Four stages, all pointed at the same thing: a working product of your own. Here's what you'll be able to do by the end of each one. Some lessons are pure technique, learned on their own; others you apply straight to your product.
A one-to-one where we find how you use AI now and where the gaps are, then plot your path through the course. You leave knowing your starting point.
How these models actually work: what they know, why they make things up, and when to reach for a lighter or heavier one. The mental model everything else rests on.
Fourteen practical techniques for reliably better results, and why each one works. Repeatable, and yours from here on.
When a normal chat isn't enough and an agent is: the task you'd otherwise keep re-explaining. You build your first one, kept deliberately simple.
Give your AI one source of truth. Connect your tools to Notion so every agent reads and writes the same place, with your brand, voice and content all living there.
Take a process you already use and rebuild it around what AI genuinely makes better. You'll validate your product idea and build an agent that ranks your riskiest assumptions first.
Turn flat, static personas into ones you can actually interview and pressure-test your thinking against. Still a best guess, only now with a voice.
Map your product before you design a single screen. Story mapping defines the features, and hands AI the context it needs to help you design them.
The basics of Claude Design, what it's good for and what it isn't, walked through in depth.
Create a visual identity, captured in a one-sheet you'll apply to everything: a logotype and mark, a type ramp, and a colour palette.
Build a tone-of-voice guide for your brand and wire it into Notion so Claude Design can read it. This one leans on the prompt and agent skills you've already learned.
Write the copy in Claude chat, build the page in Claude Design. A marketing landing page that frames your idea before you design the product itself.
Use Claude Design the way you'd have used Figma: multiple UI ideas side by side on a wide canvas, worked up from your story map.
Take your static designs and add real interactivity, using JavaScript libraries and the open-source ecosystem to make the most of them.
Build a design system for your product, extrapolated from your UI experiments. Choose your path: the Claude ecosystem, or connect Figma if that's where your team lives.
Set up Claude Code, GitHub and Vercel, and learn to run and test your build locally. Your dev studio, ready to go.
Take your Claude Design files into Claude Code and start turning them into a working build.
Delegate real build tasks to Claude Code and manage an AI that writes and ships code, keeping your judgement over what it produces. You end with a finished proof of concept.
Every lesson comes with a workbook you apply to your own product. On the full programme you also get a live consultative session at each stage, plus my personal review of your work as you build.
Bring one idea and leave with a working product of your own. After seventeen lessons, you'll know AI.