How I'd Teach a 10 Year Old to Build Agentic Workflows (Claude Code)

In This Issue

• Build agentic workflows in Claude Code that fix themselves (no coding required).
• From £1k workflows to £30k AI agent projects (how one member made the jump).
• The AI behind your automations just got 5x cheaper to run.

New Video:

How I'd Teach a 10 Year Old to Build Agentic Workflows (Claude Code)

Stop building automations step by step. This video walks you through agentic workflows in Claude Code (using the WAT framework) so you can describe what you want in plain English and let the AI figure out the rest.

The mistake most people make: It's tempting to keep adding nodes and debugging connections until the automation "works." But that approach locks you into every technical detail, and one missed edge case can break the whole thing.

How this helps you: You'll know exactly how to set up Claude Code with the WAT framework, connect tools like Firecrawl through MCP, and build workflows that fix themselves when something breaks. That means faster delivery and less time buried in error logs.

Winning Is All We Do

From £1k workflows to £30k AI agent projects

What happened: Joe Scott spent 12 months selling n8n workflows for £1,000 each, competing on price with everyone else on Upwork. He shifted his approach (stopped leading with the technology, started solving bigger problems) and recently closed 2 AI agent projects at £25-30k each for SMEs with 10-250 employees.

The opportunity for you: Businesses spending £30k don't care what tools you use. They care that you can sit down, understand what's actually hurting them, and map out a solution. If you stop selling workflows and start selling outcomes, the same skills you have today can command 25x the price.

This is one of many success stories from AI Automation Society Plus. If you're not a member, you're missing out.

In the News

Flagship AI performance just dropped to one-fifth the price

The news: Anthropic released Claude Sonnet 4.6, which matches or beats its top-tier Opus models on most benchmarks (including coding, computer use, and office tasks) while costing $3/$15 per million tokens instead of $15/$75. In Claude Code testing, users preferred Sonnet 4.6 over the previous version 70% of the time and even preferred it over the more expensive Opus 4.5.

How it affects automation: The AI powering your automations just got dramatically cheaper to run at scale. If you're building agents that make thousands of API calls per day (whether in n8n or Claude Code), your costs just dropped by 80% without sacrificing quality.