Most "I want to build something with AI" stops at the idea. This intake forces you through 5 questions that scope, sequence, and time-box the work — output is a specific build, the right weekly guide, and a ship-by date. Takes 6 minutes.
This is what filling out the intake actually looks like — questions, sample answers, the routing output at the end. The live form prefills your selections to skip the second time around.
Not "learn AI." Not "build an agent." The actual change in your day or your business. "I'd like 2 hours back from research" or "I want client onboarding to run without me touching it."
Pick all that apply. The intake routes around whatever blockers you flag.
URLs, account access, existing prompts, source documents — anything that means we're not starting from zero.
Constrains the recommendation. If you won't write code, the intake routes to a no-code variant. If you're a Claude Code user, it routes to that.
No date = no shipping. Pick a real date. The intake cuts scope to fit.
Most people who say "I want to build with AI" spend 3 weeks reading content before deciding what to build. The intake compresses that to 6 minutes.
You start with "automate research." Then "build a research agent." Then "build a research platform with auth and a UI." By week 2 you've shipped nothing.
Q5 demands a real ship date. The intake auto-trims your scope to what fits. The platform version goes on the "next quarter" list.
You leave with: a weekly guide link, an estimated build time, a ship-by date, and the supporting skill cheatsheet. Now you just execute.