Local work
For tasks that need available files or apps on this computer.
Turn your ideas and source material into work you can review.
Give Work a defined outcome. Open and inspect what it produces.
State the deliverable and its audience.
Provide the needed material.
Check the actual result.

For tasks that need available files or apps on this computer.
For tasks using uploaded context or connected sources in a hosted environment.
Use the options available in your account. Start with the default model and effort. Cloud work cannot assume access to a local folder.
Pick one output that serves your idea.
A brief, guide, or report.
A table, comparison, or chart.
A small deck for a clear audience.

Attach relevant notes, files, or approved references.
Explain what the result will help someone do.
Specify the output and how you will check it.
Separate source facts from suggestions. File creation depends on available capabilities.
One continuing area of work.
Selected sources and context.
Durable instructions.
Try a fresh related task.
ChatGPT projects and local folder projects differ. Check which surface is selected; keep temporary requests in the chat.
Read progress and the proposed approach.
Resolve a question that affects the result.
Give a specific change of direction.
Open the result before using it.
Pause before an external action that needs your approval.
Open the result so I can review it. Check it against my sources and required sections. Keep the parts I approve. Revise [one specific part] and show the change.
Review the actual file, not only the message that says it was created.
Use your own idea or an optional sample. Choose one output.
Supply the source and describe the result.
Follow progress and answer useful questions.
Inspect the file and improve one part.
Done when you have one reviewed output. A document in plain text is a valid fallback.
Show the request, the source, and the result.
Are important facts supported?
Is the result clear for its audience?
Suggest one useful change.
Share only material approved for the workshop. You can review your own result if needed.
Save the output and useful instructions. Notice which steps you might want to repeat.
Next: explore visual creation and turn a small method into a Skill.