What we need is a global security standard for support staff, with a template as to what information is accessible by staff and what isn't. And what is available to better trained 2nd-level support, etc.
And then each company can say they are certified for this particular security standard, and then you can't get social engineering attacks where you attack one large corporation, get partial information, and then feed that into another large organization to get other information. This was done previously using Amazon, again, to get enough information to take someone's Twitter account, if i remember correctly.