I would recommend some kind of microflow that will batch handle all your data and remove the names.
However, unless you process the data on the Production side you will still at some point have the sensitive data on accp or test.
Can I ask why you need to production database on test and accp, could you create something that adds a load of data to test and accp for you?
In our case, the sensitive data in our production environment is transaction-data and comes from outside, through middleware. After a interface document gets sent to test, we have our middleware send the same documents also to our test-system but first replacing the sensitive parts with anonymous variants, filling test-environment with nice and clean data. Works like a charm and not very difficult to set up. Also: we never have to restore prod-backups anymore.