I didn't notice this reply in my comments, sorry for the delay. My idea is to stop spam at its origin, not the recipient. Spammers are quite easy to identify: thousands of messages with forged headers are sent from a single host. Even if the headers are legitimate, numerous abuse reports should give the spammer's provider reason to investigate. If an ISP doesn't take action, their provider should stop their ability to send mail.
Is the provider in this case authorized to inspect company confidential e-mail when they're not storing it on their servers?
Inspecting the contents of the e-mail wouldn't be necessary, all that is needed is the header information, which isn't confidential.
Host names can be forged. Determining the point of origin of some spam is not as easy as it seems.