Authoring studio
How AI authoring works in Nucleus OS, from brief and outline to generated lessons, inline editing, regeneration, and version history.
The authoring studio is where courses are made. You describe the outcome, attach sources, and the AI drafts a structured course with lessons, examples, assessments, and media. Your experts then spend their time on judgment instead of formatting.
The authoring flow
Brief
Describe the audience, the outcome, the length, and how mastery will be shown. You can start from a prompt alone, but grounding the course in your own sources gives much better results.
Outline
Nucleus OS proposes modules and lessons. Rework the structure conversationally until it is right. Nothing is generated until you approve.
Generate
Lessons, assessments, and media are generated together so they stay consistent. Every generated claim keeps a citation to the source passage it came from.
Refine
Edit any block inline, or regenerate a single block ("make this shorter", "use a retail example", "turn this into a scenario") without rebuilding the course.
Verify and publish
Run review gates, get approval, then publish in Nucleus OS or export to SCORM.
What you can generate
| Format | Use it for |
|---|---|
| Lessons | Structured explanations with examples, broken into short sections |
| Interactive cards | Bite-sized concepts learners can flip, sort, or reveal |
| Scenarios | Branching decisions based on realistic situations |
| Quizzes and assessments | Checks for understanding and graded assessments; see Assessments |
| Narrated video and AI presenters | Explainers and announcements generated from the same source material |
| Role-play | Practice conversations with an AI partner; see Role-play practice |
Version history
Every save creates a version. Compare any two versions side by side, see who changed what, and restore an earlier version. Completion records reference the version a learner completed, which matters for compliance.
Change once, update everywhere
When a policy, product, or regulation changes, update the source. Nucleus OS runs an impact analysis across every course that cites it and proposes the edits, including in every translation. Approve the changes and, if you choose, learners are re-certified automatically.
Credits
AI generation draws from your workspace's credit pool. You see an estimate before a large generation starts. See Billing and credits.
Tips for better drafts
- Keep sections under five minutes of learner time.
- Name the audience precisely. "New support agents in Lisbon" beats "employees".
- Ask for scenarios and worked examples, not just explanations.
- Attach the real source, not a summary of it.