Creating Help Center content from ticket data

Your tickets already contain the articles you have not written. Turning them into content is the highest-value use of AI here.

Your tickets already contain the articles you have not written. Turning them into content is the highest-value use of AI here.

The idea

A ticket where an agent worked something out and explained it is a draft article. Somebody asked the question, somebody answered it, and the answer is sitting in a conversation nobody else will ever read.

Generating an article from it turns a one-off answer into something reusable.

How to work

1. Find tickets worth turning into articles.
2. Generate a draft from the ticket.
3. Edit it: remove the specifics of that customer, add the general case.
4. Have somebody who knows check it.
5. Publish and link it from related articles.

Which tickets

• Questions that recur. Your ticket subjects tell you which.
• Anything that took an agent a long time to work out. High value, written once.
• Answers agents copy from each other. That is an article waiting to be written.

Remove the customer

The editing step that matters most. A generated draft carries their name, their order number, their specific situation. All of that has to go, and it is easy to miss a reference halfway down.

See also

• Creating content with generative AI
• Which macros nobody uses: measuring macro usage
• Which subjects come up most often