Just a thought, if you were company xyz and you just scanned 8 million documents for SSN and other redaction information. How hard would it be to automatically catalogue and field those documents? I'll tell you. It's not hard at all once the document has been "taught" to the computer. Let's say lender xyz also has it's legal descriptions on page 3 of their loan at coordinates x,y. Now the computer knows every time the lender pops up on the first page the computer knows where to find all the relevant indexing information.
Getting to my point. Not only does the software company get to scan the documents, but now has a very valueable bit of information. As the Land America VP said in San Diego at the ATIM conference, data is king. What is keeping these companies from turning around and selling their data? Now they could be tripple dipping. All they have to do is update their files every week and now they have a complete county record.

George