Overly simplified example:
"Can you make this button do X?" where the existing button in so many ways is only distantly connected to X. And then they get stuck on the idea that THAT button has to be where the thing happens, and they stick with it even if you explain that the usual function of that button is Y.
I simplified it saying button, but this applies to processes and other things. I think users sometimes think picking a common thing, button or process that sort of does what they want is the right entry point to discuss changes and maybe they think that somehow saves time / developer effort. Where in reality, just a new button is in fact an easier and less risky place to start.
I didn't say that very well, but I wonder if that plays a part in the endless adding of complexity to UI where users grasp onto a given button, function, or process and "just" want to alter it a little ... and it never ends until it all breaks down.
In other words, if you need expert help, trust the expert. Ask for what you need, not how to do it.
But sometimes power structures don't allow for it. I worked tech support in a number of companies. At some companies we were empowered to investigate and solve problems... sometimes that took work, and work from the customer. It had much better outcomes for the customer, but fixes were not quick. Customers / companies with good technical staff in management understood that dynamic.
Other companies were "just fix it" and tech support were just annoying drones and the company and customer's and management treated tech support as annoying drones. They got a lot more "you got exactly what you asked for" treatment ... because management and even customers will take the self defeating quick and easy path sometimes.
On the other hand, if you are using your car for a decade and feel it needs a new belt - then get a new belt. Worst case scenario- you will lose some money but learn a bit more about an item you use everyday.
Experts don't have your instincts as a user.
So when they would come in asking for a specific part to be replaced with no context I used to tell them that we wouldn't do that until we did a diagnosis. This is because if we did do as they asked and, like in most cases, it turned out that they were wrong they would then become indignant and ask why we didn't do diagnosis for free to tell them that they were wrong.
Diagnosis takes time and, therefore, costs money. If the user was capable of it then they would also be capable enough to carry out the repair. If they're capable of carrying out the diagnosis and the repair then they wouldn't be coming to me. This has proved to be true over many years for everyone from kids with their first car to accountants and even electrical engineers working on complex systems for large corporations as their occupation. That last one is particularly surprising considering that an engineer should know the bounds of their knowledge and understand how maintenance, diagnosis and repair work on a conceptual level.
Don't trust your instincts in areas where you have no understanding. Either learn and gain the understanding or accept that paying an expert is part of owning something that requires maintenance and repair.
Sometimes the best solution is not the most widely-encouraged one.
The point of XY problems isn't to call people out on supposedly bad behaviour, it's to push them in the right direction and provide more context.
I generally try to answer the Y but also indicate that it suggests there may be an X that could be better achieved some other way, and mention Z if I'm reasonably confident in what X is. It might increase the chance that the person asking just does Y anyway even if Z would be better, but frankly that's not really my business.
Sometimes it's just me firing up some SQL queries and discovering "Well this happened 3 times ... ever ..." and we do nothing ;)
In my experience it may be solved by both parties spending the effort and time to first understand what is being asked... assuming they are both willing to stomach the costs. Sometimes it isn't worth it, and it's easier to pacify than respectfully and carefully dig.
