planning tool has a conversational interface where creators can ask general questions about their content and ask the tool to draft a script according to a given formula. If the creator is primarily visual and doesn’t talk much in the video, this tool can create storyboards with different hooks.
“The AI-based planning tool has a conversational interface, similar to other chat bots, where creators can ask general questions about their content. They can also ask the tool to create a script according to a specified formula. If the creator is more visual and doesn’t have much voice in the video, the tool can create storyboards with different hooks.”
– Shivam Kumar
Although Palo is in development, the platform’s community still has the option to exchange messages between creators. At this stage, the company is conducting a pilot test with a small group of authors.
During the beta phase, Palo worked with around 40 creators with a total of over 1 million users. The platform is currently available to creators with 100,000 followers for an initial price of $250 per month, with higher tiers potentially offering more access.
The company has raised $3.8 million in funding from Surge Peak XV (formerly Sequoia India) with participation from NFX and private investors.
Rajan Anand, managing director of Peak XV, said the introduction to Paro was possible thanks to one of his mentors at Nio. He highlighted that the team’s experience in creative collaboration and technical understanding drove the investment.
“Creators around the world are looking for tools to ease the process without taking away their voice. Jay and the team were very clear about where there was real value and where there was no value, which gave us complete confidence. AI will enable us to create a new category of systems that focus on creator identity and learning deeply from the world’s best creators.
– Rajan Anand
“Josh Constine, former TechCrunch editor and Paro investor, said the tool could help creators keep up with the rapidly changing flow of content.”
“Palo’s emergence comes at a time of notable tension between artificial intelligence and the creator community. Platforms like TikTok, Meta, and Google are expanding their AI-powered toolkits. Creators are already using these tools, but some, like MrBeast, are warning of the potential negative impact on the industry.”
– TechCrunch
The company emphasizes that the main task of building AI tools for creators is not to diminish their creativity, but to give them new paths to success without standardizing their style. Palo aims to provide creators with tools that improve decision-making and reduce tedium while preserving each author’s individuality.
“The main challenge when building AI tools for creators is avoiding the formalized rut of creating similar content. Paro aims to push creators in a direction with a higher chance of success, but their creative instincts are still important.”
– Shivam Kumar
“Here’s the parallel: When comedians test new material on stage, they consciously and subconsciously collect data about whether the audience laughs or not. Each show becomes an iteration, and new audiences can benefit from what comedians have learned before. We believe AI can give creators a similar edge,” Nio said.
“Here’s the parallel: When a comedian tests new material on stage, he consciously and subconsciously collects data about whether the audience laughs or not. Each show becomes an iteration, and new viewers benefit from what the comedian has learned before. We believe AI can give creators a similar edge.”
– Jay Nio
