The 30-Day Challenge.

Stop watching.
Start shipping.

Four working automations and one deployed web app with a live URL — in 30 days. Done with you, not for you. April sold out in 10 days.

CohortMay 2026
Seats20 only
Calls4 weekly + Skool
StartsThursday May 7
Apply for the May Cohort
Within 30 days, you'll have —

Four builds.
One live URL.

  • One deployed web app on a real URL
  • Four working automations for your own business
  • Four live group calls — diagnose, build, unblock
  • Every walkthrough I've done myself
  • Private Skool group with the cohort
One payment · No upsells
From Tom

The watchers fall behind.

You've seen the demos. You've bookmarked the threads. You've probably opened Claude Code at least once. Then something didn't work, the install had a hiccup, the first build got messy, you closed the tab and told yourself you'd come back to it.

That was three weeks ago. Meanwhile every week someone posts another thing Claude Code can do that you still can't. The space is moving. You're not.

You don't have a knowledge problem. You have a direction problem.

Watching someone else build something is a completely different skill from building it yourself. The gap between the two is small, but you can't cross it on your own by watching more videos. You cross it by actually building something, with someone there when it breaks. The Challenge fixes that.

Three phases. Thirty days. One deployed app.

Diagnose. Build. Ship.

01
Week 1 · Investigate

Pick the thing.

Most automation projects fail because people automate the wrong thing. You identify the highest-value target and build the brief you'll hand to Claude Code. This is the thinking that makes everything else work.

02
Weeks 2–3 · Build

Run the thing.

Your first working script. A PDF invoice goes in, vendor name and amount come out structured into Airtable, and you didn't write a line of it. Claude Code did. Then three more — different problems, different stacks. By the end you've stopped guessing and started shipping.

03
Week 4 · Ship

Deploy the thing.

A working script on a laptop isn't an agent. Ship takes your build to a deployed web app with a real URL and a real schedule. Something running on the internet that you could actually charge for. Most courses stop before this. This is where the proof lives.

Who this is for.

This is for you if

  • You run a real business and can already picture five repetitive tasks AI should be eating
  • You've heard about Claude Code, dabbled, maybe tried two things, but haven't shipped anything
  • You want a trusted advisor in your back pocket — not another curriculum
  • You'd rather ship something ugly this month than read about it for six

This is not for you if

  • You're a curious tinkerer without a real business yet
  • Your first message is "what's the cheapest tier"
  • You've taken three other AI courses this year and shipped nothing from any of them
  • You want to spin up a 30-day AI agency and get rich quick

Plus four bonuses most courses skip.

The unfair-advantage stack. Each one alone is worth the price of the Challenge. Together, they're the system I use every day.

Bonus 01 · $300

Build Your Own Claude Code Skills.

Most people use Claude Code exactly as it comes. You'll learn to build your own skills that extend what it can do, so it works the way your business needs.

Bonus 02 · $200

Give Claude Code Memory.

Memory and context across sessions, so you stop re-explaining your project every time you open a new build. This alone is worth the price of the Challenge.

Bonus 03 · $400

Obsidian Second Brain Setup.

A knowledge system connected to your builds. Notes, processes, and decisions all in one place. The same setup I use every day.

Bonus 04 · $500

Give Claude Code Superpowers.

A repeatable framework for building apps. Install it, run it from day one, and stop second-guessing what to build next.

Bonuses alone are worth $1,400.

From inside the cohort

What architects
are saying.

Pulled straight from the Skool group and the DMs. No edits.

Andrew, on the Claude Code Challenge
Andrew, on the Claude Code Challenge
Daniel, on the Claude Code Challenge
Doug, on the Claude Code Challenge

This is safe to try.

Show up to all four calls, complete the material, and still can't build a working agent for your business or role? You get a full refund. No argument.

If you've been sitting on the fence about whether Claude Code is worth your time, this removes the risk of finding out.

Here's how we'll do it.

You don't learn about Claude Code. You learn inside it. The tool teaches you the tool, so by the time you finish the material, you've already been using it every day.

Two to four hours of course material across 30 days. One live call per week where we build, go deeper, and unblock anything stuck. Two sessions to pick from: Thursdays at 12:00 p.m. UK and Thursdays at 8:00 p.m. EST.

The Investment

$500.

One payment. Everything included.

For context: my one-on-one rate is $900 USD an hour.

For $500, you get four group calls a month where I'm live with you. And if the time zones line up, you can sit in on both Thursday sessions and make it up to eight live calls over the 30 days. That's the same amount of direct access my paying mentorship students get.

Most architects earn this back in time saved inside the first week.
Tom Crawshaw
A bit about me

Tom Crawshaw.

Eight years building automations. Over $25 million in revenue generated for clients using AI and workflow systems across private aviation, construction, marketing agencies, and e-commerce.

I don't sell information — YouTube has plenty. What it doesn't have is someone looking at your specific setup, telling you exactly what's wrong, and getting you unstuck in the same session.

How it works

Here's what happens next.

  1. 01

    Click the button.

    Hit the Join the Challenge button below.

  2. 02

    Pay the $500.

    No forms. No application. No hoops.

  3. 03

    You're in.

    Confirmation email lands straight away. Access to the private Skool group goes live the day before we kick off on Thursday May 7.

Join the Challenge — $500

Pay → confirmation → cohort access. Takes 90 seconds.

"You don't need a course. You need one good Wednesday."

Twenty seats.
Closing May 6.

20 people per cohort so everyone gets attention on the calls. April sold out in 10 days. Applications close Wednesday, May 6th, or when the last spot is gone.

Join the Challenge

No application form. No hoops. Pay → confirmation → cohort access.

P.S. The next 30 days are happening either way. The only question is whether you come out the other side with four working automations and a deployed app, or another month of bookmarked demos. The cohort starts Thursday May 7th. Doors close the day before.