The lessons worth keeping unchanged are the ones where your teaching still produces the result — where the concept is timeless, your examples still land, and students consistently get unstuck after watching or reading it. AI can help you figure out which ones those are faster than reading through everything yourself.
Why This Is Harder Than It Sounds
When you’ve built a course yourself, you’re too close to it to evaluate it clearly. You know why every lesson is there, so everything feels essential. It’s like asking someone to declutter their own house — they can justify keeping every single thing because they remember why they got it.
AI doesn’t have that attachment. You can paste a lesson transcript or outline into Claude and ask it to evaluate whether the core concept still applies, whether the examples are dated, and whether a student who had never seen your content before would find it complete on its own. That outside perspective is what makes the process useful.
How to Run the Evaluation
Start by giving Claude the context it needs: who your students are, what result the course is meant to produce, and what the lesson is supposed to do. Then paste in the lesson outline, transcript, or slides and ask it to tell you whether the lesson still achieves its stated goal.
A prompt that works well: “Here is a lesson from my course on [topic]. The student should leave this lesson able to [outcome]. Does this lesson still accomplish that in 2026? What, if anything, would make it outdated?” Claude will flag things like specific tool names that no longer exist, statistics that are now stale, or instructions that assume software interfaces that have changed. If Claude finds nothing to flag, that lesson is a strong candidate to keep as-is.
Run each lesson through this check and keep a simple list: Keep, Update, or Rebuild. The ones that come back clean across all three checks — concept, examples, outcome — are your keepers.
What This Means for Educators
As a coach or educator, your most valuable content isn’t usually the tactical how-to stuff — it’s the foundational lessons where you explain a concept, share a framework, or tell a story that reframes how a student sees the problem. Those lessons age slowly, if at all. AI evaluation tends to confirm this: the lessons that get the “keep” rating are almost always the ones built on principles rather than platform-specific instructions.
Protecting those lessons from unnecessary rework saves you time and preserves the parts of your course that are genuinely yours — your voice, your frameworks, your hard-won perspective.
The Simple Rule
If AI can’t find a reason to change it, don’t change it. A lesson that still produces the right result for the right student is doing its job — leave it alone and spend your energy on the lessons that actually need work.