The key is getting good text recognition instead of an un-parsable image. And of course, your current database of documents, metadata, etc
Me being a fan of Free software would look for a Open solution adn then if needed look for a company or individual to support your implementation for a one time fee.
Crappy arictle, decent list of starting points w/ features -
https://www.how2shout.com/tools/best-free-open-source-document-management-system-software.html
Search terms like "open source document management" should get you plenty of other ideas, or ask on
http://softwarerecs.stackexchange.com