Another aspect is that IT may try to process data using tools that would do it properly, for example writing custom code where there is a single point of truth for the source of the data and a defined process for changing anything, which is more stable and likely to perform better in the long run. However, building something yourself is a lot faster because you don't have to find someone with the programming skills to make that, nor go through the process of explaining clearly to the programmer what you need and answering all those questions. Developing a system that's high quality often takes time that the group concerned doesn't want to spend, and hacks like Excel end up bridging the gap.
As a person who sometimes ends up writing the code for this part*, I have sympathy with the groups here. When I write code that takes data from a form they already have and performs some simple transformations on it before giving it back in another straightforward form, it often seems like there should be a way to do this that's somewhere between an Excel spreadsheet with zero error checking and a custom-built program with a lot of manually-written error checking with an environment in which to run it. I haven't yet seen a low code system that I'm confident does this well, but I definitely understand the desire for it.
* My job mostly involves writing more complex programs, but if they need some plumbing, it can get assigned to me as with anyone else. I also do volunteer work and this is by far the largest subset of programming tasks that charities make of me.