Why We Moved Beyond Manual Scriptwriting and Built Our Own AI Editor

In 2026, we set an ambitious goal: dramatically increase the amount of content we publish across our platforms. The challenge wasn't deciding what to create – it was figuring out how to scale without expanding the team at the same pace.

Our projects span multiple formats. Some require in-depth interviews, others are news shows, and some are podcast episodes. Each format has its own workflow, but they all share one bottleneck.

• It's not filming.
• It's not editing.
• It's preparation.

Researching topics, analyzing information, and writing scripts have always been the most time-consuming parts of the process. A script for a single news podcast can easily reach 90 pages. Producing that manually means hours of research, fact-checking, and writing before a camera is ever turned on.

That's why we built Screenwriter.

From Audience Insights to a Finished Script

We weren't interested in building yet another AI tool that simply generates text.

Our goal was much more ambitious: create a system that understands what matters to our audience and helps us produce content they'll actually want to watch.

Everything starts with understanding who that audience is.

Traditional analytics can tell you someone's age, interests, or browsing behavior. That's useful, but it doesn't explain why people choose one piece of content over another – or what motivates them to engage with it in the first place.

To answer those questions, we used large language models (LLMs) to combine the data we already had into a much richer psychological and behavioral profile.

Inside the system, this profile evolved into what we internally call the Strategic Creator persona – a representation of our core audience, complete with its own motivations, goals, interests, and preferred ways of consuming information.

That changed how the entire system worked.

Instead of asking: "What happened today?" 

Screenwriter started asking: "Which of today's events would actually matter to this audience?"

That shift became the foundation of everything that followed.

How Screenwriter Finds the Right Stories

Once Screenwriter understands who the content is for, it starts processing information.

The system can connect to virtually any source –Telegram channels, news websites, industry publications, and more. It isn't limited by language, either. Whether the source is in English, Russian, Chinese, Spanish, or another language, Screenwriter analyzes it the same way.

But collecting information is only the first step.

The real value comes from filtering it.

Rather than presenting editors with an endless stream of news, Screenwriter compares every story against the audience profile and eliminates anything unlikely to resonate.

What's left is a curated list of topics with the highest potential value for our audience.

During the first eight weeks after launch, Screenwriter analyzed roughly 240,000 news stories.

From that dataset, it automatically identified the most important developments, estimated their relevance to our audience, and ranked them by priority.

Instead of spending hours monitoring dozens of news sources, editors start with a shortlist of stories that are already worth exploring.
That's a fundamentally different workflow.

Instead of searching for ideas, the team spends its time developing the best ones.

From Breaking News to a Ready-to-Use Script

Finding the right story is only half the job.

Once a topic has been selected, Screenwriter can turn it into a complete script within minutes. Whether it's a single news segment, a full news show, or the foundation for a podcast episode, the system generates a structured draft that's ready for editorial review.

What once required hours of research, outlining, and writing now happens in a fraction of the time.

That doesn't mean the process is fully automated. It means editors no longer have to start from a blank page.

Instead of spending their time gathering information and organizing it into a coherent narrative, they can focus on what humans do best: refining ideas, adding context, and shaping the story.

Why We Chose a Multi-Model AI Architecture

One thing became clear early on: no single AI model excels at every task.

Some models are better at processing large volumes of information. Others are stronger at organizing complex material into a clear structure. And some consistently produce better writing than the rest.

Rather than forcing one model to do everything, we designed Screenwriter as a multi-model system.

Each stage of the workflow is handled by the model best suited for that particular job. Information analysis, topic prioritization, content planning, and script generation are distributed across different models, allowing each one to play to its strengths.
The result is better quality – and a more reliable system.

There was another reason for this architectural decision.

AI services evolve quickly. Models change, APIs are updated, pricing shifts, and providers occasionally experience outages or impose new usage limits.

If your entire workflow depends on a single model, any disruption can bring production to a halt.

We wanted to avoid that risk from day one.

By designing Screenwriter around multiple AI providers, we built a platform that remains resilient even when individual models become unavailable. As the AI landscape evolves, we can adopt new models, replace existing ones, or combine multiple systems without disrupting the workflow.

In practice, that means the platform is designed to evolve alongside the technology it depends on.

What Screenwriter Can Do Today

Today, Screenwriter supports far more than script generation.

It helps our team identify the most relevant stories, prepare news episodes, research interview topics, and generate thoughtful questions for podcast guests.

Tasks that once required hours – or even days – of manual work can now be completed in a small fraction of that time.
More importantly, every part of the workflow is connected.

Instead of switching between research tools, spreadsheets, note-taking apps, and AI chatbots, our team works inside a single system that takes content from the earliest idea all the way to a production-ready script.

For us, Screenwriter isn't just another AI assistant.

It's become the backbone of our editorial workflow.