The question that reframed the design
Not "how do we get people to use the ticket system" but "why is sending a message easier". Once the second question is asked honestly, the answer is usually obvious and slightly embarrassing: because it is.
Every internal IT team runs two queues. There is the official one, in whatever system was chosen, and there is the real one: messages, corridor conversations, and requests to whoever is nearest and looks least busy.
The shadow queue gets the work done and destroys everything else. The team looks less loaded than it is, recurring problems never surface as patterns, nothing can be planned or staffed properly, and the person who is most approachable absorbs the most interruptions.
I attacked the friction rather than the behaviour. I made submission genuinely short, did not require the user to classify their own problem into a taxonomy they have never seen, and let the team add structure on receipt where they have the context to do it correctly.
A ticketing system designed around the occasional user rather than the IT team's taxonomy, with classification pushed to triage. Once requests consistently arrive through one path, operational visibility follows for free.
Not "how do we get people to use the ticket system" but "why is sending a message easier". Once the second question is asked honestly, the answer is usually obvious and slightly embarrassing: because it is.
What it runs on