5 AI agents wrote an article about 5 AI agents: how to scale content
Ksenia Ivanchikova and her team at Generation AI built a multi-agent content pipeline: five specialized AI agents take a 20-minute YouTube talk, turn it into a
YouTube talks remain invisible to search engines: people watch videos, but there's no article. Ksenia Ivanchikova from Generation AI found a solution through a multi-agent system. A pipeline of five AI agents transforms a 20-minute talk into a ready-made case study with screenshots, SEO optimization, and WordPress formatting in 15 minutes.
What happened a year ago
A year ago, Ivanchikova was writing prompts in ChatGPT, hoping to get deep content. The results were disappointing — the text sounded like it was written by a robot, facts got mixed up, structure was vague. It quickly became clear: what was needed wasn't a powerful model, but the right system with role division. She assembled a multi-agent pipeline where each agent is responsible for one step in the chain.
Five agents, five specializations
The pipeline works like a conveyor: you take a YouTube link, upload it to the system, and then it's full automation.
- Research agent watches the video, extracts key points, structures them into an outline
- Processing agent rewrites the outline into an article (500–800 words), adds examples and context
- Illustration agent finds screenshots from the video or generates images in Midjourney
- SEO agent optimizes the text for search queries, writes meta descriptions, places keywords
- Layout agent uploads the article to WordPress, adds tags, arranges blocks, publishes
Each agent receives the output of the previous one, and no one interferes with anyone else's work.
Result doesn't require rework
Over a year, this Generation AI system published 200+ case studies. Most were published without manual editing — the pipeline handles it itself. The article you're reading right now is also the result of this conveyor. That is, the first step toward complete automation of the content business succeeded. Ivanchikova is already planning expansion: an agent for podcasts, integration with TikTok, deeper analysis for specialized media.
"Multi-agent systems are cheaper and faster than one large agent.
Specialization works," says Ivanchikova.
What this means for media
Scaling content stops being about hiring employees. Media startups, blogs of large companies, B2B publications — all will face the same problem: volume grows, speed drops. Multi-agent systems solve this through specialization, not through the power of a single tool.