This repository serves as a detailed case study for the AI-assisted workflow I developed to produce the "Before AGI" daily podcast. While the core operational scripts are proprietary, this document outlines the process, tools, and results to demonstrate my capabilities in AI content systems, workflow automation, and strategic content repurposing.
Turning Written Content into a Daily Podcast
This project demonstrates a robust, AI-assisted system designed to solve a critical business problem: "content debt," where valuable written assets are underutilized. I designed and executed a workflow to transform sourced text into a professionally produced daily podcast. The proof of concept, the "Before AGI" podcast, has successfully released over 240 episodes since October 2024, proving the system's consistency, quality, and scalability.
Most businesses invest heavily in creating a library of high-value written content. While valuable, this content is "silent"—inaccessible to the significant audience that prefers audio. The challenge was to create a system that could reliably repurpose this text into engaging audio content on a consistent schedule without requiring hours of daily manual production.
To solve this, I created the "Before AGI" podcast—a daily show exploring AI developments on the journey toward Artificial General Intelligence. This project was not just about creating a podcast; it was about engineering the underlying system as a repeatable, scalable model for any business.
The workflow's success hinges on a strategic, human-in-the-loop process:
- Source Curation & Grounding: High-quality sources (articles, research) are curated to provide a factual foundation for each episode's topic.
- AI-Assisted Scripting (NotebookLM): I use Google's NotebookLM to ground the AI in the provided sources, then leverage it to summarize key points and structure a coherent podcast script.
- Professional Voice Synthesis: The script is rendered using high-quality AI voices selected to match the show's authoritative tone.
- Human-Led Audio Engineering: This is the critical quality-control step. I personally perform a full editing and mastering pass in DaVinci Resolve, mixing in music and ensuring the final track meets professional loudness standards.
- Consistency: Over 240 episodes produced and released on a consistent schedule.
- Multi-Platform Distribution: Successfully published daily across Spotify, YouTube, and Substack.
- Proven System: The creation of a reliable, scalable content engine adaptable to any business niche.
I used this exact process to create my daily podcast. Listen to a sample episode below to experience the final audio quality.
➡️ Listen to "Before AGI" on Spotify or YouTube
- Core AI Tool: Google's NotebookLM (Summarization, Voice Generation)
- Research: ChatGPT Deep Research, Perplexity Deep Research, Gemini Deep Research
- Audio Engineering: DaVinci Resolve (Editing, Mixing, Mastering)
- Distribution: Spotify for Podcasters, YouTube, Substack