Adapting one course for different levels with AI

Almost every course has a mixed audience. Some members are complete beginners who need every term explained. Others have years of experience and joined for the finer points. Teach to the middle and both groups drift away: the beginners feel lost, the experienced members feel patronized, and your completion rates suffer.

Building separate beginner and advanced courses is one answer, but it doubles the work of creating and maintaining everything. Teachers call the alternative differentiation: one course, with different support and different challenges depending on where a learner starts. AI makes this far more practical, because the extra layers are exactly the kind of structured drafting it handles well. You keep control of the core teaching and check every layer it adds.

Keep one core, vary the support

The principle is simple. Every learner works toward the same learning objectives through the same core lessons. What changes is what surrounds those lessons:

• Foundation layer for beginners: definitions of key terms, a worked example, a checklist and a reminder of anything they should know first.
• Core lesson for everyone: your main teaching, unchanged.
• Stretch layer for experienced members: a harder task, a messy real-world case, edge cases and common expert mistakes.

Kenji runs a photography membership. His lesson on composition works for everyone at the core. Beginners get a short foundation note explaining focal points and the idea of leading lines with labeled examples. Experienced members get a stretch task: take three photos that deliberately break a composition rule, and explain why each still works.

Because the core stays the same, you maintain one course rather than two, and every learner can move between layers as they grow.

Find where levels split

Before generating anything, find the points in a lesson where beginners and experienced learners part company. AI is good at reading a lesson and spotting assumed knowledge, which is often invisible to you as the expert.

Here is a lesson from my course for [describe your audience, including the range of experience]. [paste lesson text or transcript]

First, list every term, skill or piece of background knowledge the lesson assumes without explaining. Second, identify the three places a complete beginner is most likely to get lost, and why. Third, identify the parts an experienced learner would find too basic, and suggest what would challenge them instead. Do not rewrite the lesson.

Read the results against your own experience of teaching this topic. The assumed-knowledge list is often the most valuable part. It tells you exactly what belongs in the foundation layer.

Generate the layers

With the split points identified, ask for the two extra layers. Keep the core out of it: you are adding around your teaching, not replacing it.

Using the lesson and the analysis above, draft two additions. Foundation notes: plain-language definitions of the assumed terms, one worked example, and a short checklist for completing the lesson's practice activity. Stretch material: one harder task that uses the same skill in a less tidy situation, two edge cases experienced practitioners get wrong, and one question to think about. Match the tone of my lesson. Mark anything you are not certain is accurate.

Stretch material deserves the closest review. When AI reaches for advanced content, it is more likely to invent techniques, overstate a rule or wander beyond your expertise. If you would not teach it confidently yourself, leave it out. For generating the practice itself at several levels, see personalized practice exercises generated with AI.

Help learners find their level

Layers only help if learners know which to use. Two approaches work well together.

A short self-check at the start. A handful of questions, each tied to a key skill, with a simple suggestion at the end: start with the foundation notes, go straight to the core, or try the stretch tasks. AI can draft these questions from your objectives; you should check that each one really does separate the levels. For the non-AI side of this, see prerequisites and placement.

Free choice with clear labels. Describe each layer by what it offers rather than who it is for. “New to this? Start with the key terms” is kinder than “For beginners.” Nobody should feel labeled, and members should always be able to dip into any layer.

Review before anything goes live

Go through every layer with these questions:

• Are the definitions accurate and consistent with how you use the terms in the core lesson?
• Does the worked example follow your method, or a different one the AI prefers?
• Is the stretch task genuinely harder, or just longer?
• Does anything contradict the core lesson?
• Does it sound like you?

Consistency matters more than it seems. If the foundation notes explain an idea one way and your video explains it another, beginners will be more confused than before. When you find a mismatch, fix the layer, not your core teaching, unless the AI has spotted a real weakness in the lesson.

Watch how members use the layers after launch. If almost everyone opens the foundation notes, your core lesson may assume too much. If nobody touches the stretch tasks, they may be too hard to find or not challenging in the right way. Clear learning objectives make this easier to judge, because every layer should point back to them.

First steps

1. Pick the lesson where you hear the widest range of questions.
2. Run the split-point prompt and check the assumed-knowledge list.
3. Generate foundation and stretch layers, and review them against your core lesson.
4. Cut any stretch content you would not teach confidently yourself.
5. Add a short self-check and friendly labels so learners find the right layer.
6. Watch how the layers are used and adjust before extending the approach to other lessons.