The same questions, over and over, answered by hand. The obvious answer is to write articles about them, and that's right, but not for all of them, and writing the wrong ones is how teams end up with a bigger help center and exactly the same ticket volume.
Most of the work here is deciding what not to write.
Step 1 of 4
Find what you actually answer most
Subject lines and first messages are enough. You don't need an export, a tagging system or a data project, and waiting until you have one is the reason this never gets done.
Step 2 of 4
Separate the docs problems from the product problems
This is the step that decides whether any of this works. Some repeated questions aren't documentation failures at all. If 40 people a week ask where a setting is, an article about where the setting is documents a bug. You'll have written it, the tickets won't stop, and you'll conclude documentation doesn't work.
Step 3 of 4
Turn the real ones into articles
A solved ticket is most of an article already. The question is in the customer's words, the answer is in your agent's words, and both are better source material than anything you'd write from scratch.
Step 4 of 4
Check it answers the question as asked
An article that answers a slightly different question reduces tickets by nothing. It's the most common way this work fails, and it's invisible unless you check deliberately, because the article looks fine.
Of these 4 steps, 3 are free. If the ticket volume is the thing keeping you up, start at step 1 today and you'll know by tomorrow which questions are worth writing about. Turning questions into pages is taught in the course Deciding what goes where.
Start at step 1, it's free →