Levels 0-1 | 3 lessons
Start with the right surface
Pick the job before you pick the feature.
Most ChatGPT misses start before the prompt. This module teaches the first call: plain chat, search, file upload, Project, GPT, image input, voice, task, or agent mode. You learn to ask for proof before acting on anything that depends on current facts or private context.
Proof:By Lesson 3, you can ask one current question with dated official sources and mark which claims are proven, unclear, or plan-dependent.
New ChatGPT usersTeams tired of generic answersAnyone who needs source checks
Start Module 1Level 2 | 4 lessons
Projects, files, and memory
Put context where it belongs.
This module splits context into current chat, uploaded files, Project files, account memory, and custom instructions. You learn where each type belongs, what to avoid saving, and why scheduled tasks cannot depend on Project files.
Proof:You finish with one Project setup plan that names files, instructions, memory items, and the items that should stay out.
People rebuilding context every chatProject ownersPrivacy-conscious teams
Start Module 2Level 3 | 3 lessons
GPTs and app connections
Reuse a process without hiding the source boundary.
You build a GPT only when repeat use is worth the setup. Then you decide when an app connection belongs in the workflow, what permission it needs, and how to test one no-connection path before a team depends on it.
Proof:You finish with a GPT plan, a source boundary, and one fallback route for accounts without the needed app connection.
Process buildersTeam leadsPeople sharing repeat prompts
Start Module 3Level 4 | 4 lessons
Images, voice, data, and drafts
Use the right work surface for the artifact in front of you.
This module covers image inputs, image creation requests, voice, Record, table cleanup, chart prompts, data summaries, writing blocks, and code blocks. The rule stays the same: name the source, the finish line, and what must be preserved.
Proof:You finish with one reviewable output from a file, a voice note, image work, or an editable draft, with checks attached.
Writers and analystsVisual thinkersPeople moving between meetings and docs
Start Module 4Level 5 | 3 lessons
Tasks, agent mode, and Sites handoff
Know when ChatGPT should act, and when Codex should take over.
You schedule useful tasks, supervise agent mode, and classify site requests correctly. Sites are taught as a workspace Codex handoff, not a normal ChatGPT chat feature, because current availability depends on workspace eligibility.
Proof:You finish with one scheduled task spec, one agent-mode supervision checklist, and one site handoff brief.
OperatorsPeople running recurring workTeams preparing Codex handoffs
Start Module 5