Data Mining - Can I borrow your collective brains for a bit?

So, let's say I am in a fairly restrictive environment which means user do not have global read access to all data.  200+ Product contexts containing a complete list of CAD, Parts, BOMs and document content (Program data). So, when I search for things, I only get back what I currently have read access to at that time. Works as it should.

But let's say that I need to do some data mining. I am in Engineering and I am looking to see if we've done some type of analysis or a report from user X or if we solved a failure previously on an old program. This body of content has huge potential for Engineering to learn from past as opposed to people keeping it in their heads.

So, is there a way for users to search (thinking SOLR here) for content, get hits that it exists BUT prevent them accessing content directly?  I understand that this does open crack in security wall for exploit, trade offs I guess. In my head I see that it says I got a hit on doc XYZ but user lacks access to it. Then they should follow proper channels to get access to what they are looking for.  It might be possible with a generic login account specifically configured, for example, to prevent download but can see all meta data and indexing. Thoughts?