i worked in corporate IT previously, software development currently, and specialise in linux, so i almost never need money or hardware to do a proof of concept, just some time to read and write.
i’ll take the problem and dream up some minimal criteria for the fix, maybe rank them by importance. then dream up a managable number of reasonable options. those options get examined, and everything is written down. then the everything written down is examined, and shit options get tossed. the not-shit options are taken and argued over with people. qualifications for the argument are minimal: they have an opinion or they’re a user. everyone’s points are written down, the best options are noted.
then i show all best and almost-best options to the person least likely to understand the technical parts. this person’s pick(s) for best option(s) is written down, along with their reasoning. (the idea is that they might be more likely to focus on things i might not.)
finally, i pick my favourite and write it down under the heading “Recommendation”.