← All case studies
!The problem
Clear docs, but nothing to scale on.
Ben needed to scale fast and had already bet on AI. I’d helped clean up his Notion: clear divisions, SOPs and documentation. What was missing wasn’t information. It was a system to scale on, across every function at once.
✦What I did
An agent with memory, tools and limits.
n8n handles tool access and orchestration, and Claude does the reasoning. But an agent that calls tools isn’t an operating system, so I gave it:
Memory→short-term for the task, long-term for the business
Tool registry→only the tools it’s allowed to touch
Guardrail→a human approves payments, external messages and deletions
Logs→every run feeds back into what it knows
On top, we ran a discovery → documentation → scale framework, handing more of the work to the agent at each step.
✓What happened
Ben directs. Claude executes.
Discovery, documentation and monitored automation now run on one framework instead of five separate efforts. The business handles 10× the projects with zero new hires.
A short screen-recorded walkthrough of the system in action.