← AI & Tech
The Calum Johnson Show
Open Claw Runs My $11M Business: How To Get Rich In The New Era Of AI Agents (Even As A Beginner!)
4 min read25 Apr 2026Worth watching
TL;DR
Riley Brown joins Calum Johnson to explain what AI agents are, how they work, and how to build your first one using OpenClaw. The episode covers the shift from vibe coding to autonomous AI agents, live demos a YouTube growth agent setup, and argues that right now is the best window to capitalize on this technology before it becomes commoditized.
Key points
1
An AI agent is an AI model that runs tools in a loop — it decides which tools to use, feeds results back to itself, and keeps iterating until the task is done, unlike a one-shot ChatGPT prompt.
2
Riley Brown runs 7-8 personal AI agents daily and his company spends six figures per month on AI tokens — agents handle customer support, marketing outreach, brand deal negotiation, and executive assistant tasks.
3
OpenClaw's new 'heartbeat' feature lets agents wake up every 30 minutes and decide autonomously whether to take action — a major leap toward truly autonomous, always-on AI workers.
4
A beginner can get started for as little as $20-50 per month using Claude Co-work or a hosted OpenClaw instance on a service like Chorus — no Mac Mini required.
5
The sweet spot for agent skills is 7-20 maximum; adding too many skills (above 20) causes the agent to use the wrong skill more frequently and degrades performance significantly.
Actionable insights
→
Start by giving your agent a single, specific goal rather than loading it with dozens of skills — treat it like hiring an employee who is great at two things, not a generalist who claims to do everything.
→
Use 'context engineering' not just prompting: feed your agent your last 10 video transcripts, your goals, your tools, and your style so it has the right context to produce useful outputs from the start.
→
Set up a daily cron job (scheduled automation) to have an agent analyze your email, YouTube comments, and customer data each morning and produce a one-page insight report — Riley describes this as one of the highest-ROI workflows he runs.
→
When your agent makes a mistake, edit the relevant skill markdown file directly to include the correction — this is how you improve agent performance over time without starting from scratch.
→
To monetize now: learn to build and configure AI agent workflows for businesses, since the friction in setting them up is exactly what companies are willing to pay for before plug-and-play solutions arrive.
Notable quotes
“I actually have my AI agent 24/7 negotiating all of my brand deals — I actually don't even need to be in there.”
“As soon as it is super easy, as soon as you can just go to a site and click, yep, hire a new employee and then boom, you have this employee — all the value kind of goes to the companies who create those AI agents or the model providers.”
“It is not prompt engineering, it is context engineering.”
Worth watching?
✅
Worth watching the full video?
Watch if you want to see a live hands-on OpenClaw agent being built from scratch — the core concepts and setup steps are all captured here, but the real-time demo and on-screen walkthroughs add useful visual context you will miss by only reading.
Saved you some time? The creator still deserves a like.
Want this for your own channels?
Add the channels you follow. Every new video summarised and in your inbox the moment it drops. From £4/month.
Try it free