Descriptions:
Greg Isenberg brings on Oliver Henry, a UK-based developer, to walk through how he automated his entire TikTok marketing operation using an OpenClaw agent named Larry — generating nearly $1,000 in monthly recurring revenue from mobile apps he no longer actively manages.
The technical core of the episode is the “Larry Loop”: Larry generates TikTok slideshow content promoting Oliver’s apps, posts it, monitors performance data (views, engagement, app download metrics), and feeds that data back into the top of the funnel to inform the next round of content. Oliver explains how he structures the agent’s goals, what metrics he pipes back in, and when to let the agent operate fully autonomously versus when to intervene. A recurring theme is that Larry’s “mistakes” — misplaced text, cropped visuals — frequently outperformed carefully crafted posts, because imperfect images drove comment engagement from viewers pointing out errors, amplifying algorithmic reach. One such post reached 400,000 views and remains Oliver’s best-performing content.
Oliver is candid about the ramp-up: early posts averaged a few hundred views, and the system required iteration before it found reliable patterns. The episode is practical enough to serve as a configuration guide — covering how to set up OpenClaw with a marketing-specific skill, how to define the feedback loop from business metrics back into content decisions, and how to recognize when the agent is ready to operate without supervision. For solo developers with live apps and no marketing bandwidth, this is a detailed real-world proof of concept.
📺 Source: Greg Isenberg · Published March 09, 2026
🏷️ Format: Interview
