In both my companies I've had issues replicating issues for specific users/accounts.
Seems to me to options are
1. Copy all of prod to dev
2. Maintain custom scripts to pull a significant subset
3. Override env variables to actually connect to prod & try not to mess up
Then similarly with user-emulation the only option I can see is a hacky script.
I'm considering building tooling around this & just curious how this works in other companies.
If you've really got to do it, write a script that emulates prod data with fake made up examples.