When you said "a broader topic", did you mean what algorithms to use to examine the event log with the goal of reducing future calls from the same customer on the same topic? In other words, a customer may call for help for a different topic. If a customer calls in for the same topic repetitively, something can be improved.
You may get ideas from a Coursera class Processing Mining since the issue you're solving is similar to the example of a spaghetti process in lecture 6.7:
"Spaghetti process describing the diagnosis and treatment of 2765 patients in a Dutch hospital. The process model was constructed based on an event log containing 114,592 events. There are 619 different activities (taking event types into account) executed by 266 different individuals (doctors, nurses, etc.)."
By the way, you can click the drop down menu under "Sessions", choose "April 1, 2015 to May 19, 2015" then you can register and view the lectures. On the right of each lecture, there are some icons. The first icon is to download slides for the lecture. You may find reading the slides is faster than listening to the lecture.
Suppose a customer called for installation of software release 1.5 then called a day later for running the new features of software release 1.5. Are these two issues logged as two of the 200 call categories? If so, we can use a time period (say one week) to judge whether it is about the same topic. Within a short time period, a customer is likely to work on the same topic, especially with the same key words such as "software release 1.5". We can coach the call center employees on solving possible follow up questions by saying something like "now that we finished installation, let me show you a couple of new features. It'll save you time." This will reduce the number of calls on the same topic from the same customer.