Key takeaways
- Generate an editable course draft from description, track the AI task, and review the result safely.
Open this area in SubSchool
This workflow turns description into an editable course draft. AI output must be checked against the source before learners receive it.
Step 1. Open Courses
Click Courses in the teacher menu and confirm that the intended content language is selected.
Step 2. Choose the description method
Click Create course, then choose AI course from description.
Step 3. Complete the form
Describe the audience, outcomes, topic, and draft program in concrete language.
Select the subject, language, course type, grades, and student goal. Then describe the target audience, measurable outcomes, topic boundaries, and draft program in separate fields. Do not write only a title: include prerequisites, depth, exclusions, and the intended order. A specific brief produces a more useful outline.
Step 4. Finish the detailed brief
Write outcomes as observable skills, keep the topic field focused, and use the draft-program field for the desired module sequence. Avoid contradictory instructions and promises of guaranteed results.
Step 5. Submit and watch the quick status
Click the purple generation button once. A background task is created, so you can close the dialog and continue working. Open the bell/task menu in the header to see queued and processing status. When email notifications are enabled, SubSchool also emails you after completion or failure.
Step 6. Open the complete task history
Go to AI tasks to see progress, timestamps, the task type, the generated object, and any error. Do not submit the same source again merely because the dialog closed.
Step 7. Review the generated draft
Open the generated course. Compare titles, claims, ordering, lesson descriptions, and assignments with the source. Correct unsupported facts, confirm rights and audience level, set pricing and access, and preview the learner view before publishing.
Sources and methodology
The workflow, labels, validation rules, and resulting states were checked against the live test environment. Screenshots were captured by Playwright at 3840×2160 and reviewed at full resolution.