Zero to Agent in 30 Minutes: Build a Content Engine with Max Johnson

Max Johnson, founder of the AI agency briix, regularly publishes practical AI guidance for business owners and founders to help them get the most out of AI. Max has used some of that know-how to streamline his own day-to-day work, turning what used to be a manual content creation process—researching topics, judging their relevance, and developing hooks and drafting scripts—into a single automated workflow using Claude Code. In this episode of Zero to Agent in 30 Minutes, he shows you how to do the same while remaining in the loop to make the final calls.

How to build a content creation agent

1. Give the system the context it needs to generate relevant content. Max started with a small knowledge base containing information about himself and his business, details about his audience, sample scripts, and notes on his writing voice. These files gave Claude Code reference material for evaluating topics and generating scripts that reflected his brand and style.
2. Explain the process in plain English and let AI take it from there. As Max pointed out, “Vibe coding is describing what you want clearly and let[ting] the model handle the entire building process for you.” However, you’re still ultimately responsible for what’s built. Claude Code may have generated the implementation, but Max set the context, approved permissions, reviewed results, and answered clarifying questions along the way.
3. Build the individual stages of your pipeline. Max wrote a list of prompts to take Claude Code through the process of researching topics (stage 1), scoring them and selecting the most relevant one (stage 2), then generating three hooks and expanding the strongest option into a full script based on the knowledge base and sample content (stage 3). He ran them through Claude Code and ended up with a working three-part pipeline, with each stage triggered manually.
4. Connect the stages into a single workflow you can run with one command. Max had Claude Code combine research, scoring, hook generation, and script writing into a unified content engine that produces a structured JSON file he uses in the next stage. “This is the actual moment that it stops being a few separate prompts and starts becoming an agent,” Max says.
5. Add a browser-based interface. Max built a local dashboard that displays the generated topics, scores, hooks, and scripts. This lets him inspect the results and start new runs from a browser instead of returning to the terminal each time. Here’s a tip from Max on creating a dashboard you’ll actually want to use as your mission control: Be as creative as you want, but “make it look like something designed on purpose, not a default template.”
6. Test the workflow and extend it. The first version Max created had two limitations: It produced scripts only for the top-ranked topic, and starting a new research run still required the terminal. Max solved this by asking Claude Code to generate hooks and scripts for all five topics and add a button that could start a new run from the dashboard. The revised system produced three scripts for each topic, giving him 15 script options per run.

If you’re building something similar, Max recommends starting with a repetitive task you can describe clearly. Build a working version with your preferred coding tool, use it, and then extend it in response to what you learn.

Follow along with Zero to Agent in 30 Minutes on Radar, or watch the latest episode on YouTube, Spotify, Apple, or wherever you get your podcasts. If you’re an O’Reilly member, you can watch live. Save your seat.