Lead a software project with AI on your team.
Step into a simulated workday. Interview a client, brief an AI coding agent, review what it builds and decide when to ship. Every choice changes how the project ends.
See it in action
A workday in 30 seconds.
Follow one project from the client's first email to the final report.
How it works
One desktop, seven tools.
Each project brings a new client and a new problem. The tools, and the habits they build, stay the same.
- 01
Mail
Read the brief: the client's request, a ticket from the people who'll use the system and a note from your manager.
- 02
Meet
Interview the client. They know what's really wrong, but they only answer what you ask.
- 03
Notes
Write the requirements, including how everyone will know the work succeeded.
- 04
Handbook
Look up the ideas the job depends on, explained through what you already know.
- 05
Cody
Brief the AI coding agent. It builds exactly what you describe, gaps and all.
- 06
Review
Check what comes back. Approve it, or send it back with a clear reason.
- 07
Launch
Choose how to release it, then see how it plays out.
Who you work with
AI plays the people you talk to.
The client and Cody respond to what you actually write. Your manager and the users react to the decisions you make, following the project's script.
The client
Answers your questions on a live call. Knows the business but not the technology, and only tells you what you ask.
Cody
An AI coding agent that works from your brief exactly as written: fast, confident and literal.
Your manager
Sets the budget and the deadline, and gets in touch when either starts to slip.
The users
Live with whatever you ship, and are the first to notice what you missed.
Why it works
Scripted, with AI where it counts.
Scripted, so it's fair
Every project follows a written story with real branches. It never drifts, and every learner faces the same situation.
AI where your words matter
AI reads your questions, requirements and brief: the moments where how you put something changes what happens.
Consequences, not quizzes
There are no right-answer buttons. What you miss or wave through shows up later, in the code, the release and the report.
Every choice counts
See how you worked with AI.
Skip a question and Cody builds the wrong thing. Approve too quickly and the problem ships. At the end you get a report based on what you actually did, scored on the four AI Fluency competencies.
Success
The client's problem is solved, and you can prove it.
Partly there
It mostly works, but something slipped through and cost time.
It went wrong
A gap in your brief reached real users.
Project report
How you worked with AI
Delegation
strongWhat you handed to AI, and what you kept for yourself.
Description
mixedHow clearly you described what you needed.
Discernment
strongWhether you spotted what it got wrong.
Diligence
weakWhether you took responsibility for what shipped.
Ready for your first client?
Your progress is saved as you go, so you can stop and pick up later.
Browse projects