Making $$$ with OpenClaw

2mo ago

I sit down with Nick Vasilescu, founder of Orgo, to break down exactly how people are turning OpenClaw — the open-source computer use agent — into a real revenue stream. Nick walks me through live demos of deploying OpenClaw for business clients, shows how sub-agents and parallelization multiply output, and shares his design-thinking framework for identifying and automating high-value workflows. We even build a TikTok trend-hunting agent from scratch during the episode to prove how fast you can go from idea to working prototype. Timestamps 00:00 – Intro 02:50 – Getting Set Up with OpenClaw 05:02 – Finding the Wedge: Automating Real Business Outcomes 07:39 – The Upwork Hack: Finding Paid Automation Jobs 09:41 – Andreessen Horowitz on Computer Use Agents 11:01 – Setting Up a Client Workspace in Minutes 12:41 – Design Thinking: Mapping Value vs. Effort 15:23 – Using OpenClaw to Prioritize Automations 17:57 – Building Automation Pipelines with Claude Code 19:33 – Sub-Agents vs. Tasks vs. Skills 23:22 – Automation Possibilities are huge 24:54 – Live Build: TikTok Trend Hunter from Idea Browser 32:09 – Start with an MVP Skill, Then Iterate 32:41 – Architecture of the TikTok Agent Script 36:59 – The Arbitrage Opportunity: Most Businesses Still Need Help 40:30 – Agents Are the New SaaS 42:42 – Demoing TikTok Trend Hunter 44:11 – Building Assets & the Abundance AI Will Bring 47:58 – Closing Advice: Get Your Hands Dirty Links Mentioned: Orgo: https://startup-ideas-pod.link/orgo Key Points * OpenClaw is more than a personal assistant — it is a deployable business tool that can automate end-to-end workflows for paying clients. * The fastest path to revenue is finding automation jobs on Upwork (RPA, desktop automation, workflow building) and fulfilling them with OpenClaw and Claude Code. * Sub-agents allow your main OpenClaw instance to delegate specialized tasks, keeping the orchestrator free and multiplying throughput through parallelization. * A design-thinking approach — mapping automation opportunities by value vs. effort — is essential before building anything. * Verticalizing computer use agents for a specific industry (manufacturing, real estate, distributorships) is the major startup opportunity Andreessen Horowitz is calling out. * Always start by building a lightweight MVP skill, test it, debug, and iterate before scaling. Numbered Section Summaries 1) OpenClaw Setup and Deployment Options Nick demonstrates how easy it is to install OpenClaw on a virtual machine using Orgo, though he makes clear you can use Manus, Kimi, a Mac Mini, or any setup you prefer. He spins up a workspace for me in under a minute — just a curl command in the terminal and it is ready. The point: the barrier to entry is nearly zero. 2) The Wedge: Finding Business Automation Opportunities The viral demos on Twitter are fun but toyish. The real money is in identifying a specific workflow inside a business — like downloading product data from a legacy platform and uploading it into a Zoho CRM — and automating that end to end. Nick calls this the "wedge" and it is the foundation of the entire business model. 3) Sub-Agents and Parallelization OpenClaw can spawn up to eight sub-agents, each with its own computer. Nick shows two parallelization strategies: splitting one task across multiple agents, or running the same task across multiple instances for volume. He spawned sub-agents to scrape Upwork jobs, build demo proposals, and pick the best one — all automatically. 4) The Upwork Hack If you have zero clients, Upwork is the starting point. People are posting $500–$5,000 jobs right now asking for AI workflow automation, desktop automation, and RPA replacements. Nick's approach: find the job, give the context to OpenClaw or Claude Code, build a demo, and submit the proposal. It is a lead generation machine. 5) Design Thinking for Automation Before touching any code, Nick maps every potential automation on two axes: value created and effort/cost/ti