How to Outline Your Course Using Claude

Before you open Claude, collect everything you've already created that relates to your course topic. Blog posts, newsletter issues, podcast transcripts, workshop handouts, client intake notes, slide decks — anything where you've explained, taught, or answered questions about this subject. Copy these into text files or a single document. Don't worry about organizing them yet. The goal is to have all your raw material ready to paste in.

If you're working from audio or video recordings, transcribe them first. Tools like Otter.ai or Descript produce serviceable transcripts. Rough transcripts are fine — Claude handles informal language and filler words without trouble.

Write one sentence that completes this prompt: "By the end of this course, students will be able to ___." This is your transformation promise, and it's the single most important input for the AI. Without it, Claude will organize your material topically — which gives you a reference guide, not a course. A course is a journey from point A to point B. The outcome sentence defines point B.

Be specific. "Students will understand watercolor techniques" is too broad. "Students will be able to paint a simple landscape using three watercolor techniques — wet-on-wet, dry brush, and glazing" gives Claude a concrete destination to build toward.

Start a new Claude conversation. Paste in your source material — or upload files if you're using the web interface. Then provide your framing. Here's a prompt structure that works well:

I'm creating an online course. Here's my source material
[already pasted above].

Course outcome: By the end of this course, students will be
able to [your outcome].

My students are [brief description: experience level, goals,
constraints].

Please analyze my source material and propose a course outline
with 4-6 modules. For each module, suggest 3-5 lessons. For
each lesson, include:
- The key concept
- What the student practices or applies
- Estimated time (in minutes)

Flag any gaps where my source material doesn't cover something
students would need.

The "flag any gaps" instruction is important. One of the most useful things Claude does is identify what's missing from your material. You may have covered a topic extensively in blog posts but never explained a prerequisite skill that your readers already had. Students won't have that background, so the course needs to address it.

Claude will return a structured outline. Read it not as a finished plan but as a first draft to react to. Ask yourself: Does the sequence make sense? Would a student finishing Module 2 be ready for Module 3? Are there lessons that feel redundant or that try to cover too much in one sitting?

Pay special attention to the early modules. AI tools tend to front-load foundational concepts because they're organizing by logical dependency. But students who are excited about a topic don't want to sit through three modules of theory before they get to do something. If Claude puts all the "foundations" first, ask it to restructure so students get a quick win in Module 1.

Once you have a rough structure you're happy with, zoom in. Ask Claude to expand individual modules or rework sections that don't feel right. This is where the conversation format works well — you're refining collaboratively rather than trying to get the perfect output on the first prompt.

Module 3 feels too dense. Can you split it into two modules?
The first should cover [topic A] and the second [topic B].
Keep the same lesson format.

You can also ask Claude to generate lesson-level outlines — the talking points, activities, and discussion questions for each lesson — once the module structure is settled. Work top-down: course shape first, then module details, then lesson scripts.

Activities are where learning actually happens. Once your lesson structure is solid, ask Claude to suggest a practice activity for each lesson. Be specific about your format:

For each lesson in Module 2, suggest one practice activity
that students can complete in 15-20 minutes. My students are
working professionals, so activities should use materials
they already have. Also suggest one discussion question per
lesson that would work in a small group forum.

Discussion prompts are especially worth generating. Good prompts connect the lesson content to the student's own situation — "Share one example from your practice where [concept] applies" rather than "What did you think of this lesson?" Claude is good at generating prompts that invite specific, experience-based responses.

Before you consider the outline finished, ask Claude to evaluate it from the student's perspective:

Review this outline as if you were a student taking the
course. Where would you feel confused, bored, or
overwhelmed? Where is the pacing too fast or too slow?
Are there any places where I'm assuming knowledge
the student might not have?

This prompt produces surprisingly useful feedback. Claude will often catch pacing issues — two complex lessons back-to-back, or a stretch of passive content with no practice opportunity. It won't catch everything a real student would struggle with, but it's a useful sanity check before you invest time recording.

Once you're satisfied with the structure, ask Claude to format the full outline in a clean, consistent format you can paste into your course platform or project management tool. A simple Markdown outline with modules as H2s and lessons as H3s works well. Include the estimated duration per lesson and any notes about required materials or prerequisites.

Save the entire Claude conversation, too. It contains your reasoning about why certain lessons are sequenced the way they are and what gaps you identified. That context is valuable when you return to build the course weeks later and can't remember why Module 4 comes before Module 5.

Claude can organize information into a teachable sequence. What it can't do is make the judgment calls that separate a good course from a forgettable one. Those calls are yours.

Sequencing by motivation, not just logic. Claude sequences by conceptual dependency — what must the student know before they can learn the next thing. But students also need motivation sequencing. A technically logical order that puts all the hard foundational work first and all the satisfying application at the end will lose students by week two. You know your students. You know which early wins will keep them going. Rearrange the AI's output to include at least one "I did something real" moment in the first module.

Calibrating difficulty. Claude doesn't know what's actually hard for your specific students. A lesson it labels "15 minutes" might take your audience 45 minutes because the concept conflicts with a common misconception in your field. You've seen people struggle with these ideas in real life. Adjust the time estimates and the depth of scaffolding based on your experience teaching this material to real humans.

Adding your perspective. The outline Claude produces is structurally sound but generic. What makes your course worth taking is your point of view — the things you emphasize that other instructors skip, the mistakes you've seen people make, the frameworks you've developed through experience. After you have the structure, go through each lesson and annotate it with what only you can teach.

Don't start a new chat for each module. Claude's context window is large enough to hold your entire outline conversation. Keeping everything in one thread means the AI remembers your course outcome, student profile, and previous decisions as you iterate.

If you have FAQ documents, support emails, or forum posts from past students or clients, paste those in too. They help Claude identify where students get stuck, which should inform where your course spends extra time.

"Building Your First Arrangement" is better than "Arranging Basics." Modules that sound like a journey encourage students to keep moving forward. Ask Claude to rename generic module titles into action-oriented ones.

Tell Claude your target course length upfront. "The entire course should be 8-10 hours of content" gives the AI a constraint that prevents scope creep. Without it, outlines tend to grow longer than any student will complete.

Claude tends to create broad surveys rather than deep dives. A 10-module outline where each module has two thin lessons is less useful than a 4-module outline where each module has substantive, well-scaffolded content. If the first draft has more than 6 modules, ask Claude to consolidate.

The practice activities Claude suggests are often reasonable but bland — "create a plan," "write a reflection," "share with a partner." Your students deserve activities tied to real scenarios in their specific context. After generating the initial activities, rewrite at least the first activity in each module to reference a concrete situation your students will recognize.

Left to its own devices, Claude will make every module the same length with the same number of lessons. Real courses aren't symmetrical. Some concepts need more time. Some modules have a natural "do a project" lesson that takes longer than a concept lesson. Review the outline for forced symmetry and let modules be different sizes based on what the content actually requires.

Claude doesn't plan for how students willfeel at different points in the course. The middle modules are where motivation dips. The last module should feel like a culmination, not just "more lessons." These emotional design choices aren't something AI handles — they come from your empathy with your students and your experience running courses.