When Podcraft Media Lab adopted AI tools, the goal wasn’t to reduce headcount or cut creative corners. It was to buy back time. Greg Soros, the podcaster and entrepreneur behind the Austin-based studio, designed his company’s approach to automation with that boundary clearly in mind.
Soros’s professional background Berklee College of Music training, followed by senior roles at established media organizations gave him a working understanding of both the technical and creative sides of audio production. That experience informs a skepticism about solutions that promise to automate the parts of podcasting that actually matter to listeners.
“We use artificial intelligence to handle transcription, initial audio cleanup, and metadata organization—but the creative decisions remain firmly in human hands,” he explains.
The Business Case for Human Judgment
Podcraft Media Lab’s model has produced concrete results. Production timelines are down approximately 35%, giving the team more capacity to invest in the narrative and sound design work that defines premium audio content. Client retention has held above 95%, and referrals generate 40% of new business an unusually high share that points to genuine client satisfaction rather than just price competitiveness.
Machine learning tools at Podcraft Media Lab do handle analytical tasks: identifying where ads will land well within an episode, tracking engagement patterns across a show’s catalog. But those are inputs into human decisions, not substitutes for them.
Soros has watched competitors automate more aggressively and then struggle to understand why their content stopped connecting. His diagnosis is simple: “You can’t algorithm your way to authenticity.” For a Greg Soros-led podcaster operation, that principle isn’t just a talking point. The retention numbers suggest it’s actually working. See related link for more information.
Check out his page on https://x.com/gregsorospod