These are helpful reminders, appreciate it. Right now I'm also just navigating through testing myself before sharing with our leadership team. Asking deeper questions during this process is of course helpful, I plan to dig into each problem they want to solve with AI and why.

I really was just curious about any interesting gotchas anyone else has run into. Historically, Microsoft is a balance between "this is great because it's already here, we just need to turn on a button to enable it" and the quality or reliability being not what we expected, to put it mildly. That and onboarding is a headache for external tools, but is often a tiny fraction of long-term maintenance so isn't necessarily the highest-weighted issue.