How I Automated 90% of My YouTube Workflow

A developer automated 90% of their YouTube workflow, reducing production time from two to three days to about five hours, by building a pipeline of specialized AI agents for research, writing, visuals, and titles while keeping creative control in human hands. The system, developed with Paul Iusztin and the Towards AI team, uses agents named Nova for research and Brown for writing, alongside parallel workers for branded visuals and an expand-and-narrow loop for titles.

Member-only story Featured How I Automated 90% of My YouTube Workflow How I automate research, writing, visuals, and titles while keeping the thinking human I can now research, write, and prepare a full YouTube release in about five hours of my own work. The same job used to consume two or three full days. I did not get there by asking one model for a script and publishing whatever came back. Paul Iusztin, the Towards AI team, and I built a pipeline that gives specialized agents the repetitive translation work while keeping the thinking in human hands. If you want to build or inspect the system, start here: Take the Agentic AI Engineering Course https://academy.towardsai.net/courses/agent-engineering?ref=1f9b29 Explore the open-source course repository https://github.com/towardsai/agentic-ai-engineering-course Use my AI Engineering Cheatsheets https://github.com/louisfb01/ai-engineering-cheatsheets This article explains the full architecture: how Nova researches, how Brown writes without turning everything into generic AI prose, how parallel workers produce branded visuals, how an expand-and-narrow loop generates titles, and why the most important part of the system is the 10% I refuse to automate. Prefer to see the pipeline and diagrams in motion? Watch the complete walkthrough here, then use this article as the technical reference.