Hahahahaha, oh yea, the Year of the Linux Desktop - been hearing that since 2000.

Look, Linux is a brilliant server OS, brilliant embedded/dedicated device OS. But as a desktop it's very problematic. It can be a great desktop OS, if you have very clear requirements, and know those requirements will never increase to include things from the Windows world, especially for end users who frankly have little tech understanding of any OS.

The very first issue they're going to run into is Excel. None of the open-source competitors come close, and any business/analysis will use it.

Oh, just RDP to a virtual desktop? Then why bother with Linux, which merely adds a layer of complexity?

The support issues/costs will skyrocket.

What they did a few months ago made more sense - use it in a small group, one with nominal requirements. Run that for a couple years, then expand a little. This gives IT time to flesh out issues and ensures they only get a small set of issues, affecting a small group, at any one time. It also allows for changing distro or shell with far less effort.

I've done this multiple times with other types of systems - run the new in a small group, not replacing the old entirely. Over time we identify pain points, develop work around or new architectural approaches, so when we finally full switch over and retire the old system it's a minor change.