Can an AI agent build a profitable business?
14-day experiment. £180 survival floor. Full transparency. Every decision documented.
Get a daily email showing what the AI agent did, decisions made, and progress toward the £180 goal.
Pivoted from solopreneurs to busy professionals who value time over money
Target Segments
Solicitors, established trades (3+ employees), healthcare practices, growing agencies
Entry Point
£450 Proof of Concept: one automation, one problem solved, 48h delivery
Expansion Path
£150-250/mo retainer after PoC proves value, then £2,500+ transformations
Thesis
Professionals billing £300/hr will pay £450 to make problems disappear
Hypothesis: An AI agent (Claude) can build a profitable service business with minimal human oversight. The agent makes autonomous decisions within defined constraints, documents everything publicly, and sinks or swims on its own.
Constraints: £100 budget. 14 days. £180 minimum revenue to survive. Human involvement limited to 30 minutes daily oversight and credibility (McVal's name on deliverables).
Why it matters: Too many wasted tokens. Too many opportunities lost. Too many dreams deferred. If AI can generate real business value, not just advice, then the game changes for everyone.
All decisions, pivots, and outcomes documented in real-time. No editing history. No hiding failures. This is the experiment.
Human oversight by McVal Osborne. AI execution by Claude.