Keeping course content current with AI

Courses go stale quietly. A software screen changes and your screenshots no longer match. A rule gets updated. A marketplace changes its fees. An example that felt fresh now references a tool nobody uses. None of these problems announces itself, but learners notice, and each small mismatch chips away at their trust in everything else you teach.

Most owners know they should review their courses regularly and rarely get round to it, because rereading hours of lessons is tedious. AI can take on the tedious part: scanning transcripts for time-sensitive claims, sorting learner questions and drafting updated passages. The part it cannot take on is deciding what is true now. That stays with you.

Make it a routine, not a rescue

The easiest way to keep a course current is to break maintenance into small, regular checks instead of an occasional overhaul. A simple rhythm works for most memberships:

• Every few months: a quick scan of each course for obvious problems and a look at recent learner questions.
• Once a year: a deeper review of every lesson, with decisions about what to update, re-record or retire.
• When something changes: an immediate check whenever you hear of a change in your field, such as new rules, a redesigned tool or a price change.

Carmen runs a course on selling handmade goods online. Her content depends on marketplace settings, fees and policies that change without warning, so she keeps a list of every lesson that mentions a specific platform. When one of them changes, she knows exactly which lessons to open. For the non-AI side of building this schedule, see keeping course content up to date.

Let AI find what might be stale

Start with transcripts or lesson text, since AI works best with words. If your videos lack transcripts, a transcription tool can produce them. Then ask an assistant to flag anything that could date. Tools that let you upload several documents, such as NotebookLM, Claude Projects or a custom GPT, are useful when a course spans many files; asking AI questions about your own documents covers how that works.

Below are the transcripts for [course name], a course for [describe learners]. Read them and list every statement that could become out of date. Include: named tools, apps and websites; step-by-step instructions for software; prices, fees and figures; rules, regulations and official requirements; references to recent events or trends; and words like “new,” “currently” or “recently.”

For each item, give the lesson name, quote the statement and say what kind of change could make it wrong. Do not tell me whether it is still accurate; I will check that myself. [paste transcripts]

The final instruction is deliberate. A general assistant may not know about recent changes and can confidently tell you an outdated statement is still correct. You want a list of things to check, not a verdict. Web-connected assistants can help you research each item, but always follow the trail back to the original source, such as the official help page, the regulator's site or the vendor's documentation.

Mine learner questions for hidden problems

Your learners are already telling you what is out of date, often indirectly. A question like “I can't find the settings button you mention” usually means the screen has changed. Collect questions from comments, support messages and your community, remove names and personal details, and ask the AI to group them.

Ask for groups with example questions and a count, and ask the assistant to separate confusion caused by the lesson itself from confusion caused by something in the outside world changing. Then read a sample of the original questions in each group to confirm the grouping makes sense. Clusters that point to a single lesson are your most urgent updates.

Draft updates, then verify them

Once you know what needs changing, AI can help you draft the replacement text quickly. Give it the old passage, the verified new information and your lesson's tone, and ask for a revised version that changes as little as possible. Then check every fact in the draft against your source. Assistants sometimes slip in extra claims that were not in the information you gave them, and why AI makes things up explains why.

For each flagged item, decide the lightest fix that works:

• Note: a short text update beneath a video, for small changes.
• Patch: a revised section of text or a replacement screenshot.
• Re-record: when the core teaching is wrong or badly out of date.
• Retire: when the lesson no longer serves anyone.

The same approach works for articles and other content outside your courses, which refreshing old content with AI help covers.

Tell learners what changed

Keep a simple change log for each course: the date, the lesson and what changed. Share a short update note with learners when you make significant changes, so members who took the course earlier know to revisit a lesson. Visible maintenance reassures members that the course is alive and looked after.

Your maintenance checklist

1. Set a recurring reminder for a quick review every few months and a deep review once a year.
2. Gather transcripts for every lesson in one course.
3. Run the stale-content prompt and turn the output into a checklist.
4. Verify each item against an original, official source.
5. Group recent learner questions to find hidden problems.
6. Draft fixes with AI, check every fact yourself and choose the lightest fix that works.
7. Record every change and tell learners about the important ones.